diff options
| author | Patrick Wendell <pwendell@gmail.com> | 2013-07-31 21:35:12 -0700 |
|---|---|---|
| committer | Patrick Wendell <pwendell@gmail.com> | 2013-07-31 21:35:12 -0700 |
| commit | 5cc725a0e3ef523affae8ff54dd74707e49d64e3 (patch) | |
| tree | ebd1698333d2df4194f17a9ea93a2f2eac2c7acd /core | |
| parent | b7b627d5bb1a1331ea580950834533f84735df4c (diff) | |
| parent | f3cf09491a2b63e19a15e98cf815da503e4fb69b (diff) | |
| download | spark-5cc725a0e3ef523affae8ff54dd74707e49d64e3.tar.gz spark-5cc725a0e3ef523affae8ff54dd74707e49d64e3.tar.bz2 spark-5cc725a0e3ef523affae8ff54dd74707e49d64e3.zip | |
Merge branch 'master' into ec2-updates
Conflicts:
ec2/deploy.generic/root/mesos-ec2/ec2-variables.sh
Diffstat (limited to 'core')
346 files changed, 15005 insertions, 3461 deletions
diff --git a/core/pom.xml b/core/pom.xml index da26d674ec..f0c936c86a 100644 --- a/core/pom.xml +++ b/core/pom.xml @@ -1,4 +1,21 @@ <?xml version="1.0" encoding="UTF-8"?> +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one or more + ~ contributor license agreements. See the NOTICE file distributed with + ~ this work for additional information regarding copyright ownership. + ~ The ASF licenses this file to You under the Apache License, Version 2.0 + ~ (the "License"); you may not use this file except in compliance with + ~ the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, software + ~ distributed under the License is distributed on an "AS IS" BASIS, + ~ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + ~ See the License for the specific language governing permissions and + ~ limitations under the License. + --> + <project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd"> <modelVersion>4.0.0</modelVersion> <parent> @@ -32,16 +49,22 @@ <artifactId>compress-lzf</artifactId> </dependency> <dependency> - <groupId>asm</groupId> - <artifactId>asm-all</artifactId> + <groupId>org.ow2.asm</groupId> + <artifactId>asm</artifactId> </dependency> <dependency> <groupId>com.google.protobuf</groupId> <artifactId>protobuf-java</artifactId> </dependency> <dependency> - <groupId>de.javakaffee</groupId> - <artifactId>kryo-serializers</artifactId> + <groupId>com.twitter</groupId> + <artifactId>chill_2.9.3</artifactId> + <version>0.3.0</version> + </dependency> + <dependency> + <groupId>com.twitter</groupId> + <artifactId>chill-java</artifactId> + <version>0.3.0</version> </dependency> <dependency> <groupId>com.typesafe.akka</groupId> @@ -56,43 +79,56 @@ <artifactId>akka-slf4j</artifactId> </dependency> <dependency> - <groupId>it.unimi.dsi</groupId> - <artifactId>fastutil</artifactId> - </dependency> - <dependency> - <groupId>colt</groupId> - <artifactId>colt</artifactId> + <groupId>org.scala-lang</groupId> + <artifactId>scalap</artifactId> </dependency> <dependency> - <groupId>cc.spray</groupId> - <artifactId>spray-can</artifactId> + <groupId>org.scala-lang</groupId> + <artifactId>scala-library</artifactId> </dependency> <dependency> - <groupId>cc.spray</groupId> - <artifactId>spray-server</artifactId> + <groupId>net.liftweb</groupId> + <artifactId>lift-json_2.9.2</artifactId> </dependency> <dependency> - <groupId>cc.spray</groupId> - <artifactId>spray-json_${scala.version}</artifactId> + <groupId>it.unimi.dsi</groupId> + <artifactId>fastutil</artifactId> </dependency> <dependency> - <groupId>org.tomdz.twirl</groupId> - <artifactId>twirl-api</artifactId> + <groupId>colt</groupId> + <artifactId>colt</artifactId> </dependency> <dependency> <groupId>com.github.scala-incubator.io</groupId> - <artifactId>scala-io-file_${scala.version}</artifactId> + <artifactId>scala-io-file_2.9.2</artifactId> </dependency> <dependency> <groupId>org.apache.mesos</groupId> <artifactId>mesos</artifactId> </dependency> <dependency> + <groupId>io.netty</groupId> + <artifactId>netty-all</artifactId> + </dependency> + <dependency> <groupId>log4j</groupId> <artifactId>log4j</artifactId> </dependency> + <dependency> + <groupId>com.codahale.metrics</groupId> + <artifactId>metrics-core</artifactId> + </dependency> + <dependency> + <groupId>com.codahale.metrics</groupId> + <artifactId>metrics-jvm</artifactId> + </dependency> <dependency> + <groupId>org.apache.derby</groupId> + <artifactId>derby</artifactId> + <scope>test</scope> + </dependency> + <dependency> <groupId>org.scalatest</groupId> <artifactId>scalatest_${scala.version}</artifactId> <scope>test</scope> @@ -162,10 +198,6 @@ </environmentVariables> </configuration> </plugin> - <plugin> - <groupId>org.tomdz.twirl</groupId> - <artifactId>twirl-maven-plugin</artifactId> - </plugin> </plugins> </build> diff --git a/core/src/hadoop1/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala b/core/src/hadoop1/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala index f286f2cf9c..25386b2796 100644 --- a/core/src/hadoop1/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala +++ b/core/src/hadoop1/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package org.apache.hadoop.mapred trait HadoopMapRedUtil { diff --git a/core/src/hadoop1/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala b/core/src/hadoop1/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala index 264d421d14..b1002e0cac 100644 --- a/core/src/hadoop1/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala +++ b/core/src/hadoop1/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package org.apache.hadoop.mapreduce import org.apache.hadoop.conf.Configuration diff --git a/core/src/hadoop1/scala/spark/deploy/SparkHadoopUtil.scala b/core/src/hadoop1/scala/spark/deploy/SparkHadoopUtil.scala index a0fb4fe25d..617954cb98 100644 --- a/core/src/hadoop1/scala/spark/deploy/SparkHadoopUtil.scala +++ b/core/src/hadoop1/scala/spark/deploy/SparkHadoopUtil.scala @@ -1,5 +1,23 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.mapred.JobConf /** @@ -20,4 +38,10 @@ object SparkHadoopUtil { // Return an appropriate (subclass) of Configuration. Creating config can initializes some hadoop subsystems def newConfiguration(): Configuration = new Configuration() + + // add any user credentials to the job conf which are necessary for running on a secure Hadoop cluster + def addCredentials(conf: JobConf) {} + + def isYarnMode(): Boolean = { false } + } diff --git a/core/src/hadoop2-yarn/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala b/core/src/hadoop2-yarn/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala index 875c0a220b..0f972b7a0b 100644 --- a/core/src/hadoop2-yarn/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala +++ b/core/src/hadoop2-yarn/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala @@ -1,4 +1,21 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package org.apache.hadoop.mapred import org.apache.hadoop.mapreduce.TaskType diff --git a/core/src/hadoop2-yarn/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala b/core/src/hadoop2-yarn/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala index 8bc6fb6dea..1a7cdf4788 100644 --- a/core/src/hadoop2-yarn/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala +++ b/core/src/hadoop2-yarn/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package org.apache.hadoop.mapreduce import org.apache.hadoop.conf.Configuration diff --git a/core/src/hadoop2-yarn/scala/spark/deploy/SparkHadoopUtil.scala b/core/src/hadoop2-yarn/scala/spark/deploy/SparkHadoopUtil.scala index ab1ab9d8a7..6122fdced0 100644 --- a/core/src/hadoop2-yarn/scala/spark/deploy/SparkHadoopUtil.scala +++ b/core/src/hadoop2-yarn/scala/spark/deploy/SparkHadoopUtil.scala @@ -1,6 +1,24 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy import collection.mutable.HashMap +import org.apache.hadoop.mapred.JobConf import org.apache.hadoop.security.UserGroupInformation import org.apache.hadoop.yarn.conf.YarnConfiguration import org.apache.hadoop.conf.Configuration @@ -27,18 +45,7 @@ object SparkHadoopUtil { } def runAsUser(func: (Product) => Unit, args: Product, user: String) { - - // println("running as user " + jobUserName) - - UserGroupInformation.setConfiguration(yarnConf) - val appMasterUgi: UserGroupInformation = UserGroupInformation.createRemoteUser(user) - appMasterUgi.doAs(new PrivilegedExceptionAction[AnyRef] { - def run: AnyRef = { - func(args) - // no return value ... - null - } - }) + func(args) } // Note that all params which start with SPARK are propagated all the way through, so if in yarn mode, this MUST be set to true. @@ -60,4 +67,10 @@ object SparkHadoopUtil { // Return an appropriate (subclass) of Configuration. Creating config can initializes some hadoop subsystems // Always create a new config, dont reuse yarnConf. def newConfiguration(): Configuration = new YarnConfiguration(new Configuration()) + + // add any user credentials to the job conf which are necessary for running on a secure Hadoop cluster + def addCredentials(conf: JobConf) { + val jobCreds = conf.getCredentials(); + jobCreds.mergeAll(UserGroupInformation.getCurrentUser().getCredentials()) + } } diff --git a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ApplicationMaster.scala b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ApplicationMaster.scala index ae719267e8..1b06169739 100644 --- a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ApplicationMaster.scala +++ b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ApplicationMaster.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.yarn import java.net.Socket @@ -27,26 +44,10 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) e private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) private var yarnAllocator: YarnAllocationHandler = null + private var isFinished:Boolean = false def run() { - // Initialization - val jobUserName = Utils.getUserNameFromEnvironment() - logInfo("running as user " + jobUserName) - - // run as user ... - UserGroupInformation.setConfiguration(yarnConf) - val appMasterUgi: UserGroupInformation = UserGroupInformation.createRemoteUser(jobUserName) - appMasterUgi.doAs(new PrivilegedExceptionAction[AnyRef] { - def run: AnyRef = { - runImpl() - return null - } - }) - } - - private def runImpl() { - appAttemptId = getApplicationAttemptId() resourceManager = registerWithResourceManager() val appMasterResponse: RegisterApplicationMasterResponse = registerApplicationMaster() @@ -85,10 +86,7 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) e // Wait for the user class to Finish userThread.join() - - // Finish the ApplicationMaster - finishApplicationMaster() - // TODO: Exit based on success/failure + System.exit(0) } @@ -141,30 +139,30 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) e } } } - + private def startUserClass(): Thread = { logInfo("Starting the user JAR in a separate Thread") val mainMethod = Class.forName(args.userClass, false, Thread.currentThread.getContextClassLoader) .getMethod("main", classOf[Array[String]]) val t = new Thread { override def run() { - var mainArgs: Array[String] = null - var startIndex = 0 - - // I am sure there is a better 'scala' way to do this .... but I am just trying to get things to work right now ! - if (args.userArgs.isEmpty || args.userArgs.get(0) != "yarn-standalone") { - // ensure that first param is ALWAYS "yarn-standalone" - mainArgs = new Array[String](args.userArgs.size() + 1) - mainArgs.update(0, "yarn-standalone") - startIndex = 1 - } - else { - mainArgs = new Array[String](args.userArgs.size()) + var successed = false + try { + // Copy + var mainArgs: Array[String] = new Array[String](args.userArgs.size()) + args.userArgs.copyToArray(mainArgs, 0, args.userArgs.size()) + mainMethod.invoke(null, mainArgs) + // some job script has "System.exit(0)" at the end, for example SparkPi, SparkLR + // userThread will stop here unless it has uncaught exception thrown out + // It need shutdown hook to set SUCCEEDED + successed = true + } finally { + if (successed) { + ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) + } else { + ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.FAILED) + } } - - args.userArgs.copyToArray(mainArgs, startIndex, args.userArgs.size()) - - mainMethod.invoke(null, mainArgs) } } t.start() @@ -209,7 +207,7 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) e logInfo("All workers have launched.") // Launch a progress reporter thread, else app will get killed after expiration (def: 10mins) timeout - if (userThread.isAlive){ + if (userThread.isAlive) { // ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapse. val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) @@ -227,7 +225,7 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) e val t = new Thread { override def run() { - while (userThread.isAlive){ + while (userThread.isAlive) { val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning if (missingWorkerCount > 0) { logInfo("Allocating " + missingWorkerCount + " containers to make up for (potentially ?) lost containers") @@ -265,14 +263,23 @@ class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) e } } */ - - def finishApplicationMaster() { + + def finishApplicationMaster(status: FinalApplicationStatus) { + + synchronized { + if (isFinished) { + return + } + isFinished = true + } + + logInfo("finishApplicationMaster with " + status) val finishReq = Records.newRecord(classOf[FinishApplicationMasterRequest]) .asInstanceOf[FinishApplicationMasterRequest] finishReq.setAppAttemptId(appAttemptId) - // TODO: Check if the application has failed or succeeded - finishReq.setFinishApplicationStatus(FinalApplicationStatus.SUCCEEDED) + finishReq.setFinishApplicationStatus(status) resourceManager.finishApplicationMaster(finishReq) + } } @@ -286,7 +293,7 @@ object ApplicationMaster { private val ALLOCATOR_LOOP_WAIT_COUNT = 30 def incrementAllocatorLoop(by: Int) { val count = yarnAllocatorLoop.getAndAdd(by) - if (count >= ALLOCATOR_LOOP_WAIT_COUNT){ + if (count >= ALLOCATOR_LOOP_WAIT_COUNT) { yarnAllocatorLoop.synchronized { // to wake threads off wait ... yarnAllocatorLoop.notifyAll() @@ -321,14 +328,16 @@ object ApplicationMaster { logInfo("Invoking sc stop from shutdown hook") sc.stop() // best case ... - for (master <- applicationMasters) master.finishApplicationMaster + for (master <- applicationMasters) { + master.finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) + } } } ) } // Wait for initialization to complete and atleast 'some' nodes can get allocated yarnAllocatorLoop.synchronized { - while (yarnAllocatorLoop.get() <= ALLOCATOR_LOOP_WAIT_COUNT){ + while (yarnAllocatorLoop.get() <= ALLOCATOR_LOOP_WAIT_COUNT) { yarnAllocatorLoop.wait(1000L) } } diff --git a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ApplicationMasterArguments.scala b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ApplicationMasterArguments.scala index dc89125d81..8de44b1f66 100644 --- a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ApplicationMasterArguments.scala +++ b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ApplicationMasterArguments.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.yarn import spark.util.IntParam @@ -69,7 +86,6 @@ class ApplicationMasterArguments(val args: Array[String]) { " --class CLASS_NAME Name of your application's main class (required)\n" + " --args ARGS Arguments to be passed to your application's main class.\n" + " Mutliple invocations are possible, each will be passed in order.\n" + - " Note that first argument will ALWAYS be yarn-standalone : will be added if missing.\n" + " --num-workers NUM Number of workers to start (Default: 2)\n" + " --worker-cores NUM Number of cores for the workers (Default: 1)\n" + " --worker-memory MEM Memory per Worker (e.g. 1000M, 2G) (Default: 1G)\n") diff --git a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/Client.scala b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/Client.scala index 7a881e26df..8bcbfc2735 100644 --- a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/Client.scala +++ b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/Client.scala @@ -1,9 +1,30 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.yarn import java.net.{InetSocketAddress, URI} +import java.nio.ByteBuffer import org.apache.hadoop.conf.Configuration import org.apache.hadoop.fs.{FileStatus, FileSystem, Path} +import org.apache.hadoop.mapred.Master import org.apache.hadoop.net.NetUtils +import org.apache.hadoop.io.DataOutputBuffer +import org.apache.hadoop.security.UserGroupInformation import org.apache.hadoop.yarn.api._ import org.apache.hadoop.yarn.api.records._ import org.apache.hadoop.yarn.api.protocolrecords._ @@ -23,6 +44,7 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl var rpc: YarnRPC = YarnRPC.create(conf) val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + val credentials = UserGroupInformation.getCurrentUser().getCredentials(); def run() { init(yarnConf) @@ -40,8 +62,8 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl appContext.setQueue(args.amQueue) appContext.setAMContainerSpec(amContainer) - appContext.setUser(args.amUser) - + appContext.setUser(UserGroupInformation.getCurrentUser().getShortUserName()) + submitApp(appContext) monitorApplication(appId) @@ -62,14 +84,21 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl def verifyClusterResources(app: GetNewApplicationResponse) = { val maxMem = app.getMaximumResourceCapability().getMemory() - logInfo("Max mem capabililty of resources in this cluster " + maxMem) + logInfo("Max mem capabililty of a single resource in this cluster " + maxMem) - // If the cluster does not have enough memory resources, exit. - val requestedMem = (args.amMemory + YarnAllocationHandler.MEMORY_OVERHEAD) + args.numWorkers * args.workerMemory - if (requestedMem > maxMem) { - logError("Cluster cannot satisfy memory resource request of " + requestedMem) + // if we have requested more then the clusters max for a single resource then exit. + if (args.workerMemory > maxMem) { + logError("the worker size is to large to run on this cluster " + args.workerMemory); + System.exit(1) + } + val amMem = args.amMemory + YarnAllocationHandler.MEMORY_OVERHEAD + if (amMem > maxMem) { + logError("AM size is to large to run on this cluster " + amMem) System.exit(1) } + + // We could add checks to make sure the entire cluster has enough resources but that involves getting + // all the node reports and computing ourselves } def createApplicationSubmissionContext(appId: ApplicationId): ApplicationSubmissionContext = { @@ -86,6 +115,15 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl // Upload Spark and the application JAR to the remote file system // Add them as local resources to the AM val fs = FileSystem.get(conf) + + val delegTokenRenewer = Master.getMasterPrincipal(conf); + if (UserGroupInformation.isSecurityEnabled()) { + if (delegTokenRenewer == null || delegTokenRenewer.length() == 0) { + logError("Can't get Master Kerberos principal for use as renewer") + System.exit(1) + } + } + Map("spark.jar" -> System.getenv("SPARK_JAR"), "app.jar" -> args.userJar, "log4j.properties" -> System.getenv("SPARK_LOG4J_CONF")) .foreach { case(destName, _localPath) => val localPath: String = if (_localPath != null) _localPath.trim() else "" @@ -97,6 +135,11 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl fs.copyFromLocalFile(false, true, src, dst) val destStatus = fs.getFileStatus(dst) + // get tokens for anything we upload to hdfs + if (UserGroupInformation.isSecurityEnabled()) { + fs.addDelegationTokens(delegTokenRenewer, credentials); + } + val amJarRsrc = Records.newRecord(classOf[LocalResource]).asInstanceOf[LocalResource] amJarRsrc.setType(LocalResourceType.FILE) amJarRsrc.setVisibility(LocalResourceVisibility.APPLICATION) @@ -106,6 +149,7 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl locaResources(destName) = amJarRsrc } } + UserGroupInformation.getCurrentUser().addCredentials(credentials); return locaResources } @@ -114,7 +158,6 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl val log4jConfLocalRes = localResources.getOrElse("log4j.properties", null) val env = new HashMap[String, String]() - Apps.addToEnvironment(env, Environment.USER.name, args.amUser) // If log4j present, ensure ours overrides all others if (log4jConfLocalRes != null) Apps.addToEnvironment(env, Environment.CLASSPATH.name, "./") @@ -142,6 +185,7 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl env("SPARK_YARN_LOG4J_SIZE") = log4jConfLocalRes.getSize().toString() } + // Add each SPARK-* key to the environment System.getenv().filterKeys(_.startsWith("SPARK")).foreach { case (k,v) => env(k) = v } return env @@ -195,7 +239,13 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl } // Command for the ApplicationMaster - val commands = List[String]("java " + + var javaCommand = "java"; + val javaHome = System.getenv("JAVA_HOME") + if (javaHome != null && !javaHome.isEmpty()) { + javaCommand = Environment.JAVA_HOME.$() + "/bin/java" + } + + val commands = List[String](javaCommand + " -server " + JAVA_OPTS + " spark.deploy.yarn.ApplicationMaster" + @@ -214,7 +264,12 @@ class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl // Memory for the ApplicationMaster capability.setMemory(args.amMemory + YarnAllocationHandler.MEMORY_OVERHEAD) amContainer.setResource(capability) - + + // Setup security tokens + val dob = new DataOutputBuffer() + credentials.writeTokenStorageToStream(dob) + amContainer.setContainerTokens(ByteBuffer.wrap(dob.getData())) + return amContainer } diff --git a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ClientArguments.scala b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ClientArguments.scala index 2e69fe3fb0..67aff03781 100644 --- a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ClientArguments.scala +++ b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/ClientArguments.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.yarn import spark.util.MemoryParam @@ -13,7 +30,6 @@ class ClientArguments(val args: Array[String]) { var workerMemory = 1024 var workerCores = 1 var numWorkers = 2 - var amUser = System.getProperty("user.name") var amQueue = System.getProperty("QUEUE", "default") var amMemory: Int = 512 // TODO @@ -58,10 +74,6 @@ class ClientArguments(val args: Array[String]) { workerCores = value args = tail - case ("--user") :: value :: tail => - amUser = value - args = tail - case ("--queue") :: value :: tail => amQueue = value args = tail @@ -92,13 +104,11 @@ class ClientArguments(val args: Array[String]) { " --class CLASS_NAME Name of your application's main class (required)\n" + " --args ARGS Arguments to be passed to your application's main class.\n" + " Mutliple invocations are possible, each will be passed in order.\n" + - " Note that first argument will ALWAYS be yarn-standalone : will be added if missing.\n" + " --num-workers NUM Number of workers to start (Default: 2)\n" + " --worker-cores NUM Number of cores for the workers (Default: 1). This is unsused right now.\n" + " --master-memory MEM Memory for Master (e.g. 1000M, 2G) (Default: 512 Mb)\n" + " --worker-memory MEM Memory per Worker (e.g. 1000M, 2G) (Default: 1G)\n" + - " --queue QUEUE The hadoop queue to use for allocation requests (Default: 'default')\n" + - " --user USERNAME Run the ApplicationMaster (and slaves) as a different user\n" + " --queue QUEUE The hadoop queue to use for allocation requests (Default: 'default')" ) System.exit(exitCode) } diff --git a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/WorkerRunnable.scala b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/WorkerRunnable.scala index a2bf0af762..f458f2f6a1 100644 --- a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/WorkerRunnable.scala +++ b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/WorkerRunnable.scala @@ -1,9 +1,29 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.yarn import java.net.URI +import java.nio.ByteBuffer +import java.security.PrivilegedExceptionAction import org.apache.hadoop.conf.Configuration import org.apache.hadoop.fs.{FileStatus, FileSystem, Path} +import org.apache.hadoop.io.DataOutputBuffer import org.apache.hadoop.net.NetUtils import org.apache.hadoop.security.UserGroupInformation import org.apache.hadoop.yarn.api._ @@ -11,7 +31,7 @@ import org.apache.hadoop.yarn.api.records._ import org.apache.hadoop.yarn.api.protocolrecords._ import org.apache.hadoop.yarn.conf.YarnConfiguration import org.apache.hadoop.yarn.ipc.YarnRPC -import org.apache.hadoop.yarn.util.{Apps, ConverterUtils, Records} +import org.apache.hadoop.yarn.util.{Apps, ConverterUtils, Records, ProtoUtils} import org.apache.hadoop.yarn.api.ApplicationConstants.Environment import scala.collection.JavaConversions._ @@ -76,7 +96,19 @@ class WorkerRunnable(container: Container, conf: Configuration, masterAddress: S */ ctx.setUser(UserGroupInformation.getCurrentUser().getShortUserName()) - val commands = List[String]("java " + + + val credentials = UserGroupInformation.getCurrentUser().getCredentials() + val dob = new DataOutputBuffer() + credentials.writeTokenStorageToStream(dob) + ctx.setContainerTokens(ByteBuffer.wrap(dob.getData())) + + var javaCommand = "java"; + val javaHome = System.getenv("JAVA_HOME") + if (javaHome != null && !javaHome.isEmpty()) { + javaCommand = Environment.JAVA_HOME.$() + "/bin/java" + } + + val commands = List[String](javaCommand + " -server " + // Kill if OOM is raised - leverage yarn's failure handling to cause rescheduling. // Not killing the task leaves various aspects of the worker and (to some extent) the jvm in an inconsistent state. @@ -143,8 +175,6 @@ class WorkerRunnable(container: Container, conf: Configuration, masterAddress: S def prepareEnvironment: HashMap[String, String] = { val env = new HashMap[String, String]() - // should we add this ? - Apps.addToEnvironment(env, Environment.USER.name, Utils.getUserNameFromEnvironment()) // If log4j present, ensure ours overrides all others if (System.getenv("SPARK_YARN_LOG4J_PATH") != null) { @@ -165,7 +195,23 @@ class WorkerRunnable(container: Container, conf: Configuration, masterAddress: S val cmHostPortStr = container.getNodeId().getHost() + ":" + container.getNodeId().getPort() val cmAddress = NetUtils.createSocketAddr(cmHostPortStr) logInfo("Connecting to ContainerManager at " + cmHostPortStr) - return rpc.getProxy(classOf[ContainerManager], cmAddress, conf).asInstanceOf[ContainerManager] + + // use doAs and remoteUser here so we can add the container token and not + // pollute the current users credentials with all of the individual container tokens + val user = UserGroupInformation.createRemoteUser(container.getId().toString()); + val containerToken = container.getContainerToken(); + if (containerToken != null) { + user.addToken(ProtoUtils.convertFromProtoFormat(containerToken, cmAddress)) + } + + val proxy = user + .doAs(new PrivilegedExceptionAction[ContainerManager] { + def run: ContainerManager = { + return rpc.getProxy(classOf[ContainerManager], + cmAddress, conf).asInstanceOf[ContainerManager] + } + }); + return proxy; } } diff --git a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/YarnAllocationHandler.scala b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/YarnAllocationHandler.scala index 61dd72a651..b0af8baf08 100644 --- a/core/src/hadoop2-yarn/scala/spark/deploy/yarn/YarnAllocationHandler.scala +++ b/core/src/hadoop2-yarn/scala/spark/deploy/yarn/YarnAllocationHandler.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.yarn import spark.{Logging, Utils} diff --git a/core/src/hadoop2-yarn/scala/spark/scheduler/cluster/YarnClusterScheduler.scala b/core/src/hadoop2-yarn/scala/spark/scheduler/cluster/YarnClusterScheduler.scala index ed732d36bf..307d96111c 100644 --- a/core/src/hadoop2-yarn/scala/spark/scheduler/cluster/YarnClusterScheduler.scala +++ b/core/src/hadoop2-yarn/scala/spark/scheduler/cluster/YarnClusterScheduler.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster import spark._ diff --git a/core/src/hadoop2/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala b/core/src/hadoop2/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala index a0652d7fc7..4b3d84670c 100644 --- a/core/src/hadoop2/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala +++ b/core/src/hadoop2/scala/org/apache/hadoop/mapred/HadoopMapRedUtil.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package org.apache.hadoop.mapred trait HadoopMapRedUtil { diff --git a/core/src/hadoop2/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala b/core/src/hadoop2/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala index 7fdbe322fd..aa3b1ed3a5 100644 --- a/core/src/hadoop2/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala +++ b/core/src/hadoop2/scala/org/apache/hadoop/mapreduce/HadoopMapReduceUtil.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package org.apache.hadoop.mapreduce import org.apache.hadoop.conf.Configuration diff --git a/core/src/hadoop2/scala/spark/deploy/SparkHadoopUtil.scala b/core/src/hadoop2/scala/spark/deploy/SparkHadoopUtil.scala index a0fb4fe25d..617954cb98 100644 --- a/core/src/hadoop2/scala/spark/deploy/SparkHadoopUtil.scala +++ b/core/src/hadoop2/scala/spark/deploy/SparkHadoopUtil.scala @@ -1,5 +1,23 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.mapred.JobConf /** @@ -20,4 +38,10 @@ object SparkHadoopUtil { // Return an appropriate (subclass) of Configuration. Creating config can initializes some hadoop subsystems def newConfiguration(): Configuration = new Configuration() + + // add any user credentials to the job conf which are necessary for running on a secure Hadoop cluster + def addCredentials(conf: JobConf) {} + + def isYarnMode(): Boolean = { false } + } diff --git a/core/src/main/java/spark/network/netty/FileClient.java b/core/src/main/java/spark/network/netty/FileClient.java new file mode 100644 index 0000000000..0625a6d502 --- /dev/null +++ b/core/src/main/java/spark/network/netty/FileClient.java @@ -0,0 +1,89 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty; + +import io.netty.bootstrap.Bootstrap; +import io.netty.channel.Channel; +import io.netty.channel.ChannelFuture; +import io.netty.channel.ChannelFutureListener; +import io.netty.channel.ChannelOption; +import io.netty.channel.oio.OioEventLoopGroup; +import io.netty.channel.socket.oio.OioSocketChannel; + +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +class FileClient { + + private Logger LOG = LoggerFactory.getLogger(this.getClass().getName()); + private FileClientHandler handler = null; + private Channel channel = null; + private Bootstrap bootstrap = null; + private int connectTimeout = 60*1000; // 1 min + + public FileClient(FileClientHandler handler, int connectTimeout) { + this.handler = handler; + this.connectTimeout = connectTimeout; + } + + public void init() { + bootstrap = new Bootstrap(); + bootstrap.group(new OioEventLoopGroup()) + .channel(OioSocketChannel.class) + .option(ChannelOption.SO_KEEPALIVE, true) + .option(ChannelOption.TCP_NODELAY, true) + .option(ChannelOption.CONNECT_TIMEOUT_MILLIS, connectTimeout) + .handler(new FileClientChannelInitializer(handler)); + } + + public void connect(String host, int port) { + try { + // Start the connection attempt. + channel = bootstrap.connect(host, port).sync().channel(); + // ChannelFuture cf = channel.closeFuture(); + //cf.addListener(new ChannelCloseListener(this)); + } catch (InterruptedException e) { + close(); + } + } + + public void waitForClose() { + try { + channel.closeFuture().sync(); + } catch (InterruptedException e) { + LOG.warn("FileClient interrupted", e); + } + } + + public void sendRequest(String file) { + //assert(file == null); + //assert(channel == null); + channel.write(file + "\r\n"); + } + + public void close() { + if(channel != null) { + channel.close(); + channel = null; + } + if ( bootstrap!=null) { + bootstrap.shutdown(); + bootstrap = null; + } + } +} diff --git a/core/src/main/java/spark/network/netty/FileClientChannelInitializer.java b/core/src/main/java/spark/network/netty/FileClientChannelInitializer.java new file mode 100644 index 0000000000..05ad4b61d7 --- /dev/null +++ b/core/src/main/java/spark/network/netty/FileClientChannelInitializer.java @@ -0,0 +1,41 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty; + +import io.netty.buffer.BufType; +import io.netty.channel.ChannelInitializer; +import io.netty.channel.socket.SocketChannel; +import io.netty.handler.codec.string.StringEncoder; + + +class FileClientChannelInitializer extends ChannelInitializer<SocketChannel> { + + private FileClientHandler fhandler; + + public FileClientChannelInitializer(FileClientHandler handler) { + fhandler = handler; + } + + @Override + public void initChannel(SocketChannel channel) { + // file no more than 2G + channel.pipeline() + .addLast("encoder", new StringEncoder(BufType.BYTE)) + .addLast("handler", fhandler); + } +} diff --git a/core/src/main/java/spark/network/netty/FileClientHandler.java b/core/src/main/java/spark/network/netty/FileClientHandler.java new file mode 100644 index 0000000000..e8cd9801f6 --- /dev/null +++ b/core/src/main/java/spark/network/netty/FileClientHandler.java @@ -0,0 +1,60 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty; + +import io.netty.buffer.ByteBuf; +import io.netty.channel.ChannelHandlerContext; +import io.netty.channel.ChannelInboundByteHandlerAdapter; + + +abstract class FileClientHandler extends ChannelInboundByteHandlerAdapter { + + private FileHeader currentHeader = null; + + private volatile boolean handlerCalled = false; + + public boolean isComplete() { + return handlerCalled; + } + + public abstract void handle(ChannelHandlerContext ctx, ByteBuf in, FileHeader header); + public abstract void handleError(String blockId); + + @Override + public ByteBuf newInboundBuffer(ChannelHandlerContext ctx) { + // Use direct buffer if possible. + return ctx.alloc().ioBuffer(); + } + + @Override + public void inboundBufferUpdated(ChannelHandlerContext ctx, ByteBuf in) { + // get header + if (currentHeader == null && in.readableBytes() >= FileHeader.HEADER_SIZE()) { + currentHeader = FileHeader.create(in.readBytes(FileHeader.HEADER_SIZE())); + } + // get file + if(in.readableBytes() >= currentHeader.fileLen()) { + handle(ctx, in, currentHeader); + handlerCalled = true; + currentHeader = null; + ctx.close(); + } + } + +} + diff --git a/core/src/main/java/spark/network/netty/FileServer.java b/core/src/main/java/spark/network/netty/FileServer.java new file mode 100644 index 0000000000..9f009a61d5 --- /dev/null +++ b/core/src/main/java/spark/network/netty/FileServer.java @@ -0,0 +1,103 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty; + +import java.net.InetSocketAddress; + +import io.netty.bootstrap.ServerBootstrap; +import io.netty.channel.Channel; +import io.netty.channel.ChannelFuture; +import io.netty.channel.ChannelOption; +import io.netty.channel.oio.OioEventLoopGroup; +import io.netty.channel.socket.oio.OioServerSocketChannel; + +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + + +/** + * Server that accept the path of a file an echo back its content. + */ +class FileServer { + + private Logger LOG = LoggerFactory.getLogger(this.getClass().getName()); + + private ServerBootstrap bootstrap = null; + private ChannelFuture channelFuture = null; + private int port = 0; + private Thread blockingThread = null; + + public FileServer(PathResolver pResolver, int port) { + InetSocketAddress addr = new InetSocketAddress(port); + + // Configure the server. + bootstrap = new ServerBootstrap(); + bootstrap.group(new OioEventLoopGroup(), new OioEventLoopGroup()) + .channel(OioServerSocketChannel.class) + .option(ChannelOption.SO_BACKLOG, 100) + .option(ChannelOption.SO_RCVBUF, 1500) + .childHandler(new FileServerChannelInitializer(pResolver)); + // Start the server. + channelFuture = bootstrap.bind(addr); + try { + // Get the address we bound to. + InetSocketAddress boundAddress = + ((InetSocketAddress) channelFuture.sync().channel().localAddress()); + this.port = boundAddress.getPort(); + } catch (InterruptedException ie) { + this.port = 0; + } + } + + /** + * Start the file server asynchronously in a new thread. + */ + public void start() { + blockingThread = new Thread() { + public void run() { + try { + channelFuture.channel().closeFuture().sync(); + LOG.info("FileServer exiting"); + } catch (InterruptedException e) { + LOG.error("File server start got interrupted", e); + } + // NOTE: bootstrap is shutdown in stop() + } + }; + blockingThread.setDaemon(true); + blockingThread.start(); + } + + public int getPort() { + return port; + } + + public void stop() { + // Close the bound channel. + if (channelFuture != null) { + channelFuture.channel().close(); + channelFuture = null; + } + // Shutdown bootstrap. + if (bootstrap != null) { + bootstrap.shutdown(); + bootstrap = null; + } + // TODO: Shutdown all accepted channels as well ? + } +} diff --git a/core/src/main/java/spark/network/netty/FileServerChannelInitializer.java b/core/src/main/java/spark/network/netty/FileServerChannelInitializer.java new file mode 100644 index 0000000000..50c57a81a3 --- /dev/null +++ b/core/src/main/java/spark/network/netty/FileServerChannelInitializer.java @@ -0,0 +1,42 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty; + +import io.netty.channel.ChannelInitializer; +import io.netty.channel.socket.SocketChannel; +import io.netty.handler.codec.DelimiterBasedFrameDecoder; +import io.netty.handler.codec.Delimiters; +import io.netty.handler.codec.string.StringDecoder; + + +class FileServerChannelInitializer extends ChannelInitializer<SocketChannel> { + + PathResolver pResolver; + + public FileServerChannelInitializer(PathResolver pResolver) { + this.pResolver = pResolver; + } + + @Override + public void initChannel(SocketChannel channel) { + channel.pipeline() + .addLast("framer", new DelimiterBasedFrameDecoder(8192, Delimiters.lineDelimiter())) + .addLast("strDecoder", new StringDecoder()) + .addLast("handler", new FileServerHandler(pResolver)); + } +} diff --git a/core/src/main/java/spark/network/netty/FileServerHandler.java b/core/src/main/java/spark/network/netty/FileServerHandler.java new file mode 100644 index 0000000000..176ba8da49 --- /dev/null +++ b/core/src/main/java/spark/network/netty/FileServerHandler.java @@ -0,0 +1,82 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty; + +import java.io.File; +import java.io.FileInputStream; + +import io.netty.channel.ChannelHandlerContext; +import io.netty.channel.ChannelInboundMessageHandlerAdapter; +import io.netty.channel.DefaultFileRegion; + + +class FileServerHandler extends ChannelInboundMessageHandlerAdapter<String> { + + PathResolver pResolver; + + public FileServerHandler(PathResolver pResolver){ + this.pResolver = pResolver; + } + + @Override + public void messageReceived(ChannelHandlerContext ctx, String blockId) { + String path = pResolver.getAbsolutePath(blockId); + // if getFilePath returns null, close the channel + if (path == null) { + //ctx.close(); + return; + } + File file = new File(path); + if (file.exists()) { + if (!file.isFile()) { + //logger.info("Not a file : " + file.getAbsolutePath()); + ctx.write(new FileHeader(0, blockId).buffer()); + ctx.flush(); + return; + } + long length = file.length(); + if (length > Integer.MAX_VALUE || length <= 0) { + //logger.info("too large file : " + file.getAbsolutePath() + " of size "+ length); + ctx.write(new FileHeader(0, blockId).buffer()); + ctx.flush(); + return; + } + int len = new Long(length).intValue(); + //logger.info("Sending block "+blockId+" filelen = "+len); + //logger.info("header = "+ (new FileHeader(len, blockId)).buffer()); + ctx.write((new FileHeader(len, blockId)).buffer()); + try { + ctx.sendFile(new DefaultFileRegion(new FileInputStream(file) + .getChannel(), 0, file.length())); + } catch (Exception e) { + //logger.warning("Exception when sending file : " + file.getAbsolutePath()); + e.printStackTrace(); + } + } else { + //logger.warning("File not found: " + file.getAbsolutePath()); + ctx.write(new FileHeader(0, blockId).buffer()); + } + ctx.flush(); + } + + @Override + public void exceptionCaught(ChannelHandlerContext ctx, Throwable cause) { + cause.printStackTrace(); + ctx.close(); + } +} diff --git a/core/src/main/java/spark/network/netty/PathResolver.java b/core/src/main/java/spark/network/netty/PathResolver.java new file mode 100755 index 0000000000..f446c55b19 --- /dev/null +++ b/core/src/main/java/spark/network/netty/PathResolver.java @@ -0,0 +1,29 @@ +/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package spark.network.netty;
+
+
+public interface PathResolver {
+ /**
+ * Get the absolute path of the file
+ *
+ * @param fileId
+ * @return the absolute path of file
+ */
+ public String getAbsolutePath(String fileId);
+}
diff --git a/core/src/main/resources/spark/deploy/static/bootstrap-responsive.min.css b/core/src/main/resources/spark/deploy/static/bootstrap-responsive.min.css deleted file mode 100644 index 1f55036a07..0000000000 --- a/core/src/main/resources/spark/deploy/static/bootstrap-responsive.min.css +++ /dev/null @@ -1,9 +0,0 @@ -/*! - * Bootstrap Responsive v2.0.4 - * - * Copyright 2012 Twitter, Inc - * Licensed under the Apache License v2.0 - * http://www.apache.org/licenses/LICENSE-2.0 - * - * Designed and built with all the love in the world @twitter by @mdo and @fat. - */.clearfix{*zoom:1}.clearfix:before,.clearfix:after{display:table;content:""}.clearfix:after{clear:both}.hide-text{font:0/0 a;color:transparent;text-shadow:none;background-color:transparent;border:0}.input-block-level{display:block;width:100%;min-height:28px;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;-ms-box-sizing:border-box;box-sizing:border-box}.hidden{display:none;visibility:hidden}.visible-phone{display:none!important}.visible-tablet{display:none!important}.hidden-desktop{display:none!important}@media(max-width:767px){.visible-phone{display:inherit!important}.hidden-phone{display:none!important}.hidden-desktop{display:inherit!important}.visible-desktop{display:none!important}}@media(min-width:768px) and (max-width:979px){.visible-tablet{display:inherit!important}.hidden-tablet{display:none!important}.hidden-desktop{display:inherit!important}.visible-desktop{display:none!important}}@media(max-width:480px){.nav-collapse{-webkit-transform:translate3d(0,0,0)}.page-header h1 small{display:block;line-height:18px}input[type="checkbox"],input[type="radio"]{border:1px solid #ccc}.form-horizontal .control-group>label{float:none;width:auto;padding-top:0;text-align:left}.form-horizontal .controls{margin-left:0}.form-horizontal .control-list{padding-top:0}.form-horizontal .form-actions{padding-right:10px;padding-left:10px}.modal{position:absolute;top:10px;right:10px;left:10px;width:auto;margin:0}.modal.fade.in{top:auto}.modal-header .close{padding:10px;margin:-10px}.carousel-caption{position:static}}@media(max-width:767px){body{padding-right:20px;padding-left:20px}.navbar-fixed-top,.navbar-fixed-bottom{margin-right:-20px;margin-left:-20px}.container-fluid{padding:0}.dl-horizontal dt{float:none;width:auto;clear:none;text-align:left}.dl-horizontal dd{margin-left:0}.container{width:auto}.row-fluid{width:100%}.row,.thumbnails{margin-left:0}[class*="span"],.row-fluid [class*="span"]{display:block;float:none;width:auto;margin-left:0}.input-large,.input-xlarge,.input-xxlarge,input[class*="span"],select[class*="span"],textarea[class*="span"],.uneditable-input{display:block;width:100%;min-height:28px;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;-ms-box-sizing:border-box;box-sizing:border-box}.input-prepend input,.input-append input,.input-prepend input[class*="span"],.input-append input[class*="span"]{display:inline-block;width:auto}}@media(min-width:768px) and (max-width:979px){.row{margin-left:-20px;*zoom:1}.row:before,.row:after{display:table;content:""}.row:after{clear:both}[class*="span"]{float:left;margin-left:20px}.container,.navbar-fixed-top .container,.navbar-fixed-bottom .container{width:724px}.span12{width:724px}.span11{width:662px}.span10{width:600px}.span9{width:538px}.span8{width:476px}.span7{width:414px}.span6{width:352px}.span5{width:290px}.span4{width:228px}.span3{width:166px}.span2{width:104px}.span1{width:42px}.offset12{margin-left:764px}.offset11{margin-left:702px}.offset10{margin-left:640px}.offset9{margin-left:578px}.offset8{margin-left:516px}.offset7{margin-left:454px}.offset6{margin-left:392px}.offset5{margin-left:330px}.offset4{margin-left:268px}.offset3{margin-left:206px}.offset2{margin-left:144px}.offset1{margin-left:82px}.row-fluid{width:100%;*zoom:1}.row-fluid:before,.row-fluid:after{display:table;content:""}.row-fluid:after{clear:both}.row-fluid [class*="span"]{display:block;float:left;width:100%;min-height:28px;margin-left:2.762430939%;*margin-left:2.709239449638298%;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;-ms-box-sizing:border-box;box-sizing:border-box}.row-fluid [class*="span"]:first-child{margin-left:0}.row-fluid .span12{width:99.999999993%;*width:99.9468085036383%}.row-fluid .span11{width:91.436464082%;*width:91.38327259263829%}.row-fluid .span10{width:82.87292817100001%;*width:82.8197366816383%}.row-fluid .span9{width:74.30939226%;*width:74.25620077063829%}.row-fluid .span8{width:65.74585634900001%;*width:65.6926648596383%}.row-fluid .span7{width:57.182320438000005%;*width:57.129128948638304%}.row-fluid .span6{width:48.618784527%;*width:48.5655930376383%}.row-fluid .span5{width:40.055248616%;*width:40.0020571266383%}.row-fluid .span4{width:31.491712705%;*width:31.4385212156383%}.row-fluid .span3{width:22.928176794%;*width:22.874985304638297%}.row-fluid .span2{width:14.364640883%;*width:14.311449393638298%}.row-fluid .span1{width:5.801104972%;*width:5.747913482638298%}input,textarea,.uneditable-input{margin-left:0}input.span12,textarea.span12,.uneditable-input.span12{width:714px}input.span11,textarea.span11,.uneditable-input.span11{width:652px}input.span10,textarea.span10,.uneditable-input.span10{width:590px}input.span9,textarea.span9,.uneditable-input.span9{width:528px}input.span8,textarea.span8,.uneditable-input.span8{width:466px}input.span7,textarea.span7,.uneditable-input.span7{width:404px}input.span6,textarea.span6,.uneditable-input.span6{width:342px}input.span5,textarea.span5,.uneditable-input.span5{width:280px}input.span4,textarea.span4,.uneditable-input.span4{width:218px}input.span3,textarea.span3,.uneditable-input.span3{width:156px}input.span2,textarea.span2,.uneditable-input.span2{width:94px}input.span1,textarea.span1,.uneditable-input.span1{width:32px}}@media(min-width:1200px){.row{margin-left:-30px;*zoom:1}.row:before,.row:after{display:table;content:""}.row:after{clear:both}[class*="span"]{float:left;margin-left:30px}.container,.navbar-fixed-top .container,.navbar-fixed-bottom .container{width:1170px}.span12{width:1170px}.span11{width:1070px}.span10{width:970px}.span9{width:870px}.span8{width:770px}.span7{width:670px}.span6{width:570px}.span5{width:470px}.span4{width:370px}.span3{width:270px}.span2{width:170px}.span1{width:70px}.offset12{margin-left:1230px}.offset11{margin-left:1130px}.offset10{margin-left:1030px}.offset9{margin-left:930px}.offset8{margin-left:830px}.offset7{margin-left:730px}.offset6{margin-left:630px}.offset5{margin-left:530px}.offset4{margin-left:430px}.offset3{margin-left:330px}.offset2{margin-left:230px}.offset1{margin-left:130px}.row-fluid{width:100%;*zoom:1}.row-fluid:before,.row-fluid:after{display:table;content:""}.row-fluid:after{clear:both}.row-fluid [class*="span"]{display:block;float:left;width:100%;min-height:28px;margin-left:2.564102564%;*margin-left:2.510911074638298%;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;-ms-box-sizing:border-box;box-sizing:border-box}.row-fluid [class*="span"]:first-child{margin-left:0}.row-fluid .span12{width:100%;*width:99.94680851063829%}.row-fluid .span11{width:91.45299145300001%;*width:91.3997999636383%}.row-fluid .span10{width:82.905982906%;*width:82.8527914166383%}.row-fluid .span9{width:74.358974359%;*width:74.30578286963829%}.row-fluid .span8{width:65.81196581200001%;*width:65.7587743226383%}.row-fluid .span7{width:57.264957265%;*width:57.2117657756383%}.row-fluid .span6{width:48.717948718%;*width:48.6647572286383%}.row-fluid .span5{width:40.170940171000005%;*width:40.117748681638304%}.row-fluid .span4{width:31.623931624%;*width:31.5707401346383%}.row-fluid .span3{width:23.076923077%;*width:23.0237315876383%}.row-fluid .span2{width:14.529914530000001%;*width:14.4767230406383%}.row-fluid .span1{width:5.982905983%;*width:5.929714493638298%}input,textarea,.uneditable-input{margin-left:0}input.span12,textarea.span12,.uneditable-input.span12{width:1160px}input.span11,textarea.span11,.uneditable-input.span11{width:1060px}input.span10,textarea.span10,.uneditable-input.span10{width:960px}input.span9,textarea.span9,.uneditable-input.span9{width:860px}input.span8,textarea.span8,.uneditable-input.span8{width:760px}input.span7,textarea.span7,.uneditable-input.span7{width:660px}input.span6,textarea.span6,.uneditable-input.span6{width:560px}input.span5,textarea.span5,.uneditable-input.span5{width:460px}input.span4,textarea.span4,.uneditable-input.span4{width:360px}input.span3,textarea.span3,.uneditable-input.span3{width:260px}input.span2,textarea.span2,.uneditable-input.span2{width:160px}input.span1,textarea.span1,.uneditable-input.span1{width:60px}.thumbnails{margin-left:-30px}.thumbnails>li{margin-left:30px}.row-fluid .thumbnails{margin-left:0}}@media(max-width:979px){body{padding-top:0}.navbar-fixed-top,.navbar-fixed-bottom{position:static}.navbar-fixed-top{margin-bottom:18px}.navbar-fixed-bottom{margin-top:18px}.navbar-fixed-top .navbar-inner,.navbar-fixed-bottom .navbar-inner{padding:5px}.navbar .container{width:auto;padding:0}.navbar .brand{padding-right:10px;padding-left:10px;margin:0 0 0 -5px}.nav-collapse{clear:both}.nav-collapse .nav{float:none;margin:0 0 9px}.nav-collapse .nav>li{float:none}.nav-collapse .nav>li>a{margin-bottom:2px}.nav-collapse .nav>.divider-vertical{display:none}.nav-collapse .nav .nav-header{color:#999;text-shadow:none}.nav-collapse .nav>li>a,.nav-collapse .dropdown-menu a{padding:6px 15px;font-weight:bold;color:#999;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}.nav-collapse .btn{padding:4px 10px 4px;font-weight:normal;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.nav-collapse .dropdown-menu li+li a{margin-bottom:2px}.nav-collapse .nav>li>a:hover,.nav-collapse .dropdown-menu a:hover{background-color:#222}.nav-collapse.in .btn-group{padding:0;margin-top:5px}.nav-collapse .dropdown-menu{position:static;top:auto;left:auto;display:block;float:none;max-width:none;padding:0;margin:0 15px;background-color:transparent;border:0;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0;-webkit-box-shadow:none;-moz-box-shadow:none;box-shadow:none}.nav-collapse .dropdown-menu:before,.nav-collapse .dropdown-menu:after{display:none}.nav-collapse .dropdown-menu .divider{display:none}.nav-collapse .navbar-form,.nav-collapse .navbar-search{float:none;padding:9px 15px;margin:9px 0;border-top:1px solid #222;border-bottom:1px solid #222;-webkit-box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.1);-moz-box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.1);box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.1)}.navbar .nav-collapse .nav.pull-right{float:none;margin-left:0}.nav-collapse,.nav-collapse.collapse{height:0;overflow:hidden}.navbar .btn-navbar{display:block}.navbar-static .navbar-inner{padding-right:10px;padding-left:10px}}@media(min-width:980px){.nav-collapse.collapse{height:auto!important;overflow:visible!important}} diff --git a/core/src/main/resources/spark/deploy/static/bootstrap.min.css b/core/src/main/resources/spark/deploy/static/bootstrap.min.css deleted file mode 100644 index b74b4546a6..0000000000 --- a/core/src/main/resources/spark/deploy/static/bootstrap.min.css +++ /dev/null @@ -1,9 +0,0 @@ -/*! - * Bootstrap v2.0.4 - * - * Copyright 2012 Twitter, Inc - * Licensed under the Apache License v2.0 - * http://www.apache.org/licenses/LICENSE-2.0 - * - * Designed and built with all the love in the world @twitter by @mdo and @fat. - */article,aside,details,figcaption,figure,footer,header,hgroup,nav,section{display:block}audio,canvas,video{display:inline-block;*display:inline;*zoom:1}audio:not([controls]){display:none}html{font-size:100%;-webkit-text-size-adjust:100%;-ms-text-size-adjust:100%}a:focus{outline:thin dotted #333;outline:5px auto -webkit-focus-ring-color;outline-offset:-2px}a:hover,a:active{outline:0}sub,sup{position:relative;font-size:75%;line-height:0;vertical-align:baseline}sup{top:-0.5em}sub{bottom:-0.25em}img{max-width:100%;vertical-align:middle;border:0;-ms-interpolation-mode:bicubic}#map_canvas img{max-width:none}button,input,select,textarea{margin:0;font-size:100%;vertical-align:middle}button,input{*overflow:visible;line-height:normal}button::-moz-focus-inner,input::-moz-focus-inner{padding:0;border:0}button,input[type="button"],input[type="reset"],input[type="submit"]{cursor:pointer;-webkit-appearance:button}input[type="search"]{-webkit-box-sizing:content-box;-moz-box-sizing:content-box;box-sizing:content-box;-webkit-appearance:textfield}input[type="search"]::-webkit-search-decoration,input[type="search"]::-webkit-search-cancel-button{-webkit-appearance:none}textarea{overflow:auto;vertical-align:top}.clearfix{*zoom:1}.clearfix:before,.clearfix:after{display:table;content:""}.clearfix:after{clear:both}.hide-text{font:0/0 a;color:transparent;text-shadow:none;background-color:transparent;border:0}.input-block-level{display:block;width:100%;min-height:28px;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;-ms-box-sizing:border-box;box-sizing:border-box}body{margin:0;font-family:"Helvetica Neue",Helvetica,Arial,sans-serif;font-size:13px;line-height:18px;color:#333;background-color:#fff}a{color:#08c;text-decoration:none}a:hover{color:#005580;text-decoration:underline}.row{margin-left:-20px;*zoom:1}.row:before,.row:after{display:table;content:""}.row:after{clear:both}[class*="span"]{float:left;margin-left:20px}.container,.navbar-fixed-top .container,.navbar-fixed-bottom .container{width:940px}.span12{width:940px}.span11{width:860px}.span10{width:780px}.span9{width:700px}.span8{width:620px}.span7{width:540px}.span6{width:460px}.span5{width:380px}.span4{width:300px}.span3{width:220px}.span2{width:140px}.span1{width:60px}.offset12{margin-left:980px}.offset11{margin-left:900px}.offset10{margin-left:820px}.offset9{margin-left:740px}.offset8{margin-left:660px}.offset7{margin-left:580px}.offset6{margin-left:500px}.offset5{margin-left:420px}.offset4{margin-left:340px}.offset3{margin-left:260px}.offset2{margin-left:180px}.offset1{margin-left:100px}.row-fluid{width:100%;*zoom:1}.row-fluid:before,.row-fluid:after{display:table;content:""}.row-fluid:after{clear:both}.row-fluid [class*="span"]{display:block;float:left;width:100%;min-height:28px;margin-left:2.127659574%;*margin-left:2.0744680846382977%;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;-ms-box-sizing:border-box;box-sizing:border-box}.row-fluid [class*="span"]:first-child{margin-left:0}.row-fluid .span12{width:99.99999998999999%;*width:99.94680850063828%}.row-fluid .span11{width:91.489361693%;*width:91.4361702036383%}.row-fluid .span10{width:82.97872339599999%;*width:82.92553190663828%}.row-fluid .span9{width:74.468085099%;*width:74.4148936096383%}.row-fluid .span8{width:65.95744680199999%;*width:65.90425531263828%}.row-fluid .span7{width:57.446808505%;*width:57.3936170156383%}.row-fluid .span6{width:48.93617020799999%;*width:48.88297871863829%}.row-fluid .span5{width:40.425531911%;*width:40.3723404216383%}.row-fluid .span4{width:31.914893614%;*width:31.8617021246383%}.row-fluid .span3{width:23.404255317%;*width:23.3510638276383%}.row-fluid .span2{width:14.89361702%;*width:14.8404255306383%}.row-fluid .span1{width:6.382978723%;*width:6.329787233638298%}.container{margin-right:auto;margin-left:auto;*zoom:1}.container:before,.container:after{display:table;content:""}.container:after{clear:both}.container-fluid{padding-right:20px;padding-left:20px;*zoom:1}.container-fluid:before,.container-fluid:after{display:table;content:""}.container-fluid:after{clear:both}p{margin:0 0 9px}p small{font-size:11px;color:#999}.lead{margin-bottom:18px;font-size:20px;font-weight:200;line-height:27px}h1,h2,h3,h4,h5,h6{margin:0;font-family:inherit;font-weight:bold;color:inherit;text-rendering:optimizelegibility}h1 small,h2 small,h3 small,h4 small,h5 small,h6 small{font-weight:normal;color:#999}h1{font-size:30px;line-height:36px}h1 small{font-size:18px}h2{font-size:24px;line-height:36px}h2 small{font-size:18px}h3{font-size:18px;line-height:27px}h3 small{font-size:14px}h4,h5,h6{line-height:18px}h4{font-size:14px}h4 small{font-size:12px}h5{font-size:12px}h6{font-size:11px;color:#999;text-transform:uppercase}.page-header{padding-bottom:17px;margin:18px 0;border-bottom:1px solid #eee}.page-header h1{line-height:1}ul,ol{padding:0;margin:0 0 9px 25px}ul ul,ul ol,ol ol,ol ul{margin-bottom:0}ul{list-style:disc}ol{list-style:decimal}li{line-height:18px}ul.unstyled,ol.unstyled{margin-left:0;list-style:none}dl{margin-bottom:18px}dt,dd{line-height:18px}dt{font-weight:bold;line-height:17px}dd{margin-left:9px}.dl-horizontal dt{float:left;width:120px;overflow:hidden;clear:left;text-align:right;text-overflow:ellipsis;white-space:nowrap}.dl-horizontal dd{margin-left:130px}hr{margin:18px 0;border:0;border-top:1px solid #eee;border-bottom:1px solid #fff}strong{font-weight:bold}em{font-style:italic}.muted{color:#999}abbr[title]{cursor:help;border-bottom:1px dotted #999}abbr.initialism{font-size:90%;text-transform:uppercase}blockquote{padding:0 0 0 15px;margin:0 0 18px;border-left:5px solid #eee}blockquote p{margin-bottom:0;font-size:16px;font-weight:300;line-height:22.5px}blockquote small{display:block;line-height:18px;color:#999}blockquote small:before{content:'\2014 \00A0'}blockquote.pull-right{float:right;padding-right:15px;padding-left:0;border-right:5px solid #eee;border-left:0}blockquote.pull-right p,blockquote.pull-right small{text-align:right}q:before,q:after,blockquote:before,blockquote:after{content:""}address{display:block;margin-bottom:18px;font-style:normal;line-height:18px}small{font-size:100%}cite{font-style:normal}code,pre{padding:0 3px 2px;font-family:Menlo,Monaco,Consolas,"Courier New",monospace;font-size:12px;color:#333;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}code{padding:2px 4px;color:#d14;background-color:#f7f7f9;border:1px solid #e1e1e8}pre{display:block;padding:8.5px;margin:0 0 9px;font-size:12.025px;line-height:18px;word-break:break-all;word-wrap:break-word;white-space:pre;white-space:pre-wrap;background-color:#f5f5f5;border:1px solid #ccc;border:1px solid rgba(0,0,0,0.15);-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}pre.prettyprint{margin-bottom:18px}pre code{padding:0;color:inherit;background-color:transparent;border:0}.pre-scrollable{max-height:340px;overflow-y:scroll}form{margin:0 0 18px}fieldset{padding:0;margin:0;border:0}legend{display:block;width:100%;padding:0;margin-bottom:27px;font-size:19.5px;line-height:36px;color:#333;border:0;border-bottom:1px solid #e5e5e5}legend small{font-size:13.5px;color:#999}label,input,button,select,textarea{font-size:13px;font-weight:normal;line-height:18px}input,button,select,textarea{font-family:"Helvetica Neue",Helvetica,Arial,sans-serif}label{display:block;margin-bottom:5px}select,textarea,input[type="text"],input[type="password"],input[type="datetime"],input[type="datetime-local"],input[type="date"],input[type="month"],input[type="time"],input[type="week"],input[type="number"],input[type="email"],input[type="url"],input[type="search"],input[type="tel"],input[type="color"],.uneditable-input{display:inline-block;height:18px;padding:4px;margin-bottom:9px;font-size:13px;line-height:18px;color:#555}input,textarea{width:210px}textarea{height:auto}textarea,input[type="text"],input[type="password"],input[type="datetime"],input[type="datetime-local"],input[type="date"],input[type="month"],input[type="time"],input[type="week"],input[type="number"],input[type="email"],input[type="url"],input[type="search"],input[type="tel"],input[type="color"],.uneditable-input{background-color:#fff;border:1px solid #ccc;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);-webkit-transition:border linear .2s,box-shadow linear .2s;-moz-transition:border linear .2s,box-shadow linear .2s;-ms-transition:border linear .2s,box-shadow linear .2s;-o-transition:border linear .2s,box-shadow linear .2s;transition:border linear .2s,box-shadow linear .2s}textarea:focus,input[type="text"]:focus,input[type="password"]:focus,input[type="datetime"]:focus,input[type="datetime-local"]:focus,input[type="date"]:focus,input[type="month"]:focus,input[type="time"]:focus,input[type="week"]:focus,input[type="number"]:focus,input[type="email"]:focus,input[type="url"]:focus,input[type="search"]:focus,input[type="tel"]:focus,input[type="color"]:focus,.uneditable-input:focus{border-color:rgba(82,168,236,0.8);outline:0;outline:thin dotted \9;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 8px rgba(82,168,236,0.6);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 8px rgba(82,168,236,0.6);box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 8px rgba(82,168,236,0.6)}input[type="radio"],input[type="checkbox"]{margin:3px 0;*margin-top:0;line-height:normal;cursor:pointer}input[type="submit"],input[type="reset"],input[type="button"],input[type="radio"],input[type="checkbox"]{width:auto}.uneditable-textarea{width:auto;height:auto}select,input[type="file"]{height:28px;*margin-top:4px;line-height:28px}select{width:220px;border:1px solid #bbb}select[multiple],select[size]{height:auto}select:focus,input[type="file"]:focus,input[type="radio"]:focus,input[type="checkbox"]:focus{outline:thin dotted #333;outline:5px auto -webkit-focus-ring-color;outline-offset:-2px}.radio,.checkbox{min-height:18px;padding-left:18px}.radio input[type="radio"],.checkbox input[type="checkbox"]{float:left;margin-left:-18px}.controls>.radio:first-child,.controls>.checkbox:first-child{padding-top:5px}.radio.inline,.checkbox.inline{display:inline-block;padding-top:5px;margin-bottom:0;vertical-align:middle}.radio.inline+.radio.inline,.checkbox.inline+.checkbox.inline{margin-left:10px}.input-mini{width:60px}.input-small{width:90px}.input-medium{width:150px}.input-large{width:210px}.input-xlarge{width:270px}.input-xxlarge{width:530px}input[class*="span"],select[class*="span"],textarea[class*="span"],.uneditable-input[class*="span"],.row-fluid input[class*="span"],.row-fluid select[class*="span"],.row-fluid textarea[class*="span"],.row-fluid .uneditable-input[class*="span"]{float:none;margin-left:0}.input-append input[class*="span"],.input-append .uneditable-input[class*="span"],.input-prepend input[class*="span"],.input-prepend .uneditable-input[class*="span"],.row-fluid .input-prepend [class*="span"],.row-fluid .input-append [class*="span"]{display:inline-block}input,textarea,.uneditable-input{margin-left:0}input.span12,textarea.span12,.uneditable-input.span12{width:930px}input.span11,textarea.span11,.uneditable-input.span11{width:850px}input.span10,textarea.span10,.uneditable-input.span10{width:770px}input.span9,textarea.span9,.uneditable-input.span9{width:690px}input.span8,textarea.span8,.uneditable-input.span8{width:610px}input.span7,textarea.span7,.uneditable-input.span7{width:530px}input.span6,textarea.span6,.uneditable-input.span6{width:450px}input.span5,textarea.span5,.uneditable-input.span5{width:370px}input.span4,textarea.span4,.uneditable-input.span4{width:290px}input.span3,textarea.span3,.uneditable-input.span3{width:210px}input.span2,textarea.span2,.uneditable-input.span2{width:130px}input.span1,textarea.span1,.uneditable-input.span1{width:50px}input[disabled],select[disabled],textarea[disabled],input[readonly],select[readonly],textarea[readonly]{cursor:not-allowed;background-color:#eee;border-color:#ddd}input[type="radio"][disabled],input[type="checkbox"][disabled],input[type="radio"][readonly],input[type="checkbox"][readonly]{background-color:transparent}.control-group.warning>label,.control-group.warning .help-block,.control-group.warning .help-inline{color:#c09853}.control-group.warning .checkbox,.control-group.warning .radio,.control-group.warning input,.control-group.warning select,.control-group.warning textarea{color:#c09853;border-color:#c09853}.control-group.warning .checkbox:focus,.control-group.warning .radio:focus,.control-group.warning input:focus,.control-group.warning select:focus,.control-group.warning textarea:focus{border-color:#a47e3c;-webkit-box-shadow:0 0 6px #dbc59e;-moz-box-shadow:0 0 6px #dbc59e;box-shadow:0 0 6px #dbc59e}.control-group.warning .input-prepend .add-on,.control-group.warning .input-append .add-on{color:#c09853;background-color:#fcf8e3;border-color:#c09853}.control-group.error>label,.control-group.error .help-block,.control-group.error .help-inline{color:#b94a48}.control-group.error .checkbox,.control-group.error .radio,.control-group.error input,.control-group.error select,.control-group.error textarea{color:#b94a48;border-color:#b94a48}.control-group.error .checkbox:focus,.control-group.error .radio:focus,.control-group.error input:focus,.control-group.error select:focus,.control-group.error textarea:focus{border-color:#953b39;-webkit-box-shadow:0 0 6px #d59392;-moz-box-shadow:0 0 6px #d59392;box-shadow:0 0 6px #d59392}.control-group.error .input-prepend .add-on,.control-group.error .input-append .add-on{color:#b94a48;background-color:#f2dede;border-color:#b94a48}.control-group.success>label,.control-group.success .help-block,.control-group.success .help-inline{color:#468847}.control-group.success .checkbox,.control-group.success .radio,.control-group.success input,.control-group.success select,.control-group.success textarea{color:#468847;border-color:#468847}.control-group.success .checkbox:focus,.control-group.success .radio:focus,.control-group.success input:focus,.control-group.success select:focus,.control-group.success textarea:focus{border-color:#356635;-webkit-box-shadow:0 0 6px #7aba7b;-moz-box-shadow:0 0 6px #7aba7b;box-shadow:0 0 6px #7aba7b}.control-group.success .input-prepend .add-on,.control-group.success .input-append .add-on{color:#468847;background-color:#dff0d8;border-color:#468847}input:focus:required:invalid,textarea:focus:required:invalid,select:focus:required:invalid{color:#b94a48;border-color:#ee5f5b}input:focus:required:invalid:focus,textarea:focus:required:invalid:focus,select:focus:required:invalid:focus{border-color:#e9322d;-webkit-box-shadow:0 0 6px #f8b9b7;-moz-box-shadow:0 0 6px #f8b9b7;box-shadow:0 0 6px #f8b9b7}.form-actions{padding:17px 20px 18px;margin-top:18px;margin-bottom:18px;background-color:#f5f5f5;border-top:1px solid #e5e5e5;*zoom:1}.form-actions:before,.form-actions:after{display:table;content:""}.form-actions:after{clear:both}.uneditable-input{overflow:hidden;white-space:nowrap;cursor:not-allowed;background-color:#fff;border-color:#eee;-webkit-box-shadow:inset 0 1px 2px rgba(0,0,0,0.025);-moz-box-shadow:inset 0 1px 2px rgba(0,0,0,0.025);box-shadow:inset 0 1px 2px rgba(0,0,0,0.025)}:-moz-placeholder{color:#999}:-ms-input-placeholder{color:#999}::-webkit-input-placeholder{color:#999}.help-block,.help-inline{color:#555}.help-block{display:block;margin-bottom:9px}.help-inline{display:inline-block;*display:inline;padding-left:5px;vertical-align:middle;*zoom:1}.input-prepend,.input-append{margin-bottom:5px}.input-prepend input,.input-append input,.input-prepend select,.input-append select,.input-prepend .uneditable-input,.input-append .uneditable-input{position:relative;margin-bottom:0;*margin-left:0;vertical-align:middle;-webkit-border-radius:0 3px 3px 0;-moz-border-radius:0 3px 3px 0;border-radius:0 3px 3px 0}.input-prepend input:focus,.input-append input:focus,.input-prepend select:focus,.input-append select:focus,.input-prepend .uneditable-input:focus,.input-append .uneditable-input:focus{z-index:2}.input-prepend .uneditable-input,.input-append .uneditable-input{border-left-color:#ccc}.input-prepend .add-on,.input-append .add-on{display:inline-block;width:auto;height:18px;min-width:16px;padding:4px 5px;font-weight:normal;line-height:18px;text-align:center;text-shadow:0 1px 0 #fff;vertical-align:middle;background-color:#eee;border:1px solid #ccc}.input-prepend .add-on,.input-append .add-on,.input-prepend .btn,.input-append .btn{margin-left:-1px;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.input-prepend .active,.input-append .active{background-color:#a9dba9;border-color:#46a546}.input-prepend .add-on,.input-prepend .btn{margin-right:-1px}.input-prepend .add-on:first-child,.input-prepend .btn:first-child{-webkit-border-radius:3px 0 0 3px;-moz-border-radius:3px 0 0 3px;border-radius:3px 0 0 3px}.input-append input,.input-append select,.input-append .uneditable-input{-webkit-border-radius:3px 0 0 3px;-moz-border-radius:3px 0 0 3px;border-radius:3px 0 0 3px}.input-append .uneditable-input{border-right-color:#ccc;border-left-color:#eee}.input-append .add-on:last-child,.input-append .btn:last-child{-webkit-border-radius:0 3px 3px 0;-moz-border-radius:0 3px 3px 0;border-radius:0 3px 3px 0}.input-prepend.input-append input,.input-prepend.input-append select,.input-prepend.input-append .uneditable-input{-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.input-prepend.input-append .add-on:first-child,.input-prepend.input-append .btn:first-child{margin-right:-1px;-webkit-border-radius:3px 0 0 3px;-moz-border-radius:3px 0 0 3px;border-radius:3px 0 0 3px}.input-prepend.input-append .add-on:last-child,.input-prepend.input-append .btn:last-child{margin-left:-1px;-webkit-border-radius:0 3px 3px 0;-moz-border-radius:0 3px 3px 0;border-radius:0 3px 3px 0}.search-query{padding-right:14px;padding-right:4px \9;padding-left:14px;padding-left:4px \9;margin-bottom:0;-webkit-border-radius:14px;-moz-border-radius:14px;border-radius:14px}.form-search input,.form-inline input,.form-horizontal input,.form-search textarea,.form-inline textarea,.form-horizontal textarea,.form-search select,.form-inline select,.form-horizontal select,.form-search .help-inline,.form-inline .help-inline,.form-horizontal .help-inline,.form-search .uneditable-input,.form-inline .uneditable-input,.form-horizontal .uneditable-input,.form-search .input-prepend,.form-inline .input-prepend,.form-horizontal .input-prepend,.form-search .input-append,.form-inline .input-append,.form-horizontal .input-append{display:inline-block;*display:inline;margin-bottom:0;*zoom:1}.form-search .hide,.form-inline .hide,.form-horizontal .hide{display:none}.form-search label,.form-inline label{display:inline-block}.form-search .input-append,.form-inline .input-append,.form-search .input-prepend,.form-inline .input-prepend{margin-bottom:0}.form-search .radio,.form-search .checkbox,.form-inline .radio,.form-inline .checkbox{padding-left:0;margin-bottom:0;vertical-align:middle}.form-search .radio input[type="radio"],.form-search .checkbox input[type="checkbox"],.form-inline .radio input[type="radio"],.form-inline .checkbox input[type="checkbox"]{float:left;margin-right:3px;margin-left:0}.control-group{margin-bottom:9px}legend+.control-group{margin-top:18px;-webkit-margin-top-collapse:separate}.form-horizontal .control-group{margin-bottom:18px;*zoom:1}.form-horizontal .control-group:before,.form-horizontal .control-group:after{display:table;content:""}.form-horizontal .control-group:after{clear:both}.form-horizontal .control-label{float:left;width:140px;padding-top:5px;text-align:right}.form-horizontal .controls{*display:inline-block;*padding-left:20px;margin-left:160px;*margin-left:0}.form-horizontal .controls:first-child{*padding-left:160px}.form-horizontal .help-block{margin-top:9px;margin-bottom:0}.form-horizontal .form-actions{padding-left:160px}table{max-width:100%;background-color:transparent;border-collapse:collapse;border-spacing:0}.table{width:100%;margin-bottom:18px}.table th,.table td{padding:8px;line-height:18px;text-align:left;vertical-align:top;border-top:1px solid #ddd}.table th{font-weight:bold}.table thead th{vertical-align:bottom}.table caption+thead tr:first-child th,.table caption+thead tr:first-child td,.table colgroup+thead tr:first-child th,.table colgroup+thead tr:first-child td,.table thead:first-child tr:first-child th,.table thead:first-child tr:first-child td{border-top:0}.table tbody+tbody{border-top:2px solid #ddd}.table-condensed th,.table-condensed td{padding:4px 5px}.table-bordered{border:1px solid #ddd;border-collapse:separate;*border-collapse:collapsed;border-left:0;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.table-bordered th,.table-bordered td{border-left:1px solid #ddd}.table-bordered caption+thead tr:first-child th,.table-bordered caption+tbody tr:first-child th,.table-bordered caption+tbody tr:first-child td,.table-bordered colgroup+thead tr:first-child th,.table-bordered colgroup+tbody tr:first-child th,.table-bordered colgroup+tbody tr:first-child td,.table-bordered thead:first-child tr:first-child th,.table-bordered tbody:first-child tr:first-child th,.table-bordered tbody:first-child tr:first-child td{border-top:0}.table-bordered thead:first-child tr:first-child th:first-child,.table-bordered tbody:first-child tr:first-child td:first-child{-webkit-border-top-left-radius:4px;border-top-left-radius:4px;-moz-border-radius-topleft:4px}.table-bordered thead:first-child tr:first-child th:last-child,.table-bordered tbody:first-child tr:first-child td:last-child{-webkit-border-top-right-radius:4px;border-top-right-radius:4px;-moz-border-radius-topright:4px}.table-bordered thead:last-child tr:last-child th:first-child,.table-bordered tbody:last-child tr:last-child td:first-child{-webkit-border-radius:0 0 0 4px;-moz-border-radius:0 0 0 4px;border-radius:0 0 0 4px;-webkit-border-bottom-left-radius:4px;border-bottom-left-radius:4px;-moz-border-radius-bottomleft:4px}.table-bordered thead:last-child tr:last-child th:last-child,.table-bordered tbody:last-child tr:last-child td:last-child{-webkit-border-bottom-right-radius:4px;border-bottom-right-radius:4px;-moz-border-radius-bottomright:4px}.table-striped tbody tr:nth-child(odd) td,.table-striped tbody tr:nth-child(odd) th{background-color:#f9f9f9}.table tbody tr:hover td,.table tbody tr:hover th{background-color:#f5f5f5}table .span1{float:none;width:44px;margin-left:0}table .span2{float:none;width:124px;margin-left:0}table .span3{float:none;width:204px;margin-left:0}table .span4{float:none;width:284px;margin-left:0}table .span5{float:none;width:364px;margin-left:0}table .span6{float:none;width:444px;margin-left:0}table .span7{float:none;width:524px;margin-left:0}table .span8{float:none;width:604px;margin-left:0}table .span9{float:none;width:684px;margin-left:0}table .span10{float:none;width:764px;margin-left:0}table .span11{float:none;width:844px;margin-left:0}table .span12{float:none;width:924px;margin-left:0}table .span13{float:none;width:1004px;margin-left:0}table .span14{float:none;width:1084px;margin-left:0}table .span15{float:none;width:1164px;margin-left:0}table .span16{float:none;width:1244px;margin-left:0}table .span17{float:none;width:1324px;margin-left:0}table .span18{float:none;width:1404px;margin-left:0}table .span19{float:none;width:1484px;margin-left:0}table .span20{float:none;width:1564px;margin-left:0}table .span21{float:none;width:1644px;margin-left:0}table .span22{float:none;width:1724px;margin-left:0}table .span23{float:none;width:1804px;margin-left:0}table .span24{float:none;width:1884px;margin-left:0}[class^="icon-"],[class*=" icon-"]{display:inline-block;width:14px;height:14px;*margin-right:.3em;line-height:14px;vertical-align:text-top;background-image:url("../img/glyphicons-halflings.png");background-position:14px 14px;background-repeat:no-repeat}[class^="icon-"]:last-child,[class*=" icon-"]:last-child{*margin-left:0}.icon-white{background-image:url("../img/glyphicons-halflings-white.png")}.icon-glass{background-position:0 0}.icon-music{background-position:-24px 0}.icon-search{background-position:-48px 0}.icon-envelope{background-position:-72px 0}.icon-heart{background-position:-96px 0}.icon-star{background-position:-120px 0}.icon-star-empty{background-position:-144px 0}.icon-user{background-position:-168px 0}.icon-film{background-position:-192px 0}.icon-th-large{background-position:-216px 0}.icon-th{background-position:-240px 0}.icon-th-list{background-position:-264px 0}.icon-ok{background-position:-288px 0}.icon-remove{background-position:-312px 0}.icon-zoom-in{background-position:-336px 0}.icon-zoom-out{background-position:-360px 0}.icon-off{background-position:-384px 0}.icon-signal{background-position:-408px 0}.icon-cog{background-position:-432px 0}.icon-trash{background-position:-456px 0}.icon-home{background-position:0 -24px}.icon-file{background-position:-24px -24px}.icon-time{background-position:-48px -24px}.icon-road{background-position:-72px -24px}.icon-download-alt{background-position:-96px -24px}.icon-download{background-position:-120px -24px}.icon-upload{background-position:-144px -24px}.icon-inbox{background-position:-168px -24px}.icon-play-circle{background-position:-192px -24px}.icon-repeat{background-position:-216px -24px}.icon-refresh{background-position:-240px -24px}.icon-list-alt{background-position:-264px -24px}.icon-lock{background-position:-287px -24px}.icon-flag{background-position:-312px -24px}.icon-headphones{background-position:-336px -24px}.icon-volume-off{background-position:-360px -24px}.icon-volume-down{background-position:-384px -24px}.icon-volume-up{background-position:-408px -24px}.icon-qrcode{background-position:-432px -24px}.icon-barcode{background-position:-456px -24px}.icon-tag{background-position:0 -48px}.icon-tags{background-position:-25px -48px}.icon-book{background-position:-48px -48px}.icon-bookmark{background-position:-72px -48px}.icon-print{background-position:-96px -48px}.icon-camera{background-position:-120px -48px}.icon-font{background-position:-144px -48px}.icon-bold{background-position:-167px -48px}.icon-italic{background-position:-192px -48px}.icon-text-height{background-position:-216px -48px}.icon-text-width{background-position:-240px -48px}.icon-align-left{background-position:-264px -48px}.icon-align-center{background-position:-288px -48px}.icon-align-right{background-position:-312px -48px}.icon-align-justify{background-position:-336px -48px}.icon-list{background-position:-360px -48px}.icon-indent-left{background-position:-384px -48px}.icon-indent-right{background-position:-408px -48px}.icon-facetime-video{background-position:-432px -48px}.icon-picture{background-position:-456px -48px}.icon-pencil{background-position:0 -72px}.icon-map-marker{background-position:-24px -72px}.icon-adjust{background-position:-48px -72px}.icon-tint{background-position:-72px -72px}.icon-edit{background-position:-96px -72px}.icon-share{background-position:-120px -72px}.icon-check{background-position:-144px -72px}.icon-move{background-position:-168px -72px}.icon-step-backward{background-position:-192px -72px}.icon-fast-backward{background-position:-216px -72px}.icon-backward{background-position:-240px -72px}.icon-play{background-position:-264px -72px}.icon-pause{background-position:-288px -72px}.icon-stop{background-position:-312px -72px}.icon-forward{background-position:-336px -72px}.icon-fast-forward{background-position:-360px -72px}.icon-step-forward{background-position:-384px -72px}.icon-eject{background-position:-408px -72px}.icon-chevron-left{background-position:-432px -72px}.icon-chevron-right{background-position:-456px -72px}.icon-plus-sign{background-position:0 -96px}.icon-minus-sign{background-position:-24px -96px}.icon-remove-sign{background-position:-48px -96px}.icon-ok-sign{background-position:-72px -96px}.icon-question-sign{background-position:-96px -96px}.icon-info-sign{background-position:-120px -96px}.icon-screenshot{background-position:-144px -96px}.icon-remove-circle{background-position:-168px -96px}.icon-ok-circle{background-position:-192px -96px}.icon-ban-circle{background-position:-216px -96px}.icon-arrow-left{background-position:-240px -96px}.icon-arrow-right{background-position:-264px -96px}.icon-arrow-up{background-position:-289px -96px}.icon-arrow-down{background-position:-312px -96px}.icon-share-alt{background-position:-336px -96px}.icon-resize-full{background-position:-360px -96px}.icon-resize-small{background-position:-384px -96px}.icon-plus{background-position:-408px -96px}.icon-minus{background-position:-433px -96px}.icon-asterisk{background-position:-456px -96px}.icon-exclamation-sign{background-position:0 -120px}.icon-gift{background-position:-24px -120px}.icon-leaf{background-position:-48px -120px}.icon-fire{background-position:-72px -120px}.icon-eye-open{background-position:-96px -120px}.icon-eye-close{background-position:-120px -120px}.icon-warning-sign{background-position:-144px -120px}.icon-plane{background-position:-168px -120px}.icon-calendar{background-position:-192px -120px}.icon-random{background-position:-216px -120px}.icon-comment{background-position:-240px -120px}.icon-magnet{background-position:-264px -120px}.icon-chevron-up{background-position:-288px -120px}.icon-chevron-down{background-position:-313px -119px}.icon-retweet{background-position:-336px -120px}.icon-shopping-cart{background-position:-360px -120px}.icon-folder-close{background-position:-384px -120px}.icon-folder-open{background-position:-408px -120px}.icon-resize-vertical{background-position:-432px -119px}.icon-resize-horizontal{background-position:-456px -118px}.icon-hdd{background-position:0 -144px}.icon-bullhorn{background-position:-24px -144px}.icon-bell{background-position:-48px -144px}.icon-certificate{background-position:-72px -144px}.icon-thumbs-up{background-position:-96px -144px}.icon-thumbs-down{background-position:-120px -144px}.icon-hand-right{background-position:-144px -144px}.icon-hand-left{background-position:-168px -144px}.icon-hand-up{background-position:-192px -144px}.icon-hand-down{background-position:-216px -144px}.icon-circle-arrow-right{background-position:-240px -144px}.icon-circle-arrow-left{background-position:-264px -144px}.icon-circle-arrow-up{background-position:-288px -144px}.icon-circle-arrow-down{background-position:-312px -144px}.icon-globe{background-position:-336px -144px}.icon-wrench{background-position:-360px -144px}.icon-tasks{background-position:-384px -144px}.icon-filter{background-position:-408px -144px}.icon-briefcase{background-position:-432px -144px}.icon-fullscreen{background-position:-456px -144px}.dropup,.dropdown{position:relative}.dropdown-toggle{*margin-bottom:-3px}.dropdown-toggle:active,.open .dropdown-toggle{outline:0}.caret{display:inline-block;width:0;height:0;vertical-align:top;border-top:4px solid #000;border-right:4px solid transparent;border-left:4px solid transparent;content:"";opacity:.3;filter:alpha(opacity=30)}.dropdown .caret{margin-top:8px;margin-left:2px}.dropdown:hover .caret,.open .caret{opacity:1;filter:alpha(opacity=100)}.dropdown-menu{position:absolute;top:100%;left:0;z-index:1000;display:none;float:left;min-width:160px;padding:4px 0;margin:1px 0 0;list-style:none;background-color:#fff;border:1px solid #ccc;border:1px solid rgba(0,0,0,0.2);*border-right-width:2px;*border-bottom-width:2px;-webkit-border-radius:5px;-moz-border-radius:5px;border-radius:5px;-webkit-box-shadow:0 5px 10px rgba(0,0,0,0.2);-moz-box-shadow:0 5px 10px rgba(0,0,0,0.2);box-shadow:0 5px 10px rgba(0,0,0,0.2);-webkit-background-clip:padding-box;-moz-background-clip:padding;background-clip:padding-box}.dropdown-menu.pull-right{right:0;left:auto}.dropdown-menu .divider{*width:100%;height:1px;margin:8px 1px;*margin:-5px 0 5px;overflow:hidden;background-color:#e5e5e5;border-bottom:1px solid #fff}.dropdown-menu a{display:block;padding:3px 15px;clear:both;font-weight:normal;line-height:18px;color:#333;white-space:nowrap}.dropdown-menu li>a:hover,.dropdown-menu .active>a,.dropdown-menu .active>a:hover{color:#fff;text-decoration:none;background-color:#08c}.open{*z-index:1000}.open>.dropdown-menu{display:block}.pull-right>.dropdown-menu{right:0;left:auto}.dropup .caret,.navbar-fixed-bottom .dropdown .caret{border-top:0;border-bottom:4px solid #000;content:"\2191"}.dropup .dropdown-menu,.navbar-fixed-bottom .dropdown .dropdown-menu{top:auto;bottom:100%;margin-bottom:1px}.typeahead{margin-top:2px;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.well{min-height:20px;padding:19px;margin-bottom:20px;background-color:#f5f5f5;border:1px solid #eee;border:1px solid rgba(0,0,0,0.05);-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.05);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.05);box-shadow:inset 0 1px 1px rgba(0,0,0,0.05)}.well blockquote{border-color:#ddd;border-color:rgba(0,0,0,0.15)}.well-large{padding:24px;-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px}.well-small{padding:9px;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}.fade{opacity:0;-webkit-transition:opacity .15s linear;-moz-transition:opacity .15s linear;-ms-transition:opacity .15s linear;-o-transition:opacity .15s linear;transition:opacity .15s linear}.fade.in{opacity:1}.collapse{position:relative;height:0;overflow:hidden;-webkit-transition:height .35s ease;-moz-transition:height .35s ease;-ms-transition:height .35s ease;-o-transition:height .35s ease;transition:height .35s ease}.collapse.in{height:auto}.close{float:right;font-size:20px;font-weight:bold;line-height:18px;color:#000;text-shadow:0 1px 0 #fff;opacity:.2;filter:alpha(opacity=20)}.close:hover{color:#000;text-decoration:none;cursor:pointer;opacity:.4;filter:alpha(opacity=40)}button.close{padding:0;cursor:pointer;background:transparent;border:0;-webkit-appearance:none}.btn{display:inline-block;*display:inline;padding:4px 10px 4px;margin-bottom:0;*margin-left:.3em;font-size:13px;line-height:18px;*line-height:20px;color:#333;text-align:center;text-shadow:0 1px 1px rgba(255,255,255,0.75);vertical-align:middle;cursor:pointer;background-color:#f5f5f5;*background-color:#e6e6e6;background-image:-ms-linear-gradient(top,#fff,#e6e6e6);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fff),to(#e6e6e6));background-image:-webkit-linear-gradient(top,#fff,#e6e6e6);background-image:-o-linear-gradient(top,#fff,#e6e6e6);background-image:linear-gradient(top,#fff,#e6e6e6);background-image:-moz-linear-gradient(top,#fff,#e6e6e6);background-repeat:repeat-x;border:1px solid #ccc;*border:0;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);border-color:#e6e6e6 #e6e6e6 #bfbfbf;border-bottom-color:#b3b3b3;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#ffffff',endColorstr='#e6e6e6',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false);*zoom:1;-webkit-box-shadow:inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05);box-shadow:inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05)}.btn:hover,.btn:active,.btn.active,.btn.disabled,.btn[disabled]{background-color:#e6e6e6;*background-color:#d9d9d9}.btn:active,.btn.active{background-color:#ccc \9}.btn:first-child{*margin-left:0}.btn:hover{color:#333;text-decoration:none;background-color:#e6e6e6;*background-color:#d9d9d9;background-position:0 -15px;-webkit-transition:background-position .1s linear;-moz-transition:background-position .1s linear;-ms-transition:background-position .1s linear;-o-transition:background-position .1s linear;transition:background-position .1s linear}.btn:focus{outline:thin dotted #333;outline:5px auto -webkit-focus-ring-color;outline-offset:-2px}.btn.active,.btn:active{background-color:#e6e6e6;background-color:#d9d9d9 \9;background-image:none;outline:0;-webkit-box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05);box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05)}.btn.disabled,.btn[disabled]{cursor:default;background-color:#e6e6e6;background-image:none;opacity:.65;filter:alpha(opacity=65);-webkit-box-shadow:none;-moz-box-shadow:none;box-shadow:none}.btn-large{padding:9px 14px;font-size:15px;line-height:normal;-webkit-border-radius:5px;-moz-border-radius:5px;border-radius:5px}.btn-large [class^="icon-"]{margin-top:1px}.btn-small{padding:5px 9px;font-size:11px;line-height:16px}.btn-small [class^="icon-"]{margin-top:-1px}.btn-mini{padding:2px 6px;font-size:11px;line-height:14px}.btn-primary,.btn-primary:hover,.btn-warning,.btn-warning:hover,.btn-danger,.btn-danger:hover,.btn-success,.btn-success:hover,.btn-info,.btn-info:hover,.btn-inverse,.btn-inverse:hover{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25)}.btn-primary.active,.btn-warning.active,.btn-danger.active,.btn-success.active,.btn-info.active,.btn-inverse.active{color:rgba(255,255,255,0.75)}.btn{border-color:#ccc;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25)}.btn-primary{background-color:#0074cc;*background-color:#05c;background-image:-ms-linear-gradient(top,#08c,#05c);background-image:-webkit-gradient(linear,0 0,0 100%,from(#08c),to(#05c));background-image:-webkit-linear-gradient(top,#08c,#05c);background-image:-o-linear-gradient(top,#08c,#05c);background-image:-moz-linear-gradient(top,#08c,#05c);background-image:linear-gradient(top,#08c,#05c);background-repeat:repeat-x;border-color:#05c #05c #003580;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:dximagetransform.microsoft.gradient(startColorstr='#0088cc',endColorstr='#0055cc',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-primary:hover,.btn-primary:active,.btn-primary.active,.btn-primary.disabled,.btn-primary[disabled]{background-color:#05c;*background-color:#004ab3}.btn-primary:active,.btn-primary.active{background-color:#004099 \9}.btn-warning{background-color:#faa732;*background-color:#f89406;background-image:-ms-linear-gradient(top,#fbb450,#f89406);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fbb450),to(#f89406));background-image:-webkit-linear-gradient(top,#fbb450,#f89406);background-image:-o-linear-gradient(top,#fbb450,#f89406);background-image:-moz-linear-gradient(top,#fbb450,#f89406);background-image:linear-gradient(top,#fbb450,#f89406);background-repeat:repeat-x;border-color:#f89406 #f89406 #ad6704;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:dximagetransform.microsoft.gradient(startColorstr='#fbb450',endColorstr='#f89406',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-warning:hover,.btn-warning:active,.btn-warning.active,.btn-warning.disabled,.btn-warning[disabled]{background-color:#f89406;*background-color:#df8505}.btn-warning:active,.btn-warning.active{background-color:#c67605 \9}.btn-danger{background-color:#da4f49;*background-color:#bd362f;background-image:-ms-linear-gradient(top,#ee5f5b,#bd362f);background-image:-webkit-gradient(linear,0 0,0 100%,from(#ee5f5b),to(#bd362f));background-image:-webkit-linear-gradient(top,#ee5f5b,#bd362f);background-image:-o-linear-gradient(top,#ee5f5b,#bd362f);background-image:-moz-linear-gradient(top,#ee5f5b,#bd362f);background-image:linear-gradient(top,#ee5f5b,#bd362f);background-repeat:repeat-x;border-color:#bd362f #bd362f #802420;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:dximagetransform.microsoft.gradient(startColorstr='#ee5f5b',endColorstr='#bd362f',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-danger:hover,.btn-danger:active,.btn-danger.active,.btn-danger.disabled,.btn-danger[disabled]{background-color:#bd362f;*background-color:#a9302a}.btn-danger:active,.btn-danger.active{background-color:#942a25 \9}.btn-success{background-color:#5bb75b;*background-color:#51a351;background-image:-ms-linear-gradient(top,#62c462,#51a351);background-image:-webkit-gradient(linear,0 0,0 100%,from(#62c462),to(#51a351));background-image:-webkit-linear-gradient(top,#62c462,#51a351);background-image:-o-linear-gradient(top,#62c462,#51a351);background-image:-moz-linear-gradient(top,#62c462,#51a351);background-image:linear-gradient(top,#62c462,#51a351);background-repeat:repeat-x;border-color:#51a351 #51a351 #387038;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:dximagetransform.microsoft.gradient(startColorstr='#62c462',endColorstr='#51a351',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-success:hover,.btn-success:active,.btn-success.active,.btn-success.disabled,.btn-success[disabled]{background-color:#51a351;*background-color:#499249}.btn-success:active,.btn-success.active{background-color:#408140 \9}.btn-info{background-color:#49afcd;*background-color:#2f96b4;background-image:-ms-linear-gradient(top,#5bc0de,#2f96b4);background-image:-webkit-gradient(linear,0 0,0 100%,from(#5bc0de),to(#2f96b4));background-image:-webkit-linear-gradient(top,#5bc0de,#2f96b4);background-image:-o-linear-gradient(top,#5bc0de,#2f96b4);background-image:-moz-linear-gradient(top,#5bc0de,#2f96b4);background-image:linear-gradient(top,#5bc0de,#2f96b4);background-repeat:repeat-x;border-color:#2f96b4 #2f96b4 #1f6377;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:dximagetransform.microsoft.gradient(startColorstr='#5bc0de',endColorstr='#2f96b4',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-info:hover,.btn-info:active,.btn-info.active,.btn-info.disabled,.btn-info[disabled]{background-color:#2f96b4;*background-color:#2a85a0}.btn-info:active,.btn-info.active{background-color:#24748c \9}.btn-inverse{background-color:#414141;*background-color:#222;background-image:-ms-linear-gradient(top,#555,#222);background-image:-webkit-gradient(linear,0 0,0 100%,from(#555),to(#222));background-image:-webkit-linear-gradient(top,#555,#222);background-image:-o-linear-gradient(top,#555,#222);background-image:-moz-linear-gradient(top,#555,#222);background-image:linear-gradient(top,#555,#222);background-repeat:repeat-x;border-color:#222 #222 #000;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:dximagetransform.microsoft.gradient(startColorstr='#555555',endColorstr='#222222',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-inverse:hover,.btn-inverse:active,.btn-inverse.active,.btn-inverse.disabled,.btn-inverse[disabled]{background-color:#222;*background-color:#151515}.btn-inverse:active,.btn-inverse.active{background-color:#080808 \9}button.btn,input[type="submit"].btn{*padding-top:2px;*padding-bottom:2px}button.btn::-moz-focus-inner,input[type="submit"].btn::-moz-focus-inner{padding:0;border:0}button.btn.btn-large,input[type="submit"].btn.btn-large{*padding-top:7px;*padding-bottom:7px}button.btn.btn-small,input[type="submit"].btn.btn-small{*padding-top:3px;*padding-bottom:3px}button.btn.btn-mini,input[type="submit"].btn.btn-mini{*padding-top:1px;*padding-bottom:1px}.btn-group{position:relative;*margin-left:.3em;*zoom:1}.btn-group:before,.btn-group:after{display:table;content:""}.btn-group:after{clear:both}.btn-group:first-child{*margin-left:0}.btn-group+.btn-group{margin-left:5px}.btn-toolbar{margin-top:9px;margin-bottom:9px}.btn-toolbar .btn-group{display:inline-block;*display:inline;*zoom:1}.btn-group>.btn{position:relative;float:left;margin-left:-1px;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.btn-group>.btn:first-child{margin-left:0;-webkit-border-bottom-left-radius:4px;border-bottom-left-radius:4px;-webkit-border-top-left-radius:4px;border-top-left-radius:4px;-moz-border-radius-bottomleft:4px;-moz-border-radius-topleft:4px}.btn-group>.btn:last-child,.btn-group>.dropdown-toggle{-webkit-border-top-right-radius:4px;border-top-right-radius:4px;-webkit-border-bottom-right-radius:4px;border-bottom-right-radius:4px;-moz-border-radius-topright:4px;-moz-border-radius-bottomright:4px}.btn-group>.btn.large:first-child{margin-left:0;-webkit-border-bottom-left-radius:6px;border-bottom-left-radius:6px;-webkit-border-top-left-radius:6px;border-top-left-radius:6px;-moz-border-radius-bottomleft:6px;-moz-border-radius-topleft:6px}.btn-group>.btn.large:last-child,.btn-group>.large.dropdown-toggle{-webkit-border-top-right-radius:6px;border-top-right-radius:6px;-webkit-border-bottom-right-radius:6px;border-bottom-right-radius:6px;-moz-border-radius-topright:6px;-moz-border-radius-bottomright:6px}.btn-group>.btn:hover,.btn-group>.btn:focus,.btn-group>.btn:active,.btn-group>.btn.active{z-index:2}.btn-group .dropdown-toggle:active,.btn-group.open .dropdown-toggle{outline:0}.btn-group>.dropdown-toggle{*padding-top:4px;padding-right:8px;*padding-bottom:4px;padding-left:8px;-webkit-box-shadow:inset 1px 0 0 rgba(255,255,255,0.125),inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:inset 1px 0 0 rgba(255,255,255,0.125),inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05);box-shadow:inset 1px 0 0 rgba(255,255,255,0.125),inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05)}.btn-group>.btn-mini.dropdown-toggle{padding-right:5px;padding-left:5px}.btn-group>.btn-small.dropdown-toggle{*padding-top:4px;*padding-bottom:4px}.btn-group>.btn-large.dropdown-toggle{padding-right:12px;padding-left:12px}.btn-group.open .dropdown-toggle{background-image:none;-webkit-box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05);box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05)}.btn-group.open .btn.dropdown-toggle{background-color:#e6e6e6}.btn-group.open .btn-primary.dropdown-toggle{background-color:#05c}.btn-group.open .btn-warning.dropdown-toggle{background-color:#f89406}.btn-group.open .btn-danger.dropdown-toggle{background-color:#bd362f}.btn-group.open .btn-success.dropdown-toggle{background-color:#51a351}.btn-group.open .btn-info.dropdown-toggle{background-color:#2f96b4}.btn-group.open .btn-inverse.dropdown-toggle{background-color:#222}.btn .caret{margin-top:7px;margin-left:0}.btn:hover .caret,.open.btn-group .caret{opacity:1;filter:alpha(opacity=100)}.btn-mini .caret{margin-top:5px}.btn-small .caret{margin-top:6px}.btn-large .caret{margin-top:6px;border-top-width:5px;border-right-width:5px;border-left-width:5px}.dropup .btn-large .caret{border-top:0;border-bottom:5px solid #000}.btn-primary .caret,.btn-warning .caret,.btn-danger .caret,.btn-info .caret,.btn-success .caret,.btn-inverse .caret{border-top-color:#fff;border-bottom-color:#fff;opacity:.75;filter:alpha(opacity=75)}.alert{padding:8px 35px 8px 14px;margin-bottom:18px;color:#c09853;text-shadow:0 1px 0 rgba(255,255,255,0.5);background-color:#fcf8e3;border:1px solid #fbeed5;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.alert-heading{color:inherit}.alert .close{position:relative;top:-2px;right:-21px;line-height:18px}.alert-success{color:#468847;background-color:#dff0d8;border-color:#d6e9c6}.alert-danger,.alert-error{color:#b94a48;background-color:#f2dede;border-color:#eed3d7}.alert-info{color:#3a87ad;background-color:#d9edf7;border-color:#bce8f1}.alert-block{padding-top:14px;padding-bottom:14px}.alert-block>p,.alert-block>ul{margin-bottom:0}.alert-block p+p{margin-top:5px}.nav{margin-bottom:18px;margin-left:0;list-style:none}.nav>li>a{display:block}.nav>li>a:hover{text-decoration:none;background-color:#eee}.nav>.pull-right{float:right}.nav .nav-header{display:block;padding:3px 15px;font-size:11px;font-weight:bold;line-height:18px;color:#999;text-shadow:0 1px 0 rgba(255,255,255,0.5);text-transform:uppercase}.nav li+.nav-header{margin-top:9px}.nav-list{padding-right:15px;padding-left:15px;margin-bottom:0}.nav-list>li>a,.nav-list .nav-header{margin-right:-15px;margin-left:-15px;text-shadow:0 1px 0 rgba(255,255,255,0.5)}.nav-list>li>a{padding:3px 15px}.nav-list>.active>a,.nav-list>.active>a:hover{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.2);background-color:#08c}.nav-list [class^="icon-"]{margin-right:2px}.nav-list .divider{*width:100%;height:1px;margin:8px 1px;*margin:-5px 0 5px;overflow:hidden;background-color:#e5e5e5;border-bottom:1px solid #fff}.nav-tabs,.nav-pills{*zoom:1}.nav-tabs:before,.nav-pills:before,.nav-tabs:after,.nav-pills:after{display:table;content:""}.nav-tabs:after,.nav-pills:after{clear:both}.nav-tabs>li,.nav-pills>li{float:left}.nav-tabs>li>a,.nav-pills>li>a{padding-right:12px;padding-left:12px;margin-right:2px;line-height:14px}.nav-tabs{border-bottom:1px solid #ddd}.nav-tabs>li{margin-bottom:-1px}.nav-tabs>li>a{padding-top:8px;padding-bottom:8px;line-height:18px;border:1px solid transparent;-webkit-border-radius:4px 4px 0 0;-moz-border-radius:4px 4px 0 0;border-radius:4px 4px 0 0}.nav-tabs>li>a:hover{border-color:#eee #eee #ddd}.nav-tabs>.active>a,.nav-tabs>.active>a:hover{color:#555;cursor:default;background-color:#fff;border:1px solid #ddd;border-bottom-color:transparent}.nav-pills>li>a{padding-top:8px;padding-bottom:8px;margin-top:2px;margin-bottom:2px;-webkit-border-radius:5px;-moz-border-radius:5px;border-radius:5px}.nav-pills>.active>a,.nav-pills>.active>a:hover{color:#fff;background-color:#08c}.nav-stacked>li{float:none}.nav-stacked>li>a{margin-right:0}.nav-tabs.nav-stacked{border-bottom:0}.nav-tabs.nav-stacked>li>a{border:1px solid #ddd;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.nav-tabs.nav-stacked>li:first-child>a{-webkit-border-radius:4px 4px 0 0;-moz-border-radius:4px 4px 0 0;border-radius:4px 4px 0 0}.nav-tabs.nav-stacked>li:last-child>a{-webkit-border-radius:0 0 4px 4px;-moz-border-radius:0 0 4px 4px;border-radius:0 0 4px 4px}.nav-tabs.nav-stacked>li>a:hover{z-index:2;border-color:#ddd}.nav-pills.nav-stacked>li>a{margin-bottom:3px}.nav-pills.nav-stacked>li:last-child>a{margin-bottom:1px}.nav-tabs .dropdown-menu{-webkit-border-radius:0 0 5px 5px;-moz-border-radius:0 0 5px 5px;border-radius:0 0 5px 5px}.nav-pills .dropdown-menu{-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.nav-tabs .dropdown-toggle .caret,.nav-pills .dropdown-toggle .caret{margin-top:6px;border-top-color:#08c;border-bottom-color:#08c}.nav-tabs .dropdown-toggle:hover .caret,.nav-pills .dropdown-toggle:hover .caret{border-top-color:#005580;border-bottom-color:#005580}.nav-tabs .active .dropdown-toggle .caret,.nav-pills .active .dropdown-toggle .caret{border-top-color:#333;border-bottom-color:#333}.nav>.dropdown.active>a:hover{color:#000;cursor:pointer}.nav-tabs .open .dropdown-toggle,.nav-pills .open .dropdown-toggle,.nav>li.dropdown.open.active>a:hover{color:#fff;background-color:#999;border-color:#999}.nav li.dropdown.open .caret,.nav li.dropdown.open.active .caret,.nav li.dropdown.open a:hover .caret{border-top-color:#fff;border-bottom-color:#fff;opacity:1;filter:alpha(opacity=100)}.tabs-stacked .open>a:hover{border-color:#999}.tabbable{*zoom:1}.tabbable:before,.tabbable:after{display:table;content:""}.tabbable:after{clear:both}.tab-content{overflow:auto}.tabs-below>.nav-tabs,.tabs-right>.nav-tabs,.tabs-left>.nav-tabs{border-bottom:0}.tab-content>.tab-pane,.pill-content>.pill-pane{display:none}.tab-content>.active,.pill-content>.active{display:block}.tabs-below>.nav-tabs{border-top:1px solid #ddd}.tabs-below>.nav-tabs>li{margin-top:-1px;margin-bottom:0}.tabs-below>.nav-tabs>li>a{-webkit-border-radius:0 0 4px 4px;-moz-border-radius:0 0 4px 4px;border-radius:0 0 4px 4px}.tabs-below>.nav-tabs>li>a:hover{border-top-color:#ddd;border-bottom-color:transparent}.tabs-below>.nav-tabs>.active>a,.tabs-below>.nav-tabs>.active>a:hover{border-color:transparent #ddd #ddd #ddd}.tabs-left>.nav-tabs>li,.tabs-right>.nav-tabs>li{float:none}.tabs-left>.nav-tabs>li>a,.tabs-right>.nav-tabs>li>a{min-width:74px;margin-right:0;margin-bottom:3px}.tabs-left>.nav-tabs{float:left;margin-right:19px;border-right:1px solid #ddd}.tabs-left>.nav-tabs>li>a{margin-right:-1px;-webkit-border-radius:4px 0 0 4px;-moz-border-radius:4px 0 0 4px;border-radius:4px 0 0 4px}.tabs-left>.nav-tabs>li>a:hover{border-color:#eee #ddd #eee #eee}.tabs-left>.nav-tabs .active>a,.tabs-left>.nav-tabs .active>a:hover{border-color:#ddd transparent #ddd #ddd;*border-right-color:#fff}.tabs-right>.nav-tabs{float:right;margin-left:19px;border-left:1px solid #ddd}.tabs-right>.nav-tabs>li>a{margin-left:-1px;-webkit-border-radius:0 4px 4px 0;-moz-border-radius:0 4px 4px 0;border-radius:0 4px 4px 0}.tabs-right>.nav-tabs>li>a:hover{border-color:#eee #eee #eee #ddd}.tabs-right>.nav-tabs .active>a,.tabs-right>.nav-tabs .active>a:hover{border-color:#ddd #ddd #ddd transparent;*border-left-color:#fff}.navbar{*position:relative;*z-index:2;margin-bottom:18px;overflow:visible}.navbar-inner{min-height:40px;padding-right:20px;padding-left:20px;background-color:#2c2c2c;background-image:-moz-linear-gradient(top,#333,#222);background-image:-ms-linear-gradient(top,#333,#222);background-image:-webkit-gradient(linear,0 0,0 100%,from(#333),to(#222));background-image:-webkit-linear-gradient(top,#333,#222);background-image:-o-linear-gradient(top,#333,#222);background-image:linear-gradient(top,#333,#222);background-repeat:repeat-x;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#333333',endColorstr='#222222',GradientType=0);-webkit-box-shadow:0 1px 3px rgba(0,0,0,0.25),inset 0 -1px 0 rgba(0,0,0,0.1);-moz-box-shadow:0 1px 3px rgba(0,0,0,0.25),inset 0 -1px 0 rgba(0,0,0,0.1);box-shadow:0 1px 3px rgba(0,0,0,0.25),inset 0 -1px 0 rgba(0,0,0,0.1)}.navbar .container{width:auto}.nav-collapse.collapse{height:auto}.navbar{color:#999}.navbar .brand:hover{text-decoration:none}.navbar .brand{display:block;float:left;padding:8px 20px 12px;margin-left:-20px;font-size:20px;font-weight:200;line-height:1;color:#999}.navbar .navbar-text{margin-bottom:0;line-height:40px}.navbar .navbar-link{color:#999}.navbar .navbar-link:hover{color:#fff}.navbar .btn,.navbar .btn-group{margin-top:5px}.navbar .btn-group .btn{margin:0}.navbar-form{margin-bottom:0;*zoom:1}.navbar-form:before,.navbar-form:after{display:table;content:""}.navbar-form:after{clear:both}.navbar-form input,.navbar-form select,.navbar-form .radio,.navbar-form .checkbox{margin-top:5px}.navbar-form input,.navbar-form select{display:inline-block;margin-bottom:0}.navbar-form input[type="image"],.navbar-form input[type="checkbox"],.navbar-form input[type="radio"]{margin-top:3px}.navbar-form .input-append,.navbar-form .input-prepend{margin-top:6px;white-space:nowrap}.navbar-form .input-append input,.navbar-form .input-prepend input{margin-top:0}.navbar-search{position:relative;float:left;margin-top:6px;margin-bottom:0}.navbar-search .search-query{padding:4px 9px;font-family:"Helvetica Neue",Helvetica,Arial,sans-serif;font-size:13px;font-weight:normal;line-height:1;color:#fff;background-color:#626262;border:1px solid #151515;-webkit-box-shadow:inset 0 1px 2px rgba(0,0,0,0.1),0 1px 0 rgba(255,255,255,0.15);-moz-box-shadow:inset 0 1px 2px rgba(0,0,0,0.1),0 1px 0 rgba(255,255,255,0.15);box-shadow:inset 0 1px 2px rgba(0,0,0,0.1),0 1px 0 rgba(255,255,255,0.15);-webkit-transition:none;-moz-transition:none;-ms-transition:none;-o-transition:none;transition:none}.navbar-search .search-query:-moz-placeholder{color:#ccc}.navbar-search .search-query:-ms-input-placeholder{color:#ccc}.navbar-search .search-query::-webkit-input-placeholder{color:#ccc}.navbar-search .search-query:focus,.navbar-search .search-query.focused{padding:5px 10px;color:#333;text-shadow:0 1px 0 #fff;background-color:#fff;border:0;outline:0;-webkit-box-shadow:0 0 3px rgba(0,0,0,0.15);-moz-box-shadow:0 0 3px rgba(0,0,0,0.15);box-shadow:0 0 3px rgba(0,0,0,0.15)}.navbar-fixed-top,.navbar-fixed-bottom{position:fixed;right:0;left:0;z-index:1030;margin-bottom:0}.navbar-fixed-top .navbar-inner,.navbar-fixed-bottom .navbar-inner{padding-right:0;padding-left:0;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.navbar-fixed-top .container,.navbar-fixed-bottom .container{width:940px}.navbar-fixed-top{top:0}.navbar-fixed-bottom{bottom:0}.navbar .nav{position:relative;left:0;display:block;float:left;margin:0 10px 0 0}.navbar .nav.pull-right{float:right}.navbar .nav>li{display:block;float:left}.navbar .nav>li>a{float:none;padding:9px 10px 11px;line-height:19px;color:#999;text-decoration:none;text-shadow:0 -1px 0 rgba(0,0,0,0.25)}.navbar .btn{display:inline-block;padding:4px 10px 4px;margin:5px 5px 6px;line-height:18px}.navbar .btn-group{padding:5px 5px 6px;margin:0}.navbar .nav>li>a:hover{color:#fff;text-decoration:none;background-color:transparent}.navbar .nav .active>a,.navbar .nav .active>a:hover{color:#fff;text-decoration:none;background-color:#222}.navbar .divider-vertical{width:1px;height:40px;margin:0 9px;overflow:hidden;background-color:#222;border-right:1px solid #333}.navbar .nav.pull-right{margin-right:0;margin-left:10px}.navbar .btn-navbar{display:none;float:right;padding:7px 10px;margin-right:5px;margin-left:5px;background-color:#2c2c2c;*background-color:#222;background-image:-ms-linear-gradient(top,#333,#222);background-image:-webkit-gradient(linear,0 0,0 100%,from(#333),to(#222));background-image:-webkit-linear-gradient(top,#333,#222);background-image:-o-linear-gradient(top,#333,#222);background-image:linear-gradient(top,#333,#222);background-image:-moz-linear-gradient(top,#333,#222);background-repeat:repeat-x;border-color:#222 #222 #000;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:dximagetransform.microsoft.gradient(startColorstr='#333333',endColorstr='#222222',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false);-webkit-box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.075);-moz-box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.075);box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.075)}.navbar .btn-navbar:hover,.navbar .btn-navbar:active,.navbar .btn-navbar.active,.navbar .btn-navbar.disabled,.navbar .btn-navbar[disabled]{background-color:#222;*background-color:#151515}.navbar .btn-navbar:active,.navbar .btn-navbar.active{background-color:#080808 \9}.navbar .btn-navbar .icon-bar{display:block;width:18px;height:2px;background-color:#f5f5f5;-webkit-border-radius:1px;-moz-border-radius:1px;border-radius:1px;-webkit-box-shadow:0 1px 0 rgba(0,0,0,0.25);-moz-box-shadow:0 1px 0 rgba(0,0,0,0.25);box-shadow:0 1px 0 rgba(0,0,0,0.25)}.btn-navbar .icon-bar+.icon-bar{margin-top:3px}.navbar .dropdown-menu:before{position:absolute;top:-7px;left:9px;display:inline-block;border-right:7px solid transparent;border-bottom:7px solid #ccc;border-left:7px solid transparent;border-bottom-color:rgba(0,0,0,0.2);content:''}.navbar .dropdown-menu:after{position:absolute;top:-6px;left:10px;display:inline-block;border-right:6px solid transparent;border-bottom:6px solid #fff;border-left:6px solid transparent;content:''}.navbar-fixed-bottom .dropdown-menu:before{top:auto;bottom:-7px;border-top:7px solid #ccc;border-bottom:0;border-top-color:rgba(0,0,0,0.2)}.navbar-fixed-bottom .dropdown-menu:after{top:auto;bottom:-6px;border-top:6px solid #fff;border-bottom:0}.navbar .nav li.dropdown .dropdown-toggle .caret,.navbar .nav li.dropdown.open .caret{border-top-color:#fff;border-bottom-color:#fff}.navbar .nav li.dropdown.active .caret{opacity:1;filter:alpha(opacity=100)}.navbar .nav li.dropdown.open>.dropdown-toggle,.navbar .nav li.dropdown.active>.dropdown-toggle,.navbar .nav li.dropdown.open.active>.dropdown-toggle{background-color:transparent}.navbar .nav li.dropdown.active>.dropdown-toggle:hover{color:#fff}.navbar .pull-right .dropdown-menu,.navbar .dropdown-menu.pull-right{right:0;left:auto}.navbar .pull-right .dropdown-menu:before,.navbar .dropdown-menu.pull-right:before{right:12px;left:auto}.navbar .pull-right .dropdown-menu:after,.navbar .dropdown-menu.pull-right:after{right:13px;left:auto}.breadcrumb{padding:7px 14px;margin:0 0 18px;list-style:none;background-color:#fbfbfb;background-image:-moz-linear-gradient(top,#fff,#f5f5f5);background-image:-ms-linear-gradient(top,#fff,#f5f5f5);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fff),to(#f5f5f5));background-image:-webkit-linear-gradient(top,#fff,#f5f5f5);background-image:-o-linear-gradient(top,#fff,#f5f5f5);background-image:linear-gradient(top,#fff,#f5f5f5);background-repeat:repeat-x;border:1px solid #ddd;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#ffffff',endColorstr='#f5f5f5',GradientType=0);-webkit-box-shadow:inset 0 1px 0 #fff;-moz-box-shadow:inset 0 1px 0 #fff;box-shadow:inset 0 1px 0 #fff}.breadcrumb li{display:inline-block;*display:inline;text-shadow:0 1px 0 #fff;*zoom:1}.breadcrumb .divider{padding:0 5px;color:#999}.breadcrumb .active a{color:#333}.pagination{height:36px;margin:18px 0}.pagination ul{display:inline-block;*display:inline;margin-bottom:0;margin-left:0;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px;*zoom:1;-webkit-box-shadow:0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:0 1px 2px rgba(0,0,0,0.05);box-shadow:0 1px 2px rgba(0,0,0,0.05)}.pagination li{display:inline}.pagination a{float:left;padding:0 14px;line-height:34px;text-decoration:none;border:1px solid #ddd;border-left-width:0}.pagination a:hover,.pagination .active a{background-color:#f5f5f5}.pagination .active a{color:#999;cursor:default}.pagination .disabled span,.pagination .disabled a,.pagination .disabled a:hover{color:#999;cursor:default;background-color:transparent}.pagination li:first-child a{border-left-width:1px;-webkit-border-radius:3px 0 0 3px;-moz-border-radius:3px 0 0 3px;border-radius:3px 0 0 3px}.pagination li:last-child a{-webkit-border-radius:0 3px 3px 0;-moz-border-radius:0 3px 3px 0;border-radius:0 3px 3px 0}.pagination-centered{text-align:center}.pagination-right{text-align:right}.pager{margin-bottom:18px;margin-left:0;text-align:center;list-style:none;*zoom:1}.pager:before,.pager:after{display:table;content:""}.pager:after{clear:both}.pager li{display:inline}.pager a{display:inline-block;padding:5px 14px;background-color:#fff;border:1px solid #ddd;-webkit-border-radius:15px;-moz-border-radius:15px;border-radius:15px}.pager a:hover{text-decoration:none;background-color:#f5f5f5}.pager .next a{float:right}.pager .previous a{float:left}.pager .disabled a,.pager .disabled a:hover{color:#999;cursor:default;background-color:#fff}.modal-open .dropdown-menu{z-index:2050}.modal-open .dropdown.open{*z-index:2050}.modal-open .popover{z-index:2060}.modal-open .tooltip{z-index:2070}.modal-backdrop{position:fixed;top:0;right:0;bottom:0;left:0;z-index:1040;background-color:#000}.modal-backdrop.fade{opacity:0}.modal-backdrop,.modal-backdrop.fade.in{opacity:.8;filter:alpha(opacity=80)}.modal{position:fixed;top:50%;left:50%;z-index:1050;width:560px;margin:-250px 0 0 -280px;overflow:auto;background-color:#fff;border:1px solid #999;border:1px solid rgba(0,0,0,0.3);*border:1px solid #999;-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px;-webkit-box-shadow:0 3px 7px rgba(0,0,0,0.3);-moz-box-shadow:0 3px 7px rgba(0,0,0,0.3);box-shadow:0 3px 7px rgba(0,0,0,0.3);-webkit-background-clip:padding-box;-moz-background-clip:padding-box;background-clip:padding-box}.modal.fade{top:-25%;-webkit-transition:opacity .3s linear,top .3s ease-out;-moz-transition:opacity .3s linear,top .3s ease-out;-ms-transition:opacity .3s linear,top .3s ease-out;-o-transition:opacity .3s linear,top .3s ease-out;transition:opacity .3s linear,top .3s ease-out}.modal.fade.in{top:50%}.modal-header{padding:9px 15px;border-bottom:1px solid #eee}.modal-header .close{margin-top:2px}.modal-body{max-height:400px;padding:15px;overflow-y:auto}.modal-form{margin-bottom:0}.modal-footer{padding:14px 15px 15px;margin-bottom:0;text-align:right;background-color:#f5f5f5;border-top:1px solid #ddd;-webkit-border-radius:0 0 6px 6px;-moz-border-radius:0 0 6px 6px;border-radius:0 0 6px 6px;*zoom:1;-webkit-box-shadow:inset 0 1px 0 #fff;-moz-box-shadow:inset 0 1px 0 #fff;box-shadow:inset 0 1px 0 #fff}.modal-footer:before,.modal-footer:after{display:table;content:""}.modal-footer:after{clear:both}.modal-footer .btn+.btn{margin-bottom:0;margin-left:5px}.modal-footer .btn-group .btn+.btn{margin-left:-1px}.tooltip{position:absolute;z-index:1020;display:block;padding:5px;font-size:11px;opacity:0;filter:alpha(opacity=0);visibility:visible}.tooltip.in{opacity:.8;filter:alpha(opacity=80)}.tooltip.top{margin-top:-2px}.tooltip.right{margin-left:2px}.tooltip.bottom{margin-top:2px}.tooltip.left{margin-left:-2px}.tooltip.top .tooltip-arrow{bottom:0;left:50%;margin-left:-5px;border-top:5px solid #000;border-right:5px solid transparent;border-left:5px solid transparent}.tooltip.left .tooltip-arrow{top:50%;right:0;margin-top:-5px;border-top:5px solid transparent;border-bottom:5px solid transparent;border-left:5px solid #000}.tooltip.bottom .tooltip-arrow{top:0;left:50%;margin-left:-5px;border-right:5px solid transparent;border-bottom:5px solid #000;border-left:5px solid transparent}.tooltip.right .tooltip-arrow{top:50%;left:0;margin-top:-5px;border-top:5px solid transparent;border-right:5px solid #000;border-bottom:5px solid transparent}.tooltip-inner{max-width:200px;padding:3px 8px;color:#fff;text-align:center;text-decoration:none;background-color:#000;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.tooltip-arrow{position:absolute;width:0;height:0}.popover{position:absolute;top:0;left:0;z-index:1010;display:none;padding:5px}.popover.top{margin-top:-5px}.popover.right{margin-left:5px}.popover.bottom{margin-top:5px}.popover.left{margin-left:-5px}.popover.top .arrow{bottom:0;left:50%;margin-left:-5px;border-top:5px solid #000;border-right:5px solid transparent;border-left:5px solid transparent}.popover.right .arrow{top:50%;left:0;margin-top:-5px;border-top:5px solid transparent;border-right:5px solid #000;border-bottom:5px solid transparent}.popover.bottom .arrow{top:0;left:50%;margin-left:-5px;border-right:5px solid transparent;border-bottom:5px solid #000;border-left:5px solid transparent}.popover.left .arrow{top:50%;right:0;margin-top:-5px;border-top:5px solid transparent;border-bottom:5px solid transparent;border-left:5px solid #000}.popover .arrow{position:absolute;width:0;height:0}.popover-inner{width:280px;padding:3px;overflow:hidden;background:#000;background:rgba(0,0,0,0.8);-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px;-webkit-box-shadow:0 3px 7px rgba(0,0,0,0.3);-moz-box-shadow:0 3px 7px rgba(0,0,0,0.3);box-shadow:0 3px 7px rgba(0,0,0,0.3)}.popover-title{padding:9px 15px;line-height:1;background-color:#f5f5f5;border-bottom:1px solid #eee;-webkit-border-radius:3px 3px 0 0;-moz-border-radius:3px 3px 0 0;border-radius:3px 3px 0 0}.popover-content{padding:14px;background-color:#fff;-webkit-border-radius:0 0 3px 3px;-moz-border-radius:0 0 3px 3px;border-radius:0 0 3px 3px;-webkit-background-clip:padding-box;-moz-background-clip:padding-box;background-clip:padding-box}.popover-content p,.popover-content ul,.popover-content ol{margin-bottom:0}.thumbnails{margin-left:-20px;list-style:none;*zoom:1}.thumbnails:before,.thumbnails:after{display:table;content:""}.thumbnails:after{clear:both}.row-fluid .thumbnails{margin-left:0}.thumbnails>li{float:left;margin-bottom:18px;margin-left:20px}.thumbnail{display:block;padding:4px;line-height:1;border:1px solid #ddd;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;-webkit-box-shadow:0 1px 1px rgba(0,0,0,0.075);-moz-box-shadow:0 1px 1px rgba(0,0,0,0.075);box-shadow:0 1px 1px rgba(0,0,0,0.075)}a.thumbnail:hover{border-color:#08c;-webkit-box-shadow:0 1px 4px rgba(0,105,214,0.25);-moz-box-shadow:0 1px 4px rgba(0,105,214,0.25);box-shadow:0 1px 4px rgba(0,105,214,0.25)}.thumbnail>img{display:block;max-width:100%;margin-right:auto;margin-left:auto}.thumbnail .caption{padding:9px}.label,.badge{font-size:10.998px;font-weight:bold;line-height:14px;color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);white-space:nowrap;vertical-align:baseline;background-color:#999}.label{padding:1px 4px 2px;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}.badge{padding:1px 9px 2px;-webkit-border-radius:9px;-moz-border-radius:9px;border-radius:9px}a.label:hover,a.badge:hover{color:#fff;text-decoration:none;cursor:pointer}.label-important,.badge-important{background-color:#b94a48}.label-important[href],.badge-important[href]{background-color:#953b39}.label-warning,.badge-warning{background-color:#f89406}.label-warning[href],.badge-warning[href]{background-color:#c67605}.label-success,.badge-success{background-color:#468847}.label-success[href],.badge-success[href]{background-color:#356635}.label-info,.badge-info{background-color:#3a87ad}.label-info[href],.badge-info[href]{background-color:#2d6987}.label-inverse,.badge-inverse{background-color:#333}.label-inverse[href],.badge-inverse[href]{background-color:#1a1a1a}@-webkit-keyframes progress-bar-stripes{from{background-position:40px 0}to{background-position:0 0}}@-moz-keyframes progress-bar-stripes{from{background-position:40px 0}to{background-position:0 0}}@-ms-keyframes progress-bar-stripes{from{background-position:40px 0}to{background-position:0 0}}@-o-keyframes progress-bar-stripes{from{background-position:0 0}to{background-position:40px 0}}@keyframes progress-bar-stripes{from{background-position:40px 0}to{background-position:0 0}}.progress{height:18px;margin-bottom:18px;overflow:hidden;background-color:#f7f7f7;background-image:-moz-linear-gradient(top,#f5f5f5,#f9f9f9);background-image:-ms-linear-gradient(top,#f5f5f5,#f9f9f9);background-image:-webkit-gradient(linear,0 0,0 100%,from(#f5f5f5),to(#f9f9f9));background-image:-webkit-linear-gradient(top,#f5f5f5,#f9f9f9);background-image:-o-linear-gradient(top,#f5f5f5,#f9f9f9);background-image:linear-gradient(top,#f5f5f5,#f9f9f9);background-repeat:repeat-x;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#f5f5f5',endColorstr='#f9f9f9',GradientType=0);-webkit-box-shadow:inset 0 1px 2px rgba(0,0,0,0.1);-moz-box-shadow:inset 0 1px 2px rgba(0,0,0,0.1);box-shadow:inset 0 1px 2px rgba(0,0,0,0.1)}.progress .bar{width:0;height:18px;font-size:12px;color:#fff;text-align:center;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#0e90d2;background-image:-moz-linear-gradient(top,#149bdf,#0480be);background-image:-webkit-gradient(linear,0 0,0 100%,from(#149bdf),to(#0480be));background-image:-webkit-linear-gradient(top,#149bdf,#0480be);background-image:-o-linear-gradient(top,#149bdf,#0480be);background-image:linear-gradient(top,#149bdf,#0480be);background-image:-ms-linear-gradient(top,#149bdf,#0480be);background-repeat:repeat-x;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#149bdf',endColorstr='#0480be',GradientType=0);-webkit-box-shadow:inset 0 -1px 0 rgba(0,0,0,0.15);-moz-box-shadow:inset 0 -1px 0 rgba(0,0,0,0.15);box-shadow:inset 0 -1px 0 rgba(0,0,0,0.15);-webkit-box-sizing:border-box;-moz-box-sizing:border-box;-ms-box-sizing:border-box;box-sizing:border-box;-webkit-transition:width .6s ease;-moz-transition:width .6s ease;-ms-transition:width .6s ease;-o-transition:width .6s ease;transition:width .6s ease}.progress-striped .bar{background-color:#149bdf;background-image:-o-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-webkit-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-ms-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);-webkit-background-size:40px 40px;-moz-background-size:40px 40px;-o-background-size:40px 40px;background-size:40px 40px}.progress.active .bar{-webkit-animation:progress-bar-stripes 2s linear infinite;-moz-animation:progress-bar-stripes 2s linear infinite;-ms-animation:progress-bar-stripes 2s linear infinite;-o-animation:progress-bar-stripes 2s linear infinite;animation:progress-bar-stripes 2s linear infinite}.progress-danger .bar{background-color:#dd514c;background-image:-moz-linear-gradient(top,#ee5f5b,#c43c35);background-image:-ms-linear-gradient(top,#ee5f5b,#c43c35);background-image:-webkit-gradient(linear,0 0,0 100%,from(#ee5f5b),to(#c43c35));background-image:-webkit-linear-gradient(top,#ee5f5b,#c43c35);background-image:-o-linear-gradient(top,#ee5f5b,#c43c35);background-image:linear-gradient(top,#ee5f5b,#c43c35);background-repeat:repeat-x;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#ee5f5b',endColorstr='#c43c35',GradientType=0)}.progress-danger.progress-striped .bar{background-color:#ee5f5b;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-ms-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent)}.progress-success .bar{background-color:#5eb95e;background-image:-moz-linear-gradient(top,#62c462,#57a957);background-image:-ms-linear-gradient(top,#62c462,#57a957);background-image:-webkit-gradient(linear,0 0,0 100%,from(#62c462),to(#57a957));background-image:-webkit-linear-gradient(top,#62c462,#57a957);background-image:-o-linear-gradient(top,#62c462,#57a957);background-image:linear-gradient(top,#62c462,#57a957);background-repeat:repeat-x;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#62c462',endColorstr='#57a957',GradientType=0)}.progress-success.progress-striped .bar{background-color:#62c462;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-ms-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent)}.progress-info .bar{background-color:#4bb1cf;background-image:-moz-linear-gradient(top,#5bc0de,#339bb9);background-image:-ms-linear-gradient(top,#5bc0de,#339bb9);background-image:-webkit-gradient(linear,0 0,0 100%,from(#5bc0de),to(#339bb9));background-image:-webkit-linear-gradient(top,#5bc0de,#339bb9);background-image:-o-linear-gradient(top,#5bc0de,#339bb9);background-image:linear-gradient(top,#5bc0de,#339bb9);background-repeat:repeat-x;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#5bc0de',endColorstr='#339bb9',GradientType=0)}.progress-info.progress-striped .bar{background-color:#5bc0de;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-ms-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent)}.progress-warning .bar{background-color:#faa732;background-image:-moz-linear-gradient(top,#fbb450,#f89406);background-image:-ms-linear-gradient(top,#fbb450,#f89406);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fbb450),to(#f89406));background-image:-webkit-linear-gradient(top,#fbb450,#f89406);background-image:-o-linear-gradient(top,#fbb450,#f89406);background-image:linear-gradient(top,#fbb450,#f89406);background-repeat:repeat-x;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#fbb450',endColorstr='#f89406',GradientType=0)}.progress-warning.progress-striped .bar{background-color:#fbb450;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-ms-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(-45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent)}.accordion{margin-bottom:18px}.accordion-group{margin-bottom:2px;border:1px solid #e5e5e5;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.accordion-heading{border-bottom:0}.accordion-heading .accordion-toggle{display:block;padding:8px 15px}.accordion-toggle{cursor:pointer}.accordion-inner{padding:9px 15px;border-top:1px solid #e5e5e5}.carousel{position:relative;margin-bottom:18px;line-height:1}.carousel-inner{position:relative;width:100%;overflow:hidden}.carousel .item{position:relative;display:none;-webkit-transition:.6s ease-in-out left;-moz-transition:.6s ease-in-out left;-ms-transition:.6s ease-in-out left;-o-transition:.6s ease-in-out left;transition:.6s ease-in-out left}.carousel .item>img{display:block;line-height:1}.carousel .active,.carousel .next,.carousel .prev{display:block}.carousel .active{left:0}.carousel .next,.carousel .prev{position:absolute;top:0;width:100%}.carousel .next{left:100%}.carousel .prev{left:-100%}.carousel .next.left,.carousel .prev.right{left:0}.carousel .active.left{left:-100%}.carousel .active.right{left:100%}.carousel-control{position:absolute;top:40%;left:15px;width:40px;height:40px;margin-top:-20px;font-size:60px;font-weight:100;line-height:30px;color:#fff;text-align:center;background:#222;border:3px solid #fff;-webkit-border-radius:23px;-moz-border-radius:23px;border-radius:23px;opacity:.5;filter:alpha(opacity=50)}.carousel-control.right{right:15px;left:auto}.carousel-control:hover{color:#fff;text-decoration:none;opacity:.9;filter:alpha(opacity=90)}.carousel-caption{position:absolute;right:0;bottom:0;left:0;padding:10px 15px 5px;background:#333;background:rgba(0,0,0,0.75)}.carousel-caption h4,.carousel-caption p{color:#fff}.hero-unit{padding:60px;margin-bottom:30px;background-color:#eee;-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px}.hero-unit h1{margin-bottom:0;font-size:60px;line-height:1;letter-spacing:-1px;color:inherit}.hero-unit p{font-size:18px;font-weight:200;line-height:27px;color:inherit}.pull-right{float:right}.pull-left{float:left}.hide{display:none}.show{display:block}.invisible{visibility:hidden} diff --git a/core/src/main/resources/spark/ui/static/bootstrap-responsive.min.css b/core/src/main/resources/spark/ui/static/bootstrap-responsive.min.css new file mode 100644 index 0000000000..f4ede63f32 --- /dev/null +++ b/core/src/main/resources/spark/ui/static/bootstrap-responsive.min.css @@ -0,0 +1,9 @@ +/*! + * Bootstrap Responsive v2.3.2 + * + * Copyright 2012 Twitter, Inc + * Licensed under the Apache License v2.0 + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Designed and built with all the love in the world @twitter by @mdo and @fat. + */.clearfix{*zoom:1}.clearfix:before,.clearfix:after{display:table;line-height:0;content:""}.clearfix:after{clear:both}.hide-text{font:0/0 a;color:transparent;text-shadow:none;background-color:transparent;border:0}.input-block-level{display:block;width:100%;min-height:30px;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}@-ms-viewport{width:device-width}.hidden{display:none;visibility:hidden}.visible-phone{display:none!important}.visible-tablet{display:none!important}.hidden-desktop{display:none!important}.visible-desktop{display:inherit!important}@media(min-width:768px) and (max-width:979px){.hidden-desktop{display:inherit!important}.visible-desktop{display:none!important}.visible-tablet{display:inherit!important}.hidden-tablet{display:none!important}}@media(max-width:767px){.hidden-desktop{display:inherit!important}.visible-desktop{display:none!important}.visible-phone{display:inherit!important}.hidden-phone{display:none!important}}.visible-print{display:none!important}@media print{.visible-print{display:inherit!important}.hidden-print{display:none!important}}@media(min-width:1200px){.row{margin-left:-30px;*zoom:1}.row:before,.row:after{display:table;line-height:0;content:""}.row:after{clear:both}[class*="span"]{float:left;min-height:1px;margin-left:30px}.container,.navbar-static-top .container,.navbar-fixed-top .container,.navbar-fixed-bottom .container{width:1170px}.span12{width:1170px}.span11{width:1070px}.span10{width:970px}.span9{width:870px}.span8{width:770px}.span7{width:670px}.span6{width:570px}.span5{width:470px}.span4{width:370px}.span3{width:270px}.span2{width:170px}.span1{width:70px}.offset12{margin-left:1230px}.offset11{margin-left:1130px}.offset10{margin-left:1030px}.offset9{margin-left:930px}.offset8{margin-left:830px}.offset7{margin-left:730px}.offset6{margin-left:630px}.offset5{margin-left:530px}.offset4{margin-left:430px}.offset3{margin-left:330px}.offset2{margin-left:230px}.offset1{margin-left:130px}.row-fluid{width:100%;*zoom:1}.row-fluid:before,.row-fluid:after{display:table;line-height:0;content:""}.row-fluid:after{clear:both}.row-fluid [class*="span"]{display:block;float:left;width:100%;min-height:30px;margin-left:2.564102564102564%;*margin-left:2.5109110747408616%;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}.row-fluid [class*="span"]:first-child{margin-left:0}.row-fluid .controls-row [class*="span"]+[class*="span"]{margin-left:2.564102564102564%}.row-fluid .span12{width:100%;*width:99.94680851063829%}.row-fluid .span11{width:91.45299145299145%;*width:91.39979996362975%}.row-fluid .span10{width:82.90598290598291%;*width:82.8527914166212%}.row-fluid .span9{width:74.35897435897436%;*width:74.30578286961266%}.row-fluid .span8{width:65.81196581196582%;*width:65.75877432260411%}.row-fluid .span7{width:57.26495726495726%;*width:57.21176577559556%}.row-fluid .span6{width:48.717948717948715%;*width:48.664757228587014%}.row-fluid .span5{width:40.17094017094017%;*width:40.11774868157847%}.row-fluid .span4{width:31.623931623931625%;*width:31.570740134569924%}.row-fluid .span3{width:23.076923076923077%;*width:23.023731587561375%}.row-fluid .span2{width:14.52991452991453%;*width:14.476723040552828%}.row-fluid .span1{width:5.982905982905983%;*width:5.929714493544281%}.row-fluid .offset12{margin-left:105.12820512820512%;*margin-left:105.02182214948171%}.row-fluid .offset12:first-child{margin-left:102.56410256410257%;*margin-left:102.45771958537915%}.row-fluid .offset11{margin-left:96.58119658119658%;*margin-left:96.47481360247316%}.row-fluid .offset11:first-child{margin-left:94.01709401709402%;*margin-left:93.91071103837061%}.row-fluid .offset10{margin-left:88.03418803418803%;*margin-left:87.92780505546462%}.row-fluid .offset10:first-child{margin-left:85.47008547008548%;*margin-left:85.36370249136206%}.row-fluid .offset9{margin-left:79.48717948717949%;*margin-left:79.38079650845607%}.row-fluid .offset9:first-child{margin-left:76.92307692307693%;*margin-left:76.81669394435352%}.row-fluid .offset8{margin-left:70.94017094017094%;*margin-left:70.83378796144753%}.row-fluid .offset8:first-child{margin-left:68.37606837606839%;*margin-left:68.26968539734497%}.row-fluid .offset7{margin-left:62.393162393162385%;*margin-left:62.28677941443899%}.row-fluid .offset7:first-child{margin-left:59.82905982905982%;*margin-left:59.72267685033642%}.row-fluid .offset6{margin-left:53.84615384615384%;*margin-left:53.739770867430444%}.row-fluid .offset6:first-child{margin-left:51.28205128205128%;*margin-left:51.175668303327875%}.row-fluid .offset5{margin-left:45.299145299145295%;*margin-left:45.1927623204219%}.row-fluid .offset5:first-child{margin-left:42.73504273504273%;*margin-left:42.62865975631933%}.row-fluid .offset4{margin-left:36.75213675213675%;*margin-left:36.645753773413354%}.row-fluid .offset4:first-child{margin-left:34.18803418803419%;*margin-left:34.081651209310785%}.row-fluid .offset3{margin-left:28.205128205128204%;*margin-left:28.0987452264048%}.row-fluid .offset3:first-child{margin-left:25.641025641025642%;*margin-left:25.53464266230224%}.row-fluid .offset2{margin-left:19.65811965811966%;*margin-left:19.551736679396257%}.row-fluid .offset2:first-child{margin-left:17.094017094017094%;*margin-left:16.98763411529369%}.row-fluid .offset1{margin-left:11.11111111111111%;*margin-left:11.004728132387708%}.row-fluid .offset1:first-child{margin-left:8.547008547008547%;*margin-left:8.440625568285142%}input,textarea,.uneditable-input{margin-left:0}.controls-row [class*="span"]+[class*="span"]{margin-left:30px}input.span12,textarea.span12,.uneditable-input.span12{width:1156px}input.span11,textarea.span11,.uneditable-input.span11{width:1056px}input.span10,textarea.span10,.uneditable-input.span10{width:956px}input.span9,textarea.span9,.uneditable-input.span9{width:856px}input.span8,textarea.span8,.uneditable-input.span8{width:756px}input.span7,textarea.span7,.uneditable-input.span7{width:656px}input.span6,textarea.span6,.uneditable-input.span6{width:556px}input.span5,textarea.span5,.uneditable-input.span5{width:456px}input.span4,textarea.span4,.uneditable-input.span4{width:356px}input.span3,textarea.span3,.uneditable-input.span3{width:256px}input.span2,textarea.span2,.uneditable-input.span2{width:156px}input.span1,textarea.span1,.uneditable-input.span1{width:56px}.thumbnails{margin-left:-30px}.thumbnails>li{margin-left:30px}.row-fluid .thumbnails{margin-left:0}}@media(min-width:768px) and (max-width:979px){.row{margin-left:-20px;*zoom:1}.row:before,.row:after{display:table;line-height:0;content:""}.row:after{clear:both}[class*="span"]{float:left;min-height:1px;margin-left:20px}.container,.navbar-static-top .container,.navbar-fixed-top .container,.navbar-fixed-bottom .container{width:724px}.span12{width:724px}.span11{width:662px}.span10{width:600px}.span9{width:538px}.span8{width:476px}.span7{width:414px}.span6{width:352px}.span5{width:290px}.span4{width:228px}.span3{width:166px}.span2{width:104px}.span1{width:42px}.offset12{margin-left:764px}.offset11{margin-left:702px}.offset10{margin-left:640px}.offset9{margin-left:578px}.offset8{margin-left:516px}.offset7{margin-left:454px}.offset6{margin-left:392px}.offset5{margin-left:330px}.offset4{margin-left:268px}.offset3{margin-left:206px}.offset2{margin-left:144px}.offset1{margin-left:82px}.row-fluid{width:100%;*zoom:1}.row-fluid:before,.row-fluid:after{display:table;line-height:0;content:""}.row-fluid:after{clear:both}.row-fluid [class*="span"]{display:block;float:left;width:100%;min-height:30px;margin-left:2.7624309392265194%;*margin-left:2.709239449864817%;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}.row-fluid [class*="span"]:first-child{margin-left:0}.row-fluid .controls-row [class*="span"]+[class*="span"]{margin-left:2.7624309392265194%}.row-fluid .span12{width:100%;*width:99.94680851063829%}.row-fluid .span11{width:91.43646408839778%;*width:91.38327259903608%}.row-fluid .span10{width:82.87292817679558%;*width:82.81973668743387%}.row-fluid .span9{width:74.30939226519337%;*width:74.25620077583166%}.row-fluid .span8{width:65.74585635359117%;*width:65.69266486422946%}.row-fluid .span7{width:57.18232044198895%;*width:57.12912895262725%}.row-fluid .span6{width:48.61878453038674%;*width:48.56559304102504%}.row-fluid .span5{width:40.05524861878453%;*width:40.00205712942283%}.row-fluid .span4{width:31.491712707182323%;*width:31.43852121782062%}.row-fluid .span3{width:22.92817679558011%;*width:22.87498530621841%}.row-fluid .span2{width:14.3646408839779%;*width:14.311449394616199%}.row-fluid .span1{width:5.801104972375691%;*width:5.747913483013988%}.row-fluid .offset12{margin-left:105.52486187845304%;*margin-left:105.41847889972962%}.row-fluid .offset12:first-child{margin-left:102.76243093922652%;*margin-left:102.6560479605031%}.row-fluid .offset11{margin-left:96.96132596685082%;*margin-left:96.8549429881274%}.row-fluid .offset11:first-child{margin-left:94.1988950276243%;*margin-left:94.09251204890089%}.row-fluid .offset10{margin-left:88.39779005524862%;*margin-left:88.2914070765252%}.row-fluid .offset10:first-child{margin-left:85.6353591160221%;*margin-left:85.52897613729868%}.row-fluid .offset9{margin-left:79.8342541436464%;*margin-left:79.72787116492299%}.row-fluid .offset9:first-child{margin-left:77.07182320441989%;*margin-left:76.96544022569647%}.row-fluid .offset8{margin-left:71.2707182320442%;*margin-left:71.16433525332079%}.row-fluid .offset8:first-child{margin-left:68.50828729281768%;*margin-left:68.40190431409427%}.row-fluid .offset7{margin-left:62.70718232044199%;*margin-left:62.600799341718584%}.row-fluid .offset7:first-child{margin-left:59.94475138121547%;*margin-left:59.838368402492065%}.row-fluid .offset6{margin-left:54.14364640883978%;*margin-left:54.037263430116376%}.row-fluid .offset6:first-child{margin-left:51.38121546961326%;*margin-left:51.27483249088986%}.row-fluid .offset5{margin-left:45.58011049723757%;*margin-left:45.47372751851417%}.row-fluid .offset5:first-child{margin-left:42.81767955801105%;*margin-left:42.71129657928765%}.row-fluid .offset4{margin-left:37.01657458563536%;*margin-left:36.91019160691196%}.row-fluid .offset4:first-child{margin-left:34.25414364640884%;*margin-left:34.14776066768544%}.row-fluid .offset3{margin-left:28.45303867403315%;*margin-left:28.346655695309746%}.row-fluid .offset3:first-child{margin-left:25.69060773480663%;*margin-left:25.584224756083227%}.row-fluid .offset2{margin-left:19.88950276243094%;*margin-left:19.783119783707537%}.row-fluid .offset2:first-child{margin-left:17.12707182320442%;*margin-left:17.02068884448102%}.row-fluid .offset1{margin-left:11.32596685082873%;*margin-left:11.219583872105325%}.row-fluid .offset1:first-child{margin-left:8.56353591160221%;*margin-left:8.457152932878806%}input,textarea,.uneditable-input{margin-left:0}.controls-row [class*="span"]+[class*="span"]{margin-left:20px}input.span12,textarea.span12,.uneditable-input.span12{width:710px}input.span11,textarea.span11,.uneditable-input.span11{width:648px}input.span10,textarea.span10,.uneditable-input.span10{width:586px}input.span9,textarea.span9,.uneditable-input.span9{width:524px}input.span8,textarea.span8,.uneditable-input.span8{width:462px}input.span7,textarea.span7,.uneditable-input.span7{width:400px}input.span6,textarea.span6,.uneditable-input.span6{width:338px}input.span5,textarea.span5,.uneditable-input.span5{width:276px}input.span4,textarea.span4,.uneditable-input.span4{width:214px}input.span3,textarea.span3,.uneditable-input.span3{width:152px}input.span2,textarea.span2,.uneditable-input.span2{width:90px}input.span1,textarea.span1,.uneditable-input.span1{width:28px}}@media(max-width:767px){body{padding-right:20px;padding-left:20px}.navbar-fixed-top,.navbar-fixed-bottom,.navbar-static-top{margin-right:-20px;margin-left:-20px}.container-fluid{padding:0}.dl-horizontal dt{float:none;width:auto;clear:none;text-align:left}.dl-horizontal dd{margin-left:0}.container{width:auto}.row-fluid{width:100%}.row,.thumbnails{margin-left:0}.thumbnails>li{float:none;margin-left:0}[class*="span"],.uneditable-input[class*="span"],.row-fluid [class*="span"]{display:block;float:none;width:100%;margin-left:0;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}.span12,.row-fluid .span12{width:100%;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}.row-fluid [class*="offset"]:first-child{margin-left:0}.input-large,.input-xlarge,.input-xxlarge,input[class*="span"],select[class*="span"],textarea[class*="span"],.uneditable-input{display:block;width:100%;min-height:30px;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}.input-prepend input,.input-append input,.input-prepend input[class*="span"],.input-append input[class*="span"]{display:inline-block;width:auto}.controls-row [class*="span"]+[class*="span"]{margin-left:0}.modal{position:fixed;top:20px;right:20px;left:20px;width:auto;margin:0}.modal.fade{top:-100px}.modal.fade.in{top:20px}}@media(max-width:480px){.nav-collapse{-webkit-transform:translate3d(0,0,0)}.page-header h1 small{display:block;line-height:20px}input[type="checkbox"],input[type="radio"]{border:1px solid #ccc}.form-horizontal .control-label{float:none;width:auto;padding-top:0;text-align:left}.form-horizontal .controls{margin-left:0}.form-horizontal .control-list{padding-top:0}.form-horizontal .form-actions{padding-right:10px;padding-left:10px}.media .pull-left,.media .pull-right{display:block;float:none;margin-bottom:10px}.media-object{margin-right:0;margin-left:0}.modal{top:10px;right:10px;left:10px}.modal-header .close{padding:10px;margin:-10px}.carousel-caption{position:static}}@media(max-width:979px){body{padding-top:0}.navbar-fixed-top,.navbar-fixed-bottom{position:static}.navbar-fixed-top{margin-bottom:20px}.navbar-fixed-bottom{margin-top:20px}.navbar-fixed-top .navbar-inner,.navbar-fixed-bottom .navbar-inner{padding:5px}.navbar .container{width:auto;padding:0}.navbar .brand{padding-right:10px;padding-left:10px;margin:0 0 0 -5px}.nav-collapse{clear:both}.nav-collapse .nav{float:none;margin:0 0 10px}.nav-collapse .nav>li{float:none}.nav-collapse .nav>li>a{margin-bottom:2px}.nav-collapse .nav>.divider-vertical{display:none}.nav-collapse .nav .nav-header{color:#777;text-shadow:none}.nav-collapse .nav>li>a,.nav-collapse .dropdown-menu a{padding:9px 15px;font-weight:bold;color:#777;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}.nav-collapse .btn{padding:4px 10px 4px;font-weight:normal;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.nav-collapse .dropdown-menu li+li a{margin-bottom:2px}.nav-collapse .nav>li>a:hover,.nav-collapse .nav>li>a:focus,.nav-collapse .dropdown-menu a:hover,.nav-collapse .dropdown-menu a:focus{background-color:#f2f2f2}.navbar-inverse .nav-collapse .nav>li>a,.navbar-inverse .nav-collapse .dropdown-menu a{color:#999}.navbar-inverse .nav-collapse .nav>li>a:hover,.navbar-inverse .nav-collapse .nav>li>a:focus,.navbar-inverse .nav-collapse .dropdown-menu a:hover,.navbar-inverse .nav-collapse .dropdown-menu a:focus{background-color:#111}.nav-collapse.in .btn-group{padding:0;margin-top:5px}.nav-collapse .dropdown-menu{position:static;top:auto;left:auto;display:none;float:none;max-width:none;padding:0;margin:0 15px;background-color:transparent;border:0;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0;-webkit-box-shadow:none;-moz-box-shadow:none;box-shadow:none}.nav-collapse .open>.dropdown-menu{display:block}.nav-collapse .dropdown-menu:before,.nav-collapse .dropdown-menu:after{display:none}.nav-collapse .dropdown-menu .divider{display:none}.nav-collapse .nav>li>.dropdown-menu:before,.nav-collapse .nav>li>.dropdown-menu:after{display:none}.nav-collapse .navbar-form,.nav-collapse .navbar-search{float:none;padding:10px 15px;margin:10px 0;border-top:1px solid #f2f2f2;border-bottom:1px solid #f2f2f2;-webkit-box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.1);-moz-box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.1);box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.1)}.navbar-inverse .nav-collapse .navbar-form,.navbar-inverse .nav-collapse .navbar-search{border-top-color:#111;border-bottom-color:#111}.navbar .nav-collapse .nav.pull-right{float:none;margin-left:0}.nav-collapse,.nav-collapse.collapse{height:0;overflow:hidden}.navbar .btn-navbar{display:block}.navbar-static .navbar-inner{padding-right:10px;padding-left:10px}}@media(min-width:980px){.nav-collapse.collapse{height:auto!important;overflow:visible!important}} diff --git a/core/src/main/resources/spark/ui/static/bootstrap.min.css b/core/src/main/resources/spark/ui/static/bootstrap.min.css new file mode 100644 index 0000000000..b6428e6958 --- /dev/null +++ b/core/src/main/resources/spark/ui/static/bootstrap.min.css @@ -0,0 +1,9 @@ +/*! + * Bootstrap v2.3.2 + * + * Copyright 2012 Twitter, Inc + * Licensed under the Apache License v2.0 + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Designed and built with all the love in the world @twitter by @mdo and @fat. + */.clearfix{*zoom:1}.clearfix:before,.clearfix:after{display:table;line-height:0;content:""}.clearfix:after{clear:both}.hide-text{font:0/0 a;color:transparent;text-shadow:none;background-color:transparent;border:0}.input-block-level{display:block;width:100%;min-height:30px;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}article,aside,details,figcaption,figure,footer,header,hgroup,nav,section{display:block}audio,canvas,video{display:inline-block;*display:inline;*zoom:1}audio:not([controls]){display:none}html{font-size:100%;-webkit-text-size-adjust:100%;-ms-text-size-adjust:100%}a:focus{outline:thin dotted #333;outline:5px auto -webkit-focus-ring-color;outline-offset:-2px}a:hover,a:active{outline:0}sub,sup{position:relative;font-size:75%;line-height:0;vertical-align:baseline}sup{top:-0.5em}sub{bottom:-0.25em}img{width:auto\9;height:auto;max-width:100%;vertical-align:middle;border:0;-ms-interpolation-mode:bicubic}#map_canvas img,.google-maps img{max-width:none}button,input,select,textarea{margin:0;font-size:100%;vertical-align:middle}button,input{*overflow:visible;line-height:normal}button::-moz-focus-inner,input::-moz-focus-inner{padding:0;border:0}button,html input[type="button"],input[type="reset"],input[type="submit"]{cursor:pointer;-webkit-appearance:button}label,select,button,input[type="button"],input[type="reset"],input[type="submit"],input[type="radio"],input[type="checkbox"]{cursor:pointer}input[type="search"]{-webkit-box-sizing:content-box;-moz-box-sizing:content-box;box-sizing:content-box;-webkit-appearance:textfield}input[type="search"]::-webkit-search-decoration,input[type="search"]::-webkit-search-cancel-button{-webkit-appearance:none}textarea{overflow:auto;vertical-align:top}@media print{*{color:#000!important;text-shadow:none!important;background:transparent!important;box-shadow:none!important}a,a:visited{text-decoration:underline}a[href]:after{content:" (" attr(href) ")"}abbr[title]:after{content:" (" attr(title) ")"}.ir a:after,a[href^="javascript:"]:after,a[href^="#"]:after{content:""}pre,blockquote{border:1px solid #999;page-break-inside:avoid}thead{display:table-header-group}tr,img{page-break-inside:avoid}img{max-width:100%!important}@page{margin:.5cm}p,h2,h3{orphans:3;widows:3}h2,h3{page-break-after:avoid}}body{margin:0;font-family:"Helvetica Neue",Helvetica,Arial,sans-serif;font-size:14px;line-height:20px;color:#333;background-color:#fff}a{color:#08c;text-decoration:none}a:hover,a:focus{color:#005580;text-decoration:underline}.img-rounded{-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px}.img-polaroid{padding:4px;background-color:#fff;border:1px solid #ccc;border:1px solid rgba(0,0,0,0.2);-webkit-box-shadow:0 1px 3px rgba(0,0,0,0.1);-moz-box-shadow:0 1px 3px rgba(0,0,0,0.1);box-shadow:0 1px 3px rgba(0,0,0,0.1)}.img-circle{-webkit-border-radius:500px;-moz-border-radius:500px;border-radius:500px}.row{margin-left:-20px;*zoom:1}.row:before,.row:after{display:table;line-height:0;content:""}.row:after{clear:both}[class*="span"]{float:left;min-height:1px;margin-left:20px}.container,.navbar-static-top .container,.navbar-fixed-top .container,.navbar-fixed-bottom .container{width:940px}.span12{width:940px}.span11{width:860px}.span10{width:780px}.span9{width:700px}.span8{width:620px}.span7{width:540px}.span6{width:460px}.span5{width:380px}.span4{width:300px}.span3{width:220px}.span2{width:140px}.span1{width:60px}.offset12{margin-left:980px}.offset11{margin-left:900px}.offset10{margin-left:820px}.offset9{margin-left:740px}.offset8{margin-left:660px}.offset7{margin-left:580px}.offset6{margin-left:500px}.offset5{margin-left:420px}.offset4{margin-left:340px}.offset3{margin-left:260px}.offset2{margin-left:180px}.offset1{margin-left:100px}.row-fluid{width:100%;*zoom:1}.row-fluid:before,.row-fluid:after{display:table;line-height:0;content:""}.row-fluid:after{clear:both}.row-fluid [class*="span"]{display:block;float:left;width:100%;min-height:30px;margin-left:2.127659574468085%;*margin-left:2.074468085106383%;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}.row-fluid [class*="span"]:first-child{margin-left:0}.row-fluid .controls-row [class*="span"]+[class*="span"]{margin-left:2.127659574468085%}.row-fluid .span12{width:100%;*width:99.94680851063829%}.row-fluid .span11{width:91.48936170212765%;*width:91.43617021276594%}.row-fluid .span10{width:82.97872340425532%;*width:82.92553191489361%}.row-fluid .span9{width:74.46808510638297%;*width:74.41489361702126%}.row-fluid .span8{width:65.95744680851064%;*width:65.90425531914893%}.row-fluid .span7{width:57.44680851063829%;*width:57.39361702127659%}.row-fluid .span6{width:48.93617021276595%;*width:48.88297872340425%}.row-fluid .span5{width:40.42553191489362%;*width:40.37234042553192%}.row-fluid .span4{width:31.914893617021278%;*width:31.861702127659576%}.row-fluid .span3{width:23.404255319148934%;*width:23.351063829787233%}.row-fluid .span2{width:14.893617021276595%;*width:14.840425531914894%}.row-fluid .span1{width:6.382978723404255%;*width:6.329787234042553%}.row-fluid .offset12{margin-left:104.25531914893617%;*margin-left:104.14893617021275%}.row-fluid .offset12:first-child{margin-left:102.12765957446808%;*margin-left:102.02127659574467%}.row-fluid .offset11{margin-left:95.74468085106382%;*margin-left:95.6382978723404%}.row-fluid .offset11:first-child{margin-left:93.61702127659574%;*margin-left:93.51063829787232%}.row-fluid .offset10{margin-left:87.23404255319149%;*margin-left:87.12765957446807%}.row-fluid .offset10:first-child{margin-left:85.1063829787234%;*margin-left:84.99999999999999%}.row-fluid .offset9{margin-left:78.72340425531914%;*margin-left:78.61702127659572%}.row-fluid .offset9:first-child{margin-left:76.59574468085106%;*margin-left:76.48936170212764%}.row-fluid .offset8{margin-left:70.2127659574468%;*margin-left:70.10638297872339%}.row-fluid .offset8:first-child{margin-left:68.08510638297872%;*margin-left:67.9787234042553%}.row-fluid .offset7{margin-left:61.70212765957446%;*margin-left:61.59574468085106%}.row-fluid .offset7:first-child{margin-left:59.574468085106375%;*margin-left:59.46808510638297%}.row-fluid .offset6{margin-left:53.191489361702125%;*margin-left:53.085106382978715%}.row-fluid .offset6:first-child{margin-left:51.063829787234035%;*margin-left:50.95744680851063%}.row-fluid .offset5{margin-left:44.68085106382979%;*margin-left:44.57446808510638%}.row-fluid .offset5:first-child{margin-left:42.5531914893617%;*margin-left:42.4468085106383%}.row-fluid .offset4{margin-left:36.170212765957444%;*margin-left:36.06382978723405%}.row-fluid .offset4:first-child{margin-left:34.04255319148936%;*margin-left:33.93617021276596%}.row-fluid .offset3{margin-left:27.659574468085104%;*margin-left:27.5531914893617%}.row-fluid .offset3:first-child{margin-left:25.53191489361702%;*margin-left:25.425531914893618%}.row-fluid .offset2{margin-left:19.148936170212764%;*margin-left:19.04255319148936%}.row-fluid .offset2:first-child{margin-left:17.02127659574468%;*margin-left:16.914893617021278%}.row-fluid .offset1{margin-left:10.638297872340425%;*margin-left:10.53191489361702%}.row-fluid .offset1:first-child{margin-left:8.51063829787234%;*margin-left:8.404255319148938%}[class*="span"].hide,.row-fluid [class*="span"].hide{display:none}[class*="span"].pull-right,.row-fluid [class*="span"].pull-right{float:right}.container{margin-right:auto;margin-left:auto;*zoom:1}.container:before,.container:after{display:table;line-height:0;content:""}.container:after{clear:both}.container-fluid{padding-right:20px;padding-left:20px;*zoom:1}.container-fluid:before,.container-fluid:after{display:table;line-height:0;content:""}.container-fluid:after{clear:both}p{margin:0 0 10px}.lead{margin-bottom:20px;font-size:21px;font-weight:200;line-height:30px}small{font-size:85%}strong{font-weight:bold}em{font-style:italic}cite{font-style:normal}.muted{color:#999}a.muted:hover,a.muted:focus{color:#808080}.text-warning{color:#c09853}a.text-warning:hover,a.text-warning:focus{color:#a47e3c}.text-error{color:#b94a48}a.text-error:hover,a.text-error:focus{color:#953b39}.text-info{color:#3a87ad}a.text-info:hover,a.text-info:focus{color:#2d6987}.text-success{color:#468847}a.text-success:hover,a.text-success:focus{color:#356635}.text-left{text-align:left}.text-right{text-align:right}.text-center{text-align:center}h1,h2,h3,h4,h5,h6{margin:10px 0;font-family:inherit;font-weight:bold;line-height:20px;color:inherit;text-rendering:optimizelegibility}h1 small,h2 small,h3 small,h4 small,h5 small,h6 small{font-weight:normal;line-height:1;color:#999}h1,h2,h3{line-height:40px}h1{font-size:38.5px}h2{font-size:31.5px}h3{font-size:24.5px}h4{font-size:17.5px}h5{font-size:14px}h6{font-size:11.9px}h1 small{font-size:24.5px}h2 small{font-size:17.5px}h3 small{font-size:14px}h4 small{font-size:14px}.page-header{padding-bottom:9px;margin:20px 0 30px;border-bottom:1px solid #eee}ul,ol{padding:0;margin:0 0 10px 25px}ul ul,ul ol,ol ol,ol ul{margin-bottom:0}li{line-height:20px}ul.unstyled,ol.unstyled{margin-left:0;list-style:none}ul.inline,ol.inline{margin-left:0;list-style:none}ul.inline>li,ol.inline>li{display:inline-block;*display:inline;padding-right:5px;padding-left:5px;*zoom:1}dl{margin-bottom:20px}dt,dd{line-height:20px}dt{font-weight:bold}dd{margin-left:10px}.dl-horizontal{*zoom:1}.dl-horizontal:before,.dl-horizontal:after{display:table;line-height:0;content:""}.dl-horizontal:after{clear:both}.dl-horizontal dt{float:left;width:160px;overflow:hidden;clear:left;text-align:right;text-overflow:ellipsis;white-space:nowrap}.dl-horizontal dd{margin-left:180px}hr{margin:20px 0;border:0;border-top:1px solid #eee;border-bottom:1px solid #fff}abbr[title],abbr[data-original-title]{cursor:help;border-bottom:1px dotted #999}abbr.initialism{font-size:90%;text-transform:uppercase}blockquote{padding:0 0 0 15px;margin:0 0 20px;border-left:5px solid #eee}blockquote p{margin-bottom:0;font-size:17.5px;font-weight:300;line-height:1.25}blockquote small{display:block;line-height:20px;color:#999}blockquote small:before{content:'\2014 \00A0'}blockquote.pull-right{float:right;padding-right:15px;padding-left:0;border-right:5px solid #eee;border-left:0}blockquote.pull-right p,blockquote.pull-right small{text-align:right}blockquote.pull-right small:before{content:''}blockquote.pull-right small:after{content:'\00A0 \2014'}q:before,q:after,blockquote:before,blockquote:after{content:""}address{display:block;margin-bottom:20px;font-style:normal;line-height:20px}code,pre{padding:0 3px 2px;font-family:Monaco,Menlo,Consolas,"Courier New",monospace;font-size:12px;color:#333;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}code{padding:2px 4px;color:#d14;white-space:nowrap;background-color:#f7f7f9;border:1px solid #e1e1e8}pre{display:block;padding:9.5px;margin:0 0 10px;font-size:13px;line-height:20px;word-break:break-all;word-wrap:break-word;white-space:pre;white-space:pre-wrap;background-color:#f5f5f5;border:1px solid #ccc;border:1px solid rgba(0,0,0,0.15);-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}pre.prettyprint{margin-bottom:20px}pre code{padding:0;color:inherit;white-space:pre;white-space:pre-wrap;background-color:transparent;border:0}.pre-scrollable{max-height:340px;overflow-y:scroll}form{margin:0 0 20px}fieldset{padding:0;margin:0;border:0}legend{display:block;width:100%;padding:0;margin-bottom:20px;font-size:21px;line-height:40px;color:#333;border:0;border-bottom:1px solid #e5e5e5}legend small{font-size:15px;color:#999}label,input,button,select,textarea{font-size:14px;font-weight:normal;line-height:20px}input,button,select,textarea{font-family:"Helvetica Neue",Helvetica,Arial,sans-serif}label{display:block;margin-bottom:5px}select,textarea,input[type="text"],input[type="password"],input[type="datetime"],input[type="datetime-local"],input[type="date"],input[type="month"],input[type="time"],input[type="week"],input[type="number"],input[type="email"],input[type="url"],input[type="search"],input[type="tel"],input[type="color"],.uneditable-input{display:inline-block;height:20px;padding:4px 6px;margin-bottom:10px;font-size:14px;line-height:20px;color:#555;vertical-align:middle;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}input,textarea,.uneditable-input{width:206px}textarea{height:auto}textarea,input[type="text"],input[type="password"],input[type="datetime"],input[type="datetime-local"],input[type="date"],input[type="month"],input[type="time"],input[type="week"],input[type="number"],input[type="email"],input[type="url"],input[type="search"],input[type="tel"],input[type="color"],.uneditable-input{background-color:#fff;border:1px solid #ccc;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);-webkit-transition:border linear .2s,box-shadow linear .2s;-moz-transition:border linear .2s,box-shadow linear .2s;-o-transition:border linear .2s,box-shadow linear .2s;transition:border linear .2s,box-shadow linear .2s}textarea:focus,input[type="text"]:focus,input[type="password"]:focus,input[type="datetime"]:focus,input[type="datetime-local"]:focus,input[type="date"]:focus,input[type="month"]:focus,input[type="time"]:focus,input[type="week"]:focus,input[type="number"]:focus,input[type="email"]:focus,input[type="url"]:focus,input[type="search"]:focus,input[type="tel"]:focus,input[type="color"]:focus,.uneditable-input:focus{border-color:rgba(82,168,236,0.8);outline:0;outline:thin dotted \9;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 8px rgba(82,168,236,0.6);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 8px rgba(82,168,236,0.6);box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 8px rgba(82,168,236,0.6)}input[type="radio"],input[type="checkbox"]{margin:4px 0 0;margin-top:1px \9;*margin-top:0;line-height:normal}input[type="file"],input[type="image"],input[type="submit"],input[type="reset"],input[type="button"],input[type="radio"],input[type="checkbox"]{width:auto}select,input[type="file"]{height:30px;*margin-top:4px;line-height:30px}select{width:220px;background-color:#fff;border:1px solid #ccc}select[multiple],select[size]{height:auto}select:focus,input[type="file"]:focus,input[type="radio"]:focus,input[type="checkbox"]:focus{outline:thin dotted #333;outline:5px auto -webkit-focus-ring-color;outline-offset:-2px}.uneditable-input,.uneditable-textarea{color:#999;cursor:not-allowed;background-color:#fcfcfc;border-color:#ccc;-webkit-box-shadow:inset 0 1px 2px rgba(0,0,0,0.025);-moz-box-shadow:inset 0 1px 2px rgba(0,0,0,0.025);box-shadow:inset 0 1px 2px rgba(0,0,0,0.025)}.uneditable-input{overflow:hidden;white-space:nowrap}.uneditable-textarea{width:auto;height:auto}input:-moz-placeholder,textarea:-moz-placeholder{color:#999}input:-ms-input-placeholder,textarea:-ms-input-placeholder{color:#999}input::-webkit-input-placeholder,textarea::-webkit-input-placeholder{color:#999}.radio,.checkbox{min-height:20px;padding-left:20px}.radio input[type="radio"],.checkbox input[type="checkbox"]{float:left;margin-left:-20px}.controls>.radio:first-child,.controls>.checkbox:first-child{padding-top:5px}.radio.inline,.checkbox.inline{display:inline-block;padding-top:5px;margin-bottom:0;vertical-align:middle}.radio.inline+.radio.inline,.checkbox.inline+.checkbox.inline{margin-left:10px}.input-mini{width:60px}.input-small{width:90px}.input-medium{width:150px}.input-large{width:210px}.input-xlarge{width:270px}.input-xxlarge{width:530px}input[class*="span"],select[class*="span"],textarea[class*="span"],.uneditable-input[class*="span"],.row-fluid input[class*="span"],.row-fluid select[class*="span"],.row-fluid textarea[class*="span"],.row-fluid .uneditable-input[class*="span"]{float:none;margin-left:0}.input-append input[class*="span"],.input-append .uneditable-input[class*="span"],.input-prepend input[class*="span"],.input-prepend .uneditable-input[class*="span"],.row-fluid input[class*="span"],.row-fluid select[class*="span"],.row-fluid textarea[class*="span"],.row-fluid .uneditable-input[class*="span"],.row-fluid .input-prepend [class*="span"],.row-fluid .input-append [class*="span"]{display:inline-block}input,textarea,.uneditable-input{margin-left:0}.controls-row [class*="span"]+[class*="span"]{margin-left:20px}input.span12,textarea.span12,.uneditable-input.span12{width:926px}input.span11,textarea.span11,.uneditable-input.span11{width:846px}input.span10,textarea.span10,.uneditable-input.span10{width:766px}input.span9,textarea.span9,.uneditable-input.span9{width:686px}input.span8,textarea.span8,.uneditable-input.span8{width:606px}input.span7,textarea.span7,.uneditable-input.span7{width:526px}input.span6,textarea.span6,.uneditable-input.span6{width:446px}input.span5,textarea.span5,.uneditable-input.span5{width:366px}input.span4,textarea.span4,.uneditable-input.span4{width:286px}input.span3,textarea.span3,.uneditable-input.span3{width:206px}input.span2,textarea.span2,.uneditable-input.span2{width:126px}input.span1,textarea.span1,.uneditable-input.span1{width:46px}.controls-row{*zoom:1}.controls-row:before,.controls-row:after{display:table;line-height:0;content:""}.controls-row:after{clear:both}.controls-row [class*="span"],.row-fluid .controls-row [class*="span"]{float:left}.controls-row .checkbox[class*="span"],.controls-row .radio[class*="span"]{padding-top:5px}input[disabled],select[disabled],textarea[disabled],input[readonly],select[readonly],textarea[readonly]{cursor:not-allowed;background-color:#eee}input[type="radio"][disabled],input[type="checkbox"][disabled],input[type="radio"][readonly],input[type="checkbox"][readonly]{background-color:transparent}.control-group.warning .control-label,.control-group.warning .help-block,.control-group.warning .help-inline{color:#c09853}.control-group.warning .checkbox,.control-group.warning .radio,.control-group.warning input,.control-group.warning select,.control-group.warning textarea{color:#c09853}.control-group.warning input,.control-group.warning select,.control-group.warning textarea{border-color:#c09853;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);box-shadow:inset 0 1px 1px rgba(0,0,0,0.075)}.control-group.warning input:focus,.control-group.warning select:focus,.control-group.warning textarea:focus{border-color:#a47e3c;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #dbc59e;-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #dbc59e;box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #dbc59e}.control-group.warning .input-prepend .add-on,.control-group.warning .input-append .add-on{color:#c09853;background-color:#fcf8e3;border-color:#c09853}.control-group.error .control-label,.control-group.error .help-block,.control-group.error .help-inline{color:#b94a48}.control-group.error .checkbox,.control-group.error .radio,.control-group.error input,.control-group.error select,.control-group.error textarea{color:#b94a48}.control-group.error input,.control-group.error select,.control-group.error textarea{border-color:#b94a48;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);box-shadow:inset 0 1px 1px rgba(0,0,0,0.075)}.control-group.error input:focus,.control-group.error select:focus,.control-group.error textarea:focus{border-color:#953b39;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #d59392;-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #d59392;box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #d59392}.control-group.error .input-prepend .add-on,.control-group.error .input-append .add-on{color:#b94a48;background-color:#f2dede;border-color:#b94a48}.control-group.success .control-label,.control-group.success .help-block,.control-group.success .help-inline{color:#468847}.control-group.success .checkbox,.control-group.success .radio,.control-group.success input,.control-group.success select,.control-group.success textarea{color:#468847}.control-group.success input,.control-group.success select,.control-group.success textarea{border-color:#468847;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);box-shadow:inset 0 1px 1px rgba(0,0,0,0.075)}.control-group.success input:focus,.control-group.success select:focus,.control-group.success textarea:focus{border-color:#356635;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #7aba7b;-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #7aba7b;box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #7aba7b}.control-group.success .input-prepend .add-on,.control-group.success .input-append .add-on{color:#468847;background-color:#dff0d8;border-color:#468847}.control-group.info .control-label,.control-group.info .help-block,.control-group.info .help-inline{color:#3a87ad}.control-group.info .checkbox,.control-group.info .radio,.control-group.info input,.control-group.info select,.control-group.info textarea{color:#3a87ad}.control-group.info input,.control-group.info select,.control-group.info textarea{border-color:#3a87ad;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075);box-shadow:inset 0 1px 1px rgba(0,0,0,0.075)}.control-group.info input:focus,.control-group.info select:focus,.control-group.info textarea:focus{border-color:#2d6987;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #7ab5d3;-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #7ab5d3;box-shadow:inset 0 1px 1px rgba(0,0,0,0.075),0 0 6px #7ab5d3}.control-group.info .input-prepend .add-on,.control-group.info .input-append .add-on{color:#3a87ad;background-color:#d9edf7;border-color:#3a87ad}input:focus:invalid,textarea:focus:invalid,select:focus:invalid{color:#b94a48;border-color:#ee5f5b}input:focus:invalid:focus,textarea:focus:invalid:focus,select:focus:invalid:focus{border-color:#e9322d;-webkit-box-shadow:0 0 6px #f8b9b7;-moz-box-shadow:0 0 6px #f8b9b7;box-shadow:0 0 6px #f8b9b7}.form-actions{padding:19px 20px 20px;margin-top:20px;margin-bottom:20px;background-color:#f5f5f5;border-top:1px solid #e5e5e5;*zoom:1}.form-actions:before,.form-actions:after{display:table;line-height:0;content:""}.form-actions:after{clear:both}.help-block,.help-inline{color:#595959}.help-block{display:block;margin-bottom:10px}.help-inline{display:inline-block;*display:inline;padding-left:5px;vertical-align:middle;*zoom:1}.input-append,.input-prepend{display:inline-block;margin-bottom:10px;font-size:0;white-space:nowrap;vertical-align:middle}.input-append input,.input-prepend input,.input-append select,.input-prepend select,.input-append .uneditable-input,.input-prepend .uneditable-input,.input-append .dropdown-menu,.input-prepend .dropdown-menu,.input-append .popover,.input-prepend .popover{font-size:14px}.input-append input,.input-prepend input,.input-append select,.input-prepend select,.input-append .uneditable-input,.input-prepend .uneditable-input{position:relative;margin-bottom:0;*margin-left:0;vertical-align:top;-webkit-border-radius:0 4px 4px 0;-moz-border-radius:0 4px 4px 0;border-radius:0 4px 4px 0}.input-append input:focus,.input-prepend input:focus,.input-append select:focus,.input-prepend select:focus,.input-append .uneditable-input:focus,.input-prepend .uneditable-input:focus{z-index:2}.input-append .add-on,.input-prepend .add-on{display:inline-block;width:auto;height:20px;min-width:16px;padding:4px 5px;font-size:14px;font-weight:normal;line-height:20px;text-align:center;text-shadow:0 1px 0 #fff;background-color:#eee;border:1px solid #ccc}.input-append .add-on,.input-prepend .add-on,.input-append .btn,.input-prepend .btn,.input-append .btn-group>.dropdown-toggle,.input-prepend .btn-group>.dropdown-toggle{vertical-align:top;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.input-append .active,.input-prepend .active{background-color:#a9dba9;border-color:#46a546}.input-prepend .add-on,.input-prepend .btn{margin-right:-1px}.input-prepend .add-on:first-child,.input-prepend .btn:first-child{-webkit-border-radius:4px 0 0 4px;-moz-border-radius:4px 0 0 4px;border-radius:4px 0 0 4px}.input-append input,.input-append select,.input-append .uneditable-input{-webkit-border-radius:4px 0 0 4px;-moz-border-radius:4px 0 0 4px;border-radius:4px 0 0 4px}.input-append input+.btn-group .btn:last-child,.input-append select+.btn-group .btn:last-child,.input-append .uneditable-input+.btn-group .btn:last-child{-webkit-border-radius:0 4px 4px 0;-moz-border-radius:0 4px 4px 0;border-radius:0 4px 4px 0}.input-append .add-on,.input-append .btn,.input-append .btn-group{margin-left:-1px}.input-append .add-on:last-child,.input-append .btn:last-child,.input-append .btn-group:last-child>.dropdown-toggle{-webkit-border-radius:0 4px 4px 0;-moz-border-radius:0 4px 4px 0;border-radius:0 4px 4px 0}.input-prepend.input-append input,.input-prepend.input-append select,.input-prepend.input-append .uneditable-input{-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.input-prepend.input-append input+.btn-group .btn,.input-prepend.input-append select+.btn-group .btn,.input-prepend.input-append .uneditable-input+.btn-group .btn{-webkit-border-radius:0 4px 4px 0;-moz-border-radius:0 4px 4px 0;border-radius:0 4px 4px 0}.input-prepend.input-append .add-on:first-child,.input-prepend.input-append .btn:first-child{margin-right:-1px;-webkit-border-radius:4px 0 0 4px;-moz-border-radius:4px 0 0 4px;border-radius:4px 0 0 4px}.input-prepend.input-append .add-on:last-child,.input-prepend.input-append .btn:last-child{margin-left:-1px;-webkit-border-radius:0 4px 4px 0;-moz-border-radius:0 4px 4px 0;border-radius:0 4px 4px 0}.input-prepend.input-append .btn-group:first-child{margin-left:0}input.search-query{padding-right:14px;padding-right:4px \9;padding-left:14px;padding-left:4px \9;margin-bottom:0;-webkit-border-radius:15px;-moz-border-radius:15px;border-radius:15px}.form-search .input-append .search-query,.form-search .input-prepend .search-query{-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.form-search .input-append .search-query{-webkit-border-radius:14px 0 0 14px;-moz-border-radius:14px 0 0 14px;border-radius:14px 0 0 14px}.form-search .input-append .btn{-webkit-border-radius:0 14px 14px 0;-moz-border-radius:0 14px 14px 0;border-radius:0 14px 14px 0}.form-search .input-prepend .search-query{-webkit-border-radius:0 14px 14px 0;-moz-border-radius:0 14px 14px 0;border-radius:0 14px 14px 0}.form-search .input-prepend .btn{-webkit-border-radius:14px 0 0 14px;-moz-border-radius:14px 0 0 14px;border-radius:14px 0 0 14px}.form-search input,.form-inline input,.form-horizontal input,.form-search textarea,.form-inline textarea,.form-horizontal textarea,.form-search select,.form-inline select,.form-horizontal select,.form-search .help-inline,.form-inline .help-inline,.form-horizontal .help-inline,.form-search .uneditable-input,.form-inline .uneditable-input,.form-horizontal .uneditable-input,.form-search .input-prepend,.form-inline .input-prepend,.form-horizontal .input-prepend,.form-search .input-append,.form-inline .input-append,.form-horizontal .input-append{display:inline-block;*display:inline;margin-bottom:0;vertical-align:middle;*zoom:1}.form-search .hide,.form-inline .hide,.form-horizontal .hide{display:none}.form-search label,.form-inline label,.form-search .btn-group,.form-inline .btn-group{display:inline-block}.form-search .input-append,.form-inline .input-append,.form-search .input-prepend,.form-inline .input-prepend{margin-bottom:0}.form-search .radio,.form-search .checkbox,.form-inline .radio,.form-inline .checkbox{padding-left:0;margin-bottom:0;vertical-align:middle}.form-search .radio input[type="radio"],.form-search .checkbox input[type="checkbox"],.form-inline .radio input[type="radio"],.form-inline .checkbox input[type="checkbox"]{float:left;margin-right:3px;margin-left:0}.control-group{margin-bottom:10px}legend+.control-group{margin-top:20px;-webkit-margin-top-collapse:separate}.form-horizontal .control-group{margin-bottom:20px;*zoom:1}.form-horizontal .control-group:before,.form-horizontal .control-group:after{display:table;line-height:0;content:""}.form-horizontal .control-group:after{clear:both}.form-horizontal .control-label{float:left;width:160px;padding-top:5px;text-align:right}.form-horizontal .controls{*display:inline-block;*padding-left:20px;margin-left:180px;*margin-left:0}.form-horizontal .controls:first-child{*padding-left:180px}.form-horizontal .help-block{margin-bottom:0}.form-horizontal input+.help-block,.form-horizontal select+.help-block,.form-horizontal textarea+.help-block,.form-horizontal .uneditable-input+.help-block,.form-horizontal .input-prepend+.help-block,.form-horizontal .input-append+.help-block{margin-top:10px}.form-horizontal .form-actions{padding-left:180px}table{max-width:100%;background-color:transparent;border-collapse:collapse;border-spacing:0}.table{width:100%;margin-bottom:20px}.table th,.table td{padding:8px;line-height:20px;text-align:left;vertical-align:top;border-top:1px solid #ddd}.table th{font-weight:bold}.table thead th{vertical-align:bottom}.table caption+thead tr:first-child th,.table caption+thead tr:first-child td,.table colgroup+thead tr:first-child th,.table colgroup+thead tr:first-child td,.table thead:first-child tr:first-child th,.table thead:first-child tr:first-child td{border-top:0}.table tbody+tbody{border-top:2px solid #ddd}.table .table{background-color:#fff}.table-condensed th,.table-condensed td{padding:4px 5px}.table-bordered{border:1px solid #ddd;border-collapse:separate;*border-collapse:collapse;border-left:0;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.table-bordered th,.table-bordered td{border-left:1px solid #ddd}.table-bordered caption+thead tr:first-child th,.table-bordered caption+tbody tr:first-child th,.table-bordered caption+tbody tr:first-child td,.table-bordered colgroup+thead tr:first-child th,.table-bordered colgroup+tbody tr:first-child th,.table-bordered colgroup+tbody tr:first-child td,.table-bordered thead:first-child tr:first-child th,.table-bordered tbody:first-child tr:first-child th,.table-bordered tbody:first-child tr:first-child td{border-top:0}.table-bordered thead:first-child tr:first-child>th:first-child,.table-bordered tbody:first-child tr:first-child>td:first-child,.table-bordered tbody:first-child tr:first-child>th:first-child{-webkit-border-top-left-radius:4px;border-top-left-radius:4px;-moz-border-radius-topleft:4px}.table-bordered thead:first-child tr:first-child>th:last-child,.table-bordered tbody:first-child tr:first-child>td:last-child,.table-bordered tbody:first-child tr:first-child>th:last-child{-webkit-border-top-right-radius:4px;border-top-right-radius:4px;-moz-border-radius-topright:4px}.table-bordered thead:last-child tr:last-child>th:first-child,.table-bordered tbody:last-child tr:last-child>td:first-child,.table-bordered tbody:last-child tr:last-child>th:first-child,.table-bordered tfoot:last-child tr:last-child>td:first-child,.table-bordered tfoot:last-child tr:last-child>th:first-child{-webkit-border-bottom-left-radius:4px;border-bottom-left-radius:4px;-moz-border-radius-bottomleft:4px}.table-bordered thead:last-child tr:last-child>th:last-child,.table-bordered tbody:last-child tr:last-child>td:last-child,.table-bordered tbody:last-child tr:last-child>th:last-child,.table-bordered tfoot:last-child tr:last-child>td:last-child,.table-bordered tfoot:last-child tr:last-child>th:last-child{-webkit-border-bottom-right-radius:4px;border-bottom-right-radius:4px;-moz-border-radius-bottomright:4px}.table-bordered tfoot+tbody:last-child tr:last-child td:first-child{-webkit-border-bottom-left-radius:0;border-bottom-left-radius:0;-moz-border-radius-bottomleft:0}.table-bordered tfoot+tbody:last-child tr:last-child td:last-child{-webkit-border-bottom-right-radius:0;border-bottom-right-radius:0;-moz-border-radius-bottomright:0}.table-bordered caption+thead tr:first-child th:first-child,.table-bordered caption+tbody tr:first-child td:first-child,.table-bordered colgroup+thead tr:first-child th:first-child,.table-bordered colgroup+tbody tr:first-child td:first-child{-webkit-border-top-left-radius:4px;border-top-left-radius:4px;-moz-border-radius-topleft:4px}.table-bordered caption+thead tr:first-child th:last-child,.table-bordered caption+tbody tr:first-child td:last-child,.table-bordered colgroup+thead tr:first-child th:last-child,.table-bordered colgroup+tbody tr:first-child td:last-child{-webkit-border-top-right-radius:4px;border-top-right-radius:4px;-moz-border-radius-topright:4px}.table-striped tbody>tr:nth-child(odd)>td,.table-striped tbody>tr:nth-child(odd)>th{background-color:#f9f9f9}.table-hover tbody tr:hover>td,.table-hover tbody tr:hover>th{background-color:#f5f5f5}table td[class*="span"],table th[class*="span"],.row-fluid table td[class*="span"],.row-fluid table th[class*="span"]{display:table-cell;float:none;margin-left:0}.table td.span1,.table th.span1{float:none;width:44px;margin-left:0}.table td.span2,.table th.span2{float:none;width:124px;margin-left:0}.table td.span3,.table th.span3{float:none;width:204px;margin-left:0}.table td.span4,.table th.span4{float:none;width:284px;margin-left:0}.table td.span5,.table th.span5{float:none;width:364px;margin-left:0}.table td.span6,.table th.span6{float:none;width:444px;margin-left:0}.table td.span7,.table th.span7{float:none;width:524px;margin-left:0}.table td.span8,.table th.span8{float:none;width:604px;margin-left:0}.table td.span9,.table th.span9{float:none;width:684px;margin-left:0}.table td.span10,.table th.span10{float:none;width:764px;margin-left:0}.table td.span11,.table th.span11{float:none;width:844px;margin-left:0}.table td.span12,.table th.span12{float:none;width:924px;margin-left:0}.table tbody tr.success>td{background-color:#dff0d8}.table tbody tr.error>td{background-color:#f2dede}.table tbody tr.warning>td{background-color:#fcf8e3}.table tbody tr.info>td{background-color:#d9edf7}.table-hover tbody tr.success:hover>td{background-color:#d0e9c6}.table-hover tbody tr.error:hover>td{background-color:#ebcccc}.table-hover tbody tr.warning:hover>td{background-color:#faf2cc}.table-hover tbody tr.info:hover>td{background-color:#c4e3f3}[class^="icon-"],[class*=" icon-"]{display:inline-block;width:14px;height:14px;margin-top:1px;*margin-right:.3em;line-height:14px;vertical-align:text-top;background-image:url("../img/glyphicons-halflings.png");background-position:14px 14px;background-repeat:no-repeat}.icon-white,.nav-pills>.active>a>[class^="icon-"],.nav-pills>.active>a>[class*=" icon-"],.nav-list>.active>a>[class^="icon-"],.nav-list>.active>a>[class*=" icon-"],.navbar-inverse .nav>.active>a>[class^="icon-"],.navbar-inverse .nav>.active>a>[class*=" icon-"],.dropdown-menu>li>a:hover>[class^="icon-"],.dropdown-menu>li>a:focus>[class^="icon-"],.dropdown-menu>li>a:hover>[class*=" icon-"],.dropdown-menu>li>a:focus>[class*=" icon-"],.dropdown-menu>.active>a>[class^="icon-"],.dropdown-menu>.active>a>[class*=" icon-"],.dropdown-submenu:hover>a>[class^="icon-"],.dropdown-submenu:focus>a>[class^="icon-"],.dropdown-submenu:hover>a>[class*=" icon-"],.dropdown-submenu:focus>a>[class*=" icon-"]{background-image:url("../img/glyphicons-halflings-white.png")}.icon-glass{background-position:0 0}.icon-music{background-position:-24px 0}.icon-search{background-position:-48px 0}.icon-envelope{background-position:-72px 0}.icon-heart{background-position:-96px 0}.icon-star{background-position:-120px 0}.icon-star-empty{background-position:-144px 0}.icon-user{background-position:-168px 0}.icon-film{background-position:-192px 0}.icon-th-large{background-position:-216px 0}.icon-th{background-position:-240px 0}.icon-th-list{background-position:-264px 0}.icon-ok{background-position:-288px 0}.icon-remove{background-position:-312px 0}.icon-zoom-in{background-position:-336px 0}.icon-zoom-out{background-position:-360px 0}.icon-off{background-position:-384px 0}.icon-signal{background-position:-408px 0}.icon-cog{background-position:-432px 0}.icon-trash{background-position:-456px 0}.icon-home{background-position:0 -24px}.icon-file{background-position:-24px -24px}.icon-time{background-position:-48px -24px}.icon-road{background-position:-72px -24px}.icon-download-alt{background-position:-96px -24px}.icon-download{background-position:-120px -24px}.icon-upload{background-position:-144px -24px}.icon-inbox{background-position:-168px -24px}.icon-play-circle{background-position:-192px -24px}.icon-repeat{background-position:-216px -24px}.icon-refresh{background-position:-240px -24px}.icon-list-alt{background-position:-264px -24px}.icon-lock{background-position:-287px -24px}.icon-flag{background-position:-312px -24px}.icon-headphones{background-position:-336px -24px}.icon-volume-off{background-position:-360px -24px}.icon-volume-down{background-position:-384px -24px}.icon-volume-up{background-position:-408px -24px}.icon-qrcode{background-position:-432px -24px}.icon-barcode{background-position:-456px -24px}.icon-tag{background-position:0 -48px}.icon-tags{background-position:-25px -48px}.icon-book{background-position:-48px -48px}.icon-bookmark{background-position:-72px -48px}.icon-print{background-position:-96px -48px}.icon-camera{background-position:-120px -48px}.icon-font{background-position:-144px -48px}.icon-bold{background-position:-167px -48px}.icon-italic{background-position:-192px -48px}.icon-text-height{background-position:-216px -48px}.icon-text-width{background-position:-240px -48px}.icon-align-left{background-position:-264px -48px}.icon-align-center{background-position:-288px -48px}.icon-align-right{background-position:-312px -48px}.icon-align-justify{background-position:-336px -48px}.icon-list{background-position:-360px -48px}.icon-indent-left{background-position:-384px -48px}.icon-indent-right{background-position:-408px -48px}.icon-facetime-video{background-position:-432px -48px}.icon-picture{background-position:-456px -48px}.icon-pencil{background-position:0 -72px}.icon-map-marker{background-position:-24px -72px}.icon-adjust{background-position:-48px -72px}.icon-tint{background-position:-72px -72px}.icon-edit{background-position:-96px -72px}.icon-share{background-position:-120px -72px}.icon-check{background-position:-144px -72px}.icon-move{background-position:-168px -72px}.icon-step-backward{background-position:-192px -72px}.icon-fast-backward{background-position:-216px -72px}.icon-backward{background-position:-240px -72px}.icon-play{background-position:-264px -72px}.icon-pause{background-position:-288px -72px}.icon-stop{background-position:-312px -72px}.icon-forward{background-position:-336px -72px}.icon-fast-forward{background-position:-360px -72px}.icon-step-forward{background-position:-384px -72px}.icon-eject{background-position:-408px -72px}.icon-chevron-left{background-position:-432px -72px}.icon-chevron-right{background-position:-456px -72px}.icon-plus-sign{background-position:0 -96px}.icon-minus-sign{background-position:-24px -96px}.icon-remove-sign{background-position:-48px -96px}.icon-ok-sign{background-position:-72px -96px}.icon-question-sign{background-position:-96px -96px}.icon-info-sign{background-position:-120px -96px}.icon-screenshot{background-position:-144px -96px}.icon-remove-circle{background-position:-168px -96px}.icon-ok-circle{background-position:-192px -96px}.icon-ban-circle{background-position:-216px -96px}.icon-arrow-left{background-position:-240px -96px}.icon-arrow-right{background-position:-264px -96px}.icon-arrow-up{background-position:-289px -96px}.icon-arrow-down{background-position:-312px -96px}.icon-share-alt{background-position:-336px -96px}.icon-resize-full{background-position:-360px -96px}.icon-resize-small{background-position:-384px -96px}.icon-plus{background-position:-408px -96px}.icon-minus{background-position:-433px -96px}.icon-asterisk{background-position:-456px -96px}.icon-exclamation-sign{background-position:0 -120px}.icon-gift{background-position:-24px -120px}.icon-leaf{background-position:-48px -120px}.icon-fire{background-position:-72px -120px}.icon-eye-open{background-position:-96px -120px}.icon-eye-close{background-position:-120px -120px}.icon-warning-sign{background-position:-144px -120px}.icon-plane{background-position:-168px -120px}.icon-calendar{background-position:-192px -120px}.icon-random{width:16px;background-position:-216px -120px}.icon-comment{background-position:-240px -120px}.icon-magnet{background-position:-264px -120px}.icon-chevron-up{background-position:-288px -120px}.icon-chevron-down{background-position:-313px -119px}.icon-retweet{background-position:-336px -120px}.icon-shopping-cart{background-position:-360px -120px}.icon-folder-close{width:16px;background-position:-384px -120px}.icon-folder-open{width:16px;background-position:-408px -120px}.icon-resize-vertical{background-position:-432px -119px}.icon-resize-horizontal{background-position:-456px -118px}.icon-hdd{background-position:0 -144px}.icon-bullhorn{background-position:-24px -144px}.icon-bell{background-position:-48px -144px}.icon-certificate{background-position:-72px -144px}.icon-thumbs-up{background-position:-96px -144px}.icon-thumbs-down{background-position:-120px -144px}.icon-hand-right{background-position:-144px -144px}.icon-hand-left{background-position:-168px -144px}.icon-hand-up{background-position:-192px -144px}.icon-hand-down{background-position:-216px -144px}.icon-circle-arrow-right{background-position:-240px -144px}.icon-circle-arrow-left{background-position:-264px -144px}.icon-circle-arrow-up{background-position:-288px -144px}.icon-circle-arrow-down{background-position:-312px -144px}.icon-globe{background-position:-336px -144px}.icon-wrench{background-position:-360px -144px}.icon-tasks{background-position:-384px -144px}.icon-filter{background-position:-408px -144px}.icon-briefcase{background-position:-432px -144px}.icon-fullscreen{background-position:-456px -144px}.dropup,.dropdown{position:relative}.dropdown-toggle{*margin-bottom:-3px}.dropdown-toggle:active,.open .dropdown-toggle{outline:0}.caret{display:inline-block;width:0;height:0;vertical-align:top;border-top:4px solid #000;border-right:4px solid transparent;border-left:4px solid transparent;content:""}.dropdown .caret{margin-top:8px;margin-left:2px}.dropdown-menu{position:absolute;top:100%;left:0;z-index:1000;display:none;float:left;min-width:160px;padding:5px 0;margin:2px 0 0;list-style:none;background-color:#fff;border:1px solid #ccc;border:1px solid rgba(0,0,0,0.2);*border-right-width:2px;*border-bottom-width:2px;-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px;-webkit-box-shadow:0 5px 10px rgba(0,0,0,0.2);-moz-box-shadow:0 5px 10px rgba(0,0,0,0.2);box-shadow:0 5px 10px rgba(0,0,0,0.2);-webkit-background-clip:padding-box;-moz-background-clip:padding;background-clip:padding-box}.dropdown-menu.pull-right{right:0;left:auto}.dropdown-menu .divider{*width:100%;height:1px;margin:9px 1px;*margin:-5px 0 5px;overflow:hidden;background-color:#e5e5e5;border-bottom:1px solid #fff}.dropdown-menu>li>a{display:block;padding:3px 20px;clear:both;font-weight:normal;line-height:20px;color:#333;white-space:nowrap}.dropdown-menu>li>a:hover,.dropdown-menu>li>a:focus,.dropdown-submenu:hover>a,.dropdown-submenu:focus>a{color:#fff;text-decoration:none;background-color:#0081c2;background-image:-moz-linear-gradient(top,#08c,#0077b3);background-image:-webkit-gradient(linear,0 0,0 100%,from(#08c),to(#0077b3));background-image:-webkit-linear-gradient(top,#08c,#0077b3);background-image:-o-linear-gradient(top,#08c,#0077b3);background-image:linear-gradient(to bottom,#08c,#0077b3);background-repeat:repeat-x;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff0088cc',endColorstr='#ff0077b3',GradientType=0)}.dropdown-menu>.active>a,.dropdown-menu>.active>a:hover,.dropdown-menu>.active>a:focus{color:#fff;text-decoration:none;background-color:#0081c2;background-image:-moz-linear-gradient(top,#08c,#0077b3);background-image:-webkit-gradient(linear,0 0,0 100%,from(#08c),to(#0077b3));background-image:-webkit-linear-gradient(top,#08c,#0077b3);background-image:-o-linear-gradient(top,#08c,#0077b3);background-image:linear-gradient(to bottom,#08c,#0077b3);background-repeat:repeat-x;outline:0;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff0088cc',endColorstr='#ff0077b3',GradientType=0)}.dropdown-menu>.disabled>a,.dropdown-menu>.disabled>a:hover,.dropdown-menu>.disabled>a:focus{color:#999}.dropdown-menu>.disabled>a:hover,.dropdown-menu>.disabled>a:focus{text-decoration:none;cursor:default;background-color:transparent;background-image:none;filter:progid:DXImageTransform.Microsoft.gradient(enabled=false)}.open{*z-index:1000}.open>.dropdown-menu{display:block}.dropdown-backdrop{position:fixed;top:0;right:0;bottom:0;left:0;z-index:990}.pull-right>.dropdown-menu{right:0;left:auto}.dropup .caret,.navbar-fixed-bottom .dropdown .caret{border-top:0;border-bottom:4px solid #000;content:""}.dropup .dropdown-menu,.navbar-fixed-bottom .dropdown .dropdown-menu{top:auto;bottom:100%;margin-bottom:1px}.dropdown-submenu{position:relative}.dropdown-submenu>.dropdown-menu{top:0;left:100%;margin-top:-6px;margin-left:-1px;-webkit-border-radius:0 6px 6px 6px;-moz-border-radius:0 6px 6px 6px;border-radius:0 6px 6px 6px}.dropdown-submenu:hover>.dropdown-menu{display:block}.dropup .dropdown-submenu>.dropdown-menu{top:auto;bottom:0;margin-top:0;margin-bottom:-2px;-webkit-border-radius:5px 5px 5px 0;-moz-border-radius:5px 5px 5px 0;border-radius:5px 5px 5px 0}.dropdown-submenu>a:after{display:block;float:right;width:0;height:0;margin-top:5px;margin-right:-10px;border-color:transparent;border-left-color:#ccc;border-style:solid;border-width:5px 0 5px 5px;content:" "}.dropdown-submenu:hover>a:after{border-left-color:#fff}.dropdown-submenu.pull-left{float:none}.dropdown-submenu.pull-left>.dropdown-menu{left:-100%;margin-left:10px;-webkit-border-radius:6px 0 6px 6px;-moz-border-radius:6px 0 6px 6px;border-radius:6px 0 6px 6px}.dropdown .dropdown-menu .nav-header{padding-right:20px;padding-left:20px}.typeahead{z-index:1051;margin-top:2px;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.well{min-height:20px;padding:19px;margin-bottom:20px;background-color:#f5f5f5;border:1px solid #e3e3e3;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;-webkit-box-shadow:inset 0 1px 1px rgba(0,0,0,0.05);-moz-box-shadow:inset 0 1px 1px rgba(0,0,0,0.05);box-shadow:inset 0 1px 1px rgba(0,0,0,0.05)}.well blockquote{border-color:#ddd;border-color:rgba(0,0,0,0.15)}.well-large{padding:24px;-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px}.well-small{padding:9px;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}.fade{opacity:0;-webkit-transition:opacity .15s linear;-moz-transition:opacity .15s linear;-o-transition:opacity .15s linear;transition:opacity .15s linear}.fade.in{opacity:1}.collapse{position:relative;height:0;overflow:hidden;-webkit-transition:height .35s ease;-moz-transition:height .35s ease;-o-transition:height .35s ease;transition:height .35s ease}.collapse.in{height:auto}.close{float:right;font-size:20px;font-weight:bold;line-height:20px;color:#000;text-shadow:0 1px 0 #fff;opacity:.2;filter:alpha(opacity=20)}.close:hover,.close:focus{color:#000;text-decoration:none;cursor:pointer;opacity:.4;filter:alpha(opacity=40)}button.close{padding:0;cursor:pointer;background:transparent;border:0;-webkit-appearance:none}.btn{display:inline-block;*display:inline;padding:4px 12px;margin-bottom:0;*margin-left:.3em;font-size:14px;line-height:20px;color:#333;text-align:center;text-shadow:0 1px 1px rgba(255,255,255,0.75);vertical-align:middle;cursor:pointer;background-color:#f5f5f5;*background-color:#e6e6e6;background-image:-moz-linear-gradient(top,#fff,#e6e6e6);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fff),to(#e6e6e6));background-image:-webkit-linear-gradient(top,#fff,#e6e6e6);background-image:-o-linear-gradient(top,#fff,#e6e6e6);background-image:linear-gradient(to bottom,#fff,#e6e6e6);background-repeat:repeat-x;border:1px solid #ccc;*border:0;border-color:#e6e6e6 #e6e6e6 #bfbfbf;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);border-bottom-color:#b3b3b3;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ffffffff',endColorstr='#ffe6e6e6',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false);*zoom:1;-webkit-box-shadow:inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05);box-shadow:inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05)}.btn:hover,.btn:focus,.btn:active,.btn.active,.btn.disabled,.btn[disabled]{color:#333;background-color:#e6e6e6;*background-color:#d9d9d9}.btn:active,.btn.active{background-color:#ccc \9}.btn:first-child{*margin-left:0}.btn:hover,.btn:focus{color:#333;text-decoration:none;background-position:0 -15px;-webkit-transition:background-position .1s linear;-moz-transition:background-position .1s linear;-o-transition:background-position .1s linear;transition:background-position .1s linear}.btn:focus{outline:thin dotted #333;outline:5px auto -webkit-focus-ring-color;outline-offset:-2px}.btn.active,.btn:active{background-image:none;outline:0;-webkit-box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05);box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05)}.btn.disabled,.btn[disabled]{cursor:default;background-image:none;opacity:.65;filter:alpha(opacity=65);-webkit-box-shadow:none;-moz-box-shadow:none;box-shadow:none}.btn-large{padding:11px 19px;font-size:17.5px;-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px}.btn-large [class^="icon-"],.btn-large [class*=" icon-"]{margin-top:4px}.btn-small{padding:2px 10px;font-size:11.9px;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}.btn-small [class^="icon-"],.btn-small [class*=" icon-"]{margin-top:0}.btn-mini [class^="icon-"],.btn-mini [class*=" icon-"]{margin-top:-1px}.btn-mini{padding:0 6px;font-size:10.5px;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}.btn-block{display:block;width:100%;padding-right:0;padding-left:0;-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box}.btn-block+.btn-block{margin-top:5px}input[type="submit"].btn-block,input[type="reset"].btn-block,input[type="button"].btn-block{width:100%}.btn-primary.active,.btn-warning.active,.btn-danger.active,.btn-success.active,.btn-info.active,.btn-inverse.active{color:rgba(255,255,255,0.75)}.btn-primary{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#006dcc;*background-color:#04c;background-image:-moz-linear-gradient(top,#08c,#04c);background-image:-webkit-gradient(linear,0 0,0 100%,from(#08c),to(#04c));background-image:-webkit-linear-gradient(top,#08c,#04c);background-image:-o-linear-gradient(top,#08c,#04c);background-image:linear-gradient(to bottom,#08c,#04c);background-repeat:repeat-x;border-color:#04c #04c #002a80;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff0088cc',endColorstr='#ff0044cc',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false)}.btn-primary:hover,.btn-primary:focus,.btn-primary:active,.btn-primary.active,.btn-primary.disabled,.btn-primary[disabled]{color:#fff;background-color:#04c;*background-color:#003bb3}.btn-primary:active,.btn-primary.active{background-color:#039 \9}.btn-warning{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#faa732;*background-color:#f89406;background-image:-moz-linear-gradient(top,#fbb450,#f89406);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fbb450),to(#f89406));background-image:-webkit-linear-gradient(top,#fbb450,#f89406);background-image:-o-linear-gradient(top,#fbb450,#f89406);background-image:linear-gradient(to bottom,#fbb450,#f89406);background-repeat:repeat-x;border-color:#f89406 #f89406 #ad6704;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#fffbb450',endColorstr='#fff89406',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false)}.btn-warning:hover,.btn-warning:focus,.btn-warning:active,.btn-warning.active,.btn-warning.disabled,.btn-warning[disabled]{color:#fff;background-color:#f89406;*background-color:#df8505}.btn-warning:active,.btn-warning.active{background-color:#c67605 \9}.btn-danger{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#da4f49;*background-color:#bd362f;background-image:-moz-linear-gradient(top,#ee5f5b,#bd362f);background-image:-webkit-gradient(linear,0 0,0 100%,from(#ee5f5b),to(#bd362f));background-image:-webkit-linear-gradient(top,#ee5f5b,#bd362f);background-image:-o-linear-gradient(top,#ee5f5b,#bd362f);background-image:linear-gradient(to bottom,#ee5f5b,#bd362f);background-repeat:repeat-x;border-color:#bd362f #bd362f #802420;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ffee5f5b',endColorstr='#ffbd362f',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false)}.btn-danger:hover,.btn-danger:focus,.btn-danger:active,.btn-danger.active,.btn-danger.disabled,.btn-danger[disabled]{color:#fff;background-color:#bd362f;*background-color:#a9302a}.btn-danger:active,.btn-danger.active{background-color:#942a25 \9}.btn-success{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#5bb75b;*background-color:#51a351;background-image:-moz-linear-gradient(top,#62c462,#51a351);background-image:-webkit-gradient(linear,0 0,0 100%,from(#62c462),to(#51a351));background-image:-webkit-linear-gradient(top,#62c462,#51a351);background-image:-o-linear-gradient(top,#62c462,#51a351);background-image:linear-gradient(to bottom,#62c462,#51a351);background-repeat:repeat-x;border-color:#51a351 #51a351 #387038;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff62c462',endColorstr='#ff51a351',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false)}.btn-success:hover,.btn-success:focus,.btn-success:active,.btn-success.active,.btn-success.disabled,.btn-success[disabled]{color:#fff;background-color:#51a351;*background-color:#499249}.btn-success:active,.btn-success.active{background-color:#408140 \9}.btn-info{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#49afcd;*background-color:#2f96b4;background-image:-moz-linear-gradient(top,#5bc0de,#2f96b4);background-image:-webkit-gradient(linear,0 0,0 100%,from(#5bc0de),to(#2f96b4));background-image:-webkit-linear-gradient(top,#5bc0de,#2f96b4);background-image:-o-linear-gradient(top,#5bc0de,#2f96b4);background-image:linear-gradient(to bottom,#5bc0de,#2f96b4);background-repeat:repeat-x;border-color:#2f96b4 #2f96b4 #1f6377;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff5bc0de',endColorstr='#ff2f96b4',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false)}.btn-info:hover,.btn-info:focus,.btn-info:active,.btn-info.active,.btn-info.disabled,.btn-info[disabled]{color:#fff;background-color:#2f96b4;*background-color:#2a85a0}.btn-info:active,.btn-info.active{background-color:#24748c \9}.btn-inverse{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#363636;*background-color:#222;background-image:-moz-linear-gradient(top,#444,#222);background-image:-webkit-gradient(linear,0 0,0 100%,from(#444),to(#222));background-image:-webkit-linear-gradient(top,#444,#222);background-image:-o-linear-gradient(top,#444,#222);background-image:linear-gradient(to bottom,#444,#222);background-repeat:repeat-x;border-color:#222 #222 #000;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff444444',endColorstr='#ff222222',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false)}.btn-inverse:hover,.btn-inverse:focus,.btn-inverse:active,.btn-inverse.active,.btn-inverse.disabled,.btn-inverse[disabled]{color:#fff;background-color:#222;*background-color:#151515}.btn-inverse:active,.btn-inverse.active{background-color:#080808 \9}button.btn,input[type="submit"].btn{*padding-top:3px;*padding-bottom:3px}button.btn::-moz-focus-inner,input[type="submit"].btn::-moz-focus-inner{padding:0;border:0}button.btn.btn-large,input[type="submit"].btn.btn-large{*padding-top:7px;*padding-bottom:7px}button.btn.btn-small,input[type="submit"].btn.btn-small{*padding-top:3px;*padding-bottom:3px}button.btn.btn-mini,input[type="submit"].btn.btn-mini{*padding-top:1px;*padding-bottom:1px}.btn-link,.btn-link:active,.btn-link[disabled]{background-color:transparent;background-image:none;-webkit-box-shadow:none;-moz-box-shadow:none;box-shadow:none}.btn-link{color:#08c;cursor:pointer;border-color:transparent;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.btn-link:hover,.btn-link:focus{color:#005580;text-decoration:underline;background-color:transparent}.btn-link[disabled]:hover,.btn-link[disabled]:focus{color:#333;text-decoration:none}.btn-group{position:relative;display:inline-block;*display:inline;*margin-left:.3em;font-size:0;white-space:nowrap;vertical-align:middle;*zoom:1}.btn-group:first-child{*margin-left:0}.btn-group+.btn-group{margin-left:5px}.btn-toolbar{margin-top:10px;margin-bottom:10px;font-size:0}.btn-toolbar>.btn+.btn,.btn-toolbar>.btn-group+.btn,.btn-toolbar>.btn+.btn-group{margin-left:5px}.btn-group>.btn{position:relative;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.btn-group>.btn+.btn{margin-left:-1px}.btn-group>.btn,.btn-group>.dropdown-menu,.btn-group>.popover{font-size:14px}.btn-group>.btn-mini{font-size:10.5px}.btn-group>.btn-small{font-size:11.9px}.btn-group>.btn-large{font-size:17.5px}.btn-group>.btn:first-child{margin-left:0;-webkit-border-bottom-left-radius:4px;border-bottom-left-radius:4px;-webkit-border-top-left-radius:4px;border-top-left-radius:4px;-moz-border-radius-bottomleft:4px;-moz-border-radius-topleft:4px}.btn-group>.btn:last-child,.btn-group>.dropdown-toggle{-webkit-border-top-right-radius:4px;border-top-right-radius:4px;-webkit-border-bottom-right-radius:4px;border-bottom-right-radius:4px;-moz-border-radius-topright:4px;-moz-border-radius-bottomright:4px}.btn-group>.btn.large:first-child{margin-left:0;-webkit-border-bottom-left-radius:6px;border-bottom-left-radius:6px;-webkit-border-top-left-radius:6px;border-top-left-radius:6px;-moz-border-radius-bottomleft:6px;-moz-border-radius-topleft:6px}.btn-group>.btn.large:last-child,.btn-group>.large.dropdown-toggle{-webkit-border-top-right-radius:6px;border-top-right-radius:6px;-webkit-border-bottom-right-radius:6px;border-bottom-right-radius:6px;-moz-border-radius-topright:6px;-moz-border-radius-bottomright:6px}.btn-group>.btn:hover,.btn-group>.btn:focus,.btn-group>.btn:active,.btn-group>.btn.active{z-index:2}.btn-group .dropdown-toggle:active,.btn-group.open .dropdown-toggle{outline:0}.btn-group>.btn+.dropdown-toggle{*padding-top:5px;padding-right:8px;*padding-bottom:5px;padding-left:8px;-webkit-box-shadow:inset 1px 0 0 rgba(255,255,255,0.125),inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:inset 1px 0 0 rgba(255,255,255,0.125),inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05);box-shadow:inset 1px 0 0 rgba(255,255,255,0.125),inset 0 1px 0 rgba(255,255,255,0.2),0 1px 2px rgba(0,0,0,0.05)}.btn-group>.btn-mini+.dropdown-toggle{*padding-top:2px;padding-right:5px;*padding-bottom:2px;padding-left:5px}.btn-group>.btn-small+.dropdown-toggle{*padding-top:5px;*padding-bottom:4px}.btn-group>.btn-large+.dropdown-toggle{*padding-top:7px;padding-right:12px;*padding-bottom:7px;padding-left:12px}.btn-group.open .dropdown-toggle{background-image:none;-webkit-box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05);box-shadow:inset 0 2px 4px rgba(0,0,0,0.15),0 1px 2px rgba(0,0,0,0.05)}.btn-group.open .btn.dropdown-toggle{background-color:#e6e6e6}.btn-group.open .btn-primary.dropdown-toggle{background-color:#04c}.btn-group.open .btn-warning.dropdown-toggle{background-color:#f89406}.btn-group.open .btn-danger.dropdown-toggle{background-color:#bd362f}.btn-group.open .btn-success.dropdown-toggle{background-color:#51a351}.btn-group.open .btn-info.dropdown-toggle{background-color:#2f96b4}.btn-group.open .btn-inverse.dropdown-toggle{background-color:#222}.btn .caret{margin-top:8px;margin-left:0}.btn-large .caret{margin-top:6px}.btn-large .caret{border-top-width:5px;border-right-width:5px;border-left-width:5px}.btn-mini .caret,.btn-small .caret{margin-top:8px}.dropup .btn-large .caret{border-bottom-width:5px}.btn-primary .caret,.btn-warning .caret,.btn-danger .caret,.btn-info .caret,.btn-success .caret,.btn-inverse .caret{border-top-color:#fff;border-bottom-color:#fff}.btn-group-vertical{display:inline-block;*display:inline;*zoom:1}.btn-group-vertical>.btn{display:block;float:none;max-width:100%;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.btn-group-vertical>.btn+.btn{margin-top:-1px;margin-left:0}.btn-group-vertical>.btn:first-child{-webkit-border-radius:4px 4px 0 0;-moz-border-radius:4px 4px 0 0;border-radius:4px 4px 0 0}.btn-group-vertical>.btn:last-child{-webkit-border-radius:0 0 4px 4px;-moz-border-radius:0 0 4px 4px;border-radius:0 0 4px 4px}.btn-group-vertical>.btn-large:first-child{-webkit-border-radius:6px 6px 0 0;-moz-border-radius:6px 6px 0 0;border-radius:6px 6px 0 0}.btn-group-vertical>.btn-large:last-child{-webkit-border-radius:0 0 6px 6px;-moz-border-radius:0 0 6px 6px;border-radius:0 0 6px 6px}.alert{padding:8px 35px 8px 14px;margin-bottom:20px;text-shadow:0 1px 0 rgba(255,255,255,0.5);background-color:#fcf8e3;border:1px solid #fbeed5;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.alert,.alert h4{color:#c09853}.alert h4{margin:0}.alert .close{position:relative;top:-2px;right:-21px;line-height:20px}.alert-success{color:#468847;background-color:#dff0d8;border-color:#d6e9c6}.alert-success h4{color:#468847}.alert-danger,.alert-error{color:#b94a48;background-color:#f2dede;border-color:#eed3d7}.alert-danger h4,.alert-error h4{color:#b94a48}.alert-info{color:#3a87ad;background-color:#d9edf7;border-color:#bce8f1}.alert-info h4{color:#3a87ad}.alert-block{padding-top:14px;padding-bottom:14px}.alert-block>p,.alert-block>ul{margin-bottom:0}.alert-block p+p{margin-top:5px}.nav{margin-bottom:20px;margin-left:0;list-style:none}.nav>li>a{display:block}.nav>li>a:hover,.nav>li>a:focus{text-decoration:none;background-color:#eee}.nav>li>a>img{max-width:none}.nav>.pull-right{float:right}.nav-header{display:block;padding:3px 15px;font-size:11px;font-weight:bold;line-height:20px;color:#999;text-shadow:0 1px 0 rgba(255,255,255,0.5);text-transform:uppercase}.nav li+.nav-header{margin-top:9px}.nav-list{padding-right:15px;padding-left:15px;margin-bottom:0}.nav-list>li>a,.nav-list .nav-header{margin-right:-15px;margin-left:-15px;text-shadow:0 1px 0 rgba(255,255,255,0.5)}.nav-list>li>a{padding:3px 15px}.nav-list>.active>a,.nav-list>.active>a:hover,.nav-list>.active>a:focus{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.2);background-color:#08c}.nav-list [class^="icon-"],.nav-list [class*=" icon-"]{margin-right:2px}.nav-list .divider{*width:100%;height:1px;margin:9px 1px;*margin:-5px 0 5px;overflow:hidden;background-color:#e5e5e5;border-bottom:1px solid #fff}.nav-tabs,.nav-pills{*zoom:1}.nav-tabs:before,.nav-pills:before,.nav-tabs:after,.nav-pills:after{display:table;line-height:0;content:""}.nav-tabs:after,.nav-pills:after{clear:both}.nav-tabs>li,.nav-pills>li{float:left}.nav-tabs>li>a,.nav-pills>li>a{padding-right:12px;padding-left:12px;margin-right:2px;line-height:14px}.nav-tabs{border-bottom:1px solid #ddd}.nav-tabs>li{margin-bottom:-1px}.nav-tabs>li>a{padding-top:8px;padding-bottom:8px;line-height:20px;border:1px solid transparent;-webkit-border-radius:4px 4px 0 0;-moz-border-radius:4px 4px 0 0;border-radius:4px 4px 0 0}.nav-tabs>li>a:hover,.nav-tabs>li>a:focus{border-color:#eee #eee #ddd}.nav-tabs>.active>a,.nav-tabs>.active>a:hover,.nav-tabs>.active>a:focus{color:#555;cursor:default;background-color:#fff;border:1px solid #ddd;border-bottom-color:transparent}.nav-pills>li>a{padding-top:8px;padding-bottom:8px;margin-top:2px;margin-bottom:2px;-webkit-border-radius:5px;-moz-border-radius:5px;border-radius:5px}.nav-pills>.active>a,.nav-pills>.active>a:hover,.nav-pills>.active>a:focus{color:#fff;background-color:#08c}.nav-stacked>li{float:none}.nav-stacked>li>a{margin-right:0}.nav-tabs.nav-stacked{border-bottom:0}.nav-tabs.nav-stacked>li>a{border:1px solid #ddd;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.nav-tabs.nav-stacked>li:first-child>a{-webkit-border-top-right-radius:4px;border-top-right-radius:4px;-webkit-border-top-left-radius:4px;border-top-left-radius:4px;-moz-border-radius-topright:4px;-moz-border-radius-topleft:4px}.nav-tabs.nav-stacked>li:last-child>a{-webkit-border-bottom-right-radius:4px;border-bottom-right-radius:4px;-webkit-border-bottom-left-radius:4px;border-bottom-left-radius:4px;-moz-border-radius-bottomright:4px;-moz-border-radius-bottomleft:4px}.nav-tabs.nav-stacked>li>a:hover,.nav-tabs.nav-stacked>li>a:focus{z-index:2;border-color:#ddd}.nav-pills.nav-stacked>li>a{margin-bottom:3px}.nav-pills.nav-stacked>li:last-child>a{margin-bottom:1px}.nav-tabs .dropdown-menu{-webkit-border-radius:0 0 6px 6px;-moz-border-radius:0 0 6px 6px;border-radius:0 0 6px 6px}.nav-pills .dropdown-menu{-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px}.nav .dropdown-toggle .caret{margin-top:6px;border-top-color:#08c;border-bottom-color:#08c}.nav .dropdown-toggle:hover .caret,.nav .dropdown-toggle:focus .caret{border-top-color:#005580;border-bottom-color:#005580}.nav-tabs .dropdown-toggle .caret{margin-top:8px}.nav .active .dropdown-toggle .caret{border-top-color:#fff;border-bottom-color:#fff}.nav-tabs .active .dropdown-toggle .caret{border-top-color:#555;border-bottom-color:#555}.nav>.dropdown.active>a:hover,.nav>.dropdown.active>a:focus{cursor:pointer}.nav-tabs .open .dropdown-toggle,.nav-pills .open .dropdown-toggle,.nav>li.dropdown.open.active>a:hover,.nav>li.dropdown.open.active>a:focus{color:#fff;background-color:#999;border-color:#999}.nav li.dropdown.open .caret,.nav li.dropdown.open.active .caret,.nav li.dropdown.open a:hover .caret,.nav li.dropdown.open a:focus .caret{border-top-color:#fff;border-bottom-color:#fff;opacity:1;filter:alpha(opacity=100)}.tabs-stacked .open>a:hover,.tabs-stacked .open>a:focus{border-color:#999}.tabbable{*zoom:1}.tabbable:before,.tabbable:after{display:table;line-height:0;content:""}.tabbable:after{clear:both}.tab-content{overflow:auto}.tabs-below>.nav-tabs,.tabs-right>.nav-tabs,.tabs-left>.nav-tabs{border-bottom:0}.tab-content>.tab-pane,.pill-content>.pill-pane{display:none}.tab-content>.active,.pill-content>.active{display:block}.tabs-below>.nav-tabs{border-top:1px solid #ddd}.tabs-below>.nav-tabs>li{margin-top:-1px;margin-bottom:0}.tabs-below>.nav-tabs>li>a{-webkit-border-radius:0 0 4px 4px;-moz-border-radius:0 0 4px 4px;border-radius:0 0 4px 4px}.tabs-below>.nav-tabs>li>a:hover,.tabs-below>.nav-tabs>li>a:focus{border-top-color:#ddd;border-bottom-color:transparent}.tabs-below>.nav-tabs>.active>a,.tabs-below>.nav-tabs>.active>a:hover,.tabs-below>.nav-tabs>.active>a:focus{border-color:transparent #ddd #ddd #ddd}.tabs-left>.nav-tabs>li,.tabs-right>.nav-tabs>li{float:none}.tabs-left>.nav-tabs>li>a,.tabs-right>.nav-tabs>li>a{min-width:74px;margin-right:0;margin-bottom:3px}.tabs-left>.nav-tabs{float:left;margin-right:19px;border-right:1px solid #ddd}.tabs-left>.nav-tabs>li>a{margin-right:-1px;-webkit-border-radius:4px 0 0 4px;-moz-border-radius:4px 0 0 4px;border-radius:4px 0 0 4px}.tabs-left>.nav-tabs>li>a:hover,.tabs-left>.nav-tabs>li>a:focus{border-color:#eee #ddd #eee #eee}.tabs-left>.nav-tabs .active>a,.tabs-left>.nav-tabs .active>a:hover,.tabs-left>.nav-tabs .active>a:focus{border-color:#ddd transparent #ddd #ddd;*border-right-color:#fff}.tabs-right>.nav-tabs{float:right;margin-left:19px;border-left:1px solid #ddd}.tabs-right>.nav-tabs>li>a{margin-left:-1px;-webkit-border-radius:0 4px 4px 0;-moz-border-radius:0 4px 4px 0;border-radius:0 4px 4px 0}.tabs-right>.nav-tabs>li>a:hover,.tabs-right>.nav-tabs>li>a:focus{border-color:#eee #eee #eee #ddd}.tabs-right>.nav-tabs .active>a,.tabs-right>.nav-tabs .active>a:hover,.tabs-right>.nav-tabs .active>a:focus{border-color:#ddd #ddd #ddd transparent;*border-left-color:#fff}.nav>.disabled>a{color:#999}.nav>.disabled>a:hover,.nav>.disabled>a:focus{text-decoration:none;cursor:default;background-color:transparent}.navbar{*position:relative;*z-index:2;margin-bottom:20px;overflow:visible}.navbar-inner{min-height:40px;padding-right:20px;padding-left:20px;background-color:#fafafa;background-image:-moz-linear-gradient(top,#fff,#f2f2f2);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fff),to(#f2f2f2));background-image:-webkit-linear-gradient(top,#fff,#f2f2f2);background-image:-o-linear-gradient(top,#fff,#f2f2f2);background-image:linear-gradient(to bottom,#fff,#f2f2f2);background-repeat:repeat-x;border:1px solid #d4d4d4;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ffffffff',endColorstr='#fff2f2f2',GradientType=0);*zoom:1;-webkit-box-shadow:0 1px 4px rgba(0,0,0,0.065);-moz-box-shadow:0 1px 4px rgba(0,0,0,0.065);box-shadow:0 1px 4px rgba(0,0,0,0.065)}.navbar-inner:before,.navbar-inner:after{display:table;line-height:0;content:""}.navbar-inner:after{clear:both}.navbar .container{width:auto}.nav-collapse.collapse{height:auto;overflow:visible}.navbar .brand{display:block;float:left;padding:10px 20px 10px;margin-left:-20px;font-size:20px;font-weight:200;color:#777;text-shadow:0 1px 0 #fff}.navbar .brand:hover,.navbar .brand:focus{text-decoration:none}.navbar-text{margin-bottom:0;line-height:40px;color:#777}.navbar-link{color:#777}.navbar-link:hover,.navbar-link:focus{color:#333}.navbar .divider-vertical{height:40px;margin:0 9px;border-right:1px solid #fff;border-left:1px solid #f2f2f2}.navbar .btn,.navbar .btn-group{margin-top:5px}.navbar .btn-group .btn,.navbar .input-prepend .btn,.navbar .input-append .btn,.navbar .input-prepend .btn-group,.navbar .input-append .btn-group{margin-top:0}.navbar-form{margin-bottom:0;*zoom:1}.navbar-form:before,.navbar-form:after{display:table;line-height:0;content:""}.navbar-form:after{clear:both}.navbar-form input,.navbar-form select,.navbar-form .radio,.navbar-form .checkbox{margin-top:5px}.navbar-form input,.navbar-form select,.navbar-form .btn{display:inline-block;margin-bottom:0}.navbar-form input[type="image"],.navbar-form input[type="checkbox"],.navbar-form input[type="radio"]{margin-top:3px}.navbar-form .input-append,.navbar-form .input-prepend{margin-top:5px;white-space:nowrap}.navbar-form .input-append input,.navbar-form .input-prepend input{margin-top:0}.navbar-search{position:relative;float:left;margin-top:5px;margin-bottom:0}.navbar-search .search-query{padding:4px 14px;margin-bottom:0;font-family:"Helvetica Neue",Helvetica,Arial,sans-serif;font-size:13px;font-weight:normal;line-height:1;-webkit-border-radius:15px;-moz-border-radius:15px;border-radius:15px}.navbar-static-top{position:static;margin-bottom:0}.navbar-static-top .navbar-inner{-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.navbar-fixed-top,.navbar-fixed-bottom{position:fixed;right:0;left:0;z-index:1030;margin-bottom:0}.navbar-fixed-top .navbar-inner,.navbar-static-top .navbar-inner{border-width:0 0 1px}.navbar-fixed-bottom .navbar-inner{border-width:1px 0 0}.navbar-fixed-top .navbar-inner,.navbar-fixed-bottom .navbar-inner{padding-right:0;padding-left:0;-webkit-border-radius:0;-moz-border-radius:0;border-radius:0}.navbar-static-top .container,.navbar-fixed-top .container,.navbar-fixed-bottom .container{width:940px}.navbar-fixed-top{top:0}.navbar-fixed-top .navbar-inner,.navbar-static-top .navbar-inner{-webkit-box-shadow:0 1px 10px rgba(0,0,0,0.1);-moz-box-shadow:0 1px 10px rgba(0,0,0,0.1);box-shadow:0 1px 10px rgba(0,0,0,0.1)}.navbar-fixed-bottom{bottom:0}.navbar-fixed-bottom .navbar-inner{-webkit-box-shadow:0 -1px 10px rgba(0,0,0,0.1);-moz-box-shadow:0 -1px 10px rgba(0,0,0,0.1);box-shadow:0 -1px 10px rgba(0,0,0,0.1)}.navbar .nav{position:relative;left:0;display:block;float:left;margin:0 10px 0 0}.navbar .nav.pull-right{float:right;margin-right:0}.navbar .nav>li{float:left}.navbar .nav>li>a{float:none;padding:10px 15px 10px;color:#777;text-decoration:none;text-shadow:0 1px 0 #fff}.navbar .nav .dropdown-toggle .caret{margin-top:8px}.navbar .nav>li>a:focus,.navbar .nav>li>a:hover{color:#333;text-decoration:none;background-color:transparent}.navbar .nav>.active>a,.navbar .nav>.active>a:hover,.navbar .nav>.active>a:focus{color:#555;text-decoration:none;background-color:#e5e5e5;-webkit-box-shadow:inset 0 3px 8px rgba(0,0,0,0.125);-moz-box-shadow:inset 0 3px 8px rgba(0,0,0,0.125);box-shadow:inset 0 3px 8px rgba(0,0,0,0.125)}.navbar .btn-navbar{display:none;float:right;padding:7px 10px;margin-right:5px;margin-left:5px;color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#ededed;*background-color:#e5e5e5;background-image:-moz-linear-gradient(top,#f2f2f2,#e5e5e5);background-image:-webkit-gradient(linear,0 0,0 100%,from(#f2f2f2),to(#e5e5e5));background-image:-webkit-linear-gradient(top,#f2f2f2,#e5e5e5);background-image:-o-linear-gradient(top,#f2f2f2,#e5e5e5);background-image:linear-gradient(to bottom,#f2f2f2,#e5e5e5);background-repeat:repeat-x;border-color:#e5e5e5 #e5e5e5 #bfbfbf;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#fff2f2f2',endColorstr='#ffe5e5e5',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false);-webkit-box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.075);-moz-box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.075);box-shadow:inset 0 1px 0 rgba(255,255,255,0.1),0 1px 0 rgba(255,255,255,0.075)}.navbar .btn-navbar:hover,.navbar .btn-navbar:focus,.navbar .btn-navbar:active,.navbar .btn-navbar.active,.navbar .btn-navbar.disabled,.navbar .btn-navbar[disabled]{color:#fff;background-color:#e5e5e5;*background-color:#d9d9d9}.navbar .btn-navbar:active,.navbar .btn-navbar.active{background-color:#ccc \9}.navbar .btn-navbar .icon-bar{display:block;width:18px;height:2px;background-color:#f5f5f5;-webkit-border-radius:1px;-moz-border-radius:1px;border-radius:1px;-webkit-box-shadow:0 1px 0 rgba(0,0,0,0.25);-moz-box-shadow:0 1px 0 rgba(0,0,0,0.25);box-shadow:0 1px 0 rgba(0,0,0,0.25)}.btn-navbar .icon-bar+.icon-bar{margin-top:3px}.navbar .nav>li>.dropdown-menu:before{position:absolute;top:-7px;left:9px;display:inline-block;border-right:7px solid transparent;border-bottom:7px solid #ccc;border-left:7px solid transparent;border-bottom-color:rgba(0,0,0,0.2);content:''}.navbar .nav>li>.dropdown-menu:after{position:absolute;top:-6px;left:10px;display:inline-block;border-right:6px solid transparent;border-bottom:6px solid #fff;border-left:6px solid transparent;content:''}.navbar-fixed-bottom .nav>li>.dropdown-menu:before{top:auto;bottom:-7px;border-top:7px solid #ccc;border-bottom:0;border-top-color:rgba(0,0,0,0.2)}.navbar-fixed-bottom .nav>li>.dropdown-menu:after{top:auto;bottom:-6px;border-top:6px solid #fff;border-bottom:0}.navbar .nav li.dropdown>a:hover .caret,.navbar .nav li.dropdown>a:focus .caret{border-top-color:#333;border-bottom-color:#333}.navbar .nav li.dropdown.open>.dropdown-toggle,.navbar .nav li.dropdown.active>.dropdown-toggle,.navbar .nav li.dropdown.open.active>.dropdown-toggle{color:#555;background-color:#e5e5e5}.navbar .nav li.dropdown>.dropdown-toggle .caret{border-top-color:#777;border-bottom-color:#777}.navbar .nav li.dropdown.open>.dropdown-toggle .caret,.navbar .nav li.dropdown.active>.dropdown-toggle .caret,.navbar .nav li.dropdown.open.active>.dropdown-toggle .caret{border-top-color:#555;border-bottom-color:#555}.navbar .pull-right>li>.dropdown-menu,.navbar .nav>li>.dropdown-menu.pull-right{right:0;left:auto}.navbar .pull-right>li>.dropdown-menu:before,.navbar .nav>li>.dropdown-menu.pull-right:before{right:12px;left:auto}.navbar .pull-right>li>.dropdown-menu:after,.navbar .nav>li>.dropdown-menu.pull-right:after{right:13px;left:auto}.navbar .pull-right>li>.dropdown-menu .dropdown-menu,.navbar .nav>li>.dropdown-menu.pull-right .dropdown-menu{right:100%;left:auto;margin-right:-1px;margin-left:0;-webkit-border-radius:6px 0 6px 6px;-moz-border-radius:6px 0 6px 6px;border-radius:6px 0 6px 6px}.navbar-inverse .navbar-inner{background-color:#1b1b1b;background-image:-moz-linear-gradient(top,#222,#111);background-image:-webkit-gradient(linear,0 0,0 100%,from(#222),to(#111));background-image:-webkit-linear-gradient(top,#222,#111);background-image:-o-linear-gradient(top,#222,#111);background-image:linear-gradient(to bottom,#222,#111);background-repeat:repeat-x;border-color:#252525;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff222222',endColorstr='#ff111111',GradientType=0)}.navbar-inverse .brand,.navbar-inverse .nav>li>a{color:#999;text-shadow:0 -1px 0 rgba(0,0,0,0.25)}.navbar-inverse .brand:hover,.navbar-inverse .nav>li>a:hover,.navbar-inverse .brand:focus,.navbar-inverse .nav>li>a:focus{color:#fff}.navbar-inverse .brand{color:#999}.navbar-inverse .navbar-text{color:#999}.navbar-inverse .nav>li>a:focus,.navbar-inverse .nav>li>a:hover{color:#fff;background-color:transparent}.navbar-inverse .nav .active>a,.navbar-inverse .nav .active>a:hover,.navbar-inverse .nav .active>a:focus{color:#fff;background-color:#111}.navbar-inverse .navbar-link{color:#999}.navbar-inverse .navbar-link:hover,.navbar-inverse .navbar-link:focus{color:#fff}.navbar-inverse .divider-vertical{border-right-color:#222;border-left-color:#111}.navbar-inverse .nav li.dropdown.open>.dropdown-toggle,.navbar-inverse .nav li.dropdown.active>.dropdown-toggle,.navbar-inverse .nav li.dropdown.open.active>.dropdown-toggle{color:#fff;background-color:#111}.navbar-inverse .nav li.dropdown>a:hover .caret,.navbar-inverse .nav li.dropdown>a:focus .caret{border-top-color:#fff;border-bottom-color:#fff}.navbar-inverse .nav li.dropdown>.dropdown-toggle .caret{border-top-color:#999;border-bottom-color:#999}.navbar-inverse .nav li.dropdown.open>.dropdown-toggle .caret,.navbar-inverse .nav li.dropdown.active>.dropdown-toggle .caret,.navbar-inverse .nav li.dropdown.open.active>.dropdown-toggle .caret{border-top-color:#fff;border-bottom-color:#fff}.navbar-inverse .navbar-search .search-query{color:#fff;background-color:#515151;border-color:#111;-webkit-box-shadow:inset 0 1px 2px rgba(0,0,0,0.1),0 1px 0 rgba(255,255,255,0.15);-moz-box-shadow:inset 0 1px 2px rgba(0,0,0,0.1),0 1px 0 rgba(255,255,255,0.15);box-shadow:inset 0 1px 2px rgba(0,0,0,0.1),0 1px 0 rgba(255,255,255,0.15);-webkit-transition:none;-moz-transition:none;-o-transition:none;transition:none}.navbar-inverse .navbar-search .search-query:-moz-placeholder{color:#ccc}.navbar-inverse .navbar-search .search-query:-ms-input-placeholder{color:#ccc}.navbar-inverse .navbar-search .search-query::-webkit-input-placeholder{color:#ccc}.navbar-inverse .navbar-search .search-query:focus,.navbar-inverse .navbar-search .search-query.focused{padding:5px 15px;color:#333;text-shadow:0 1px 0 #fff;background-color:#fff;border:0;outline:0;-webkit-box-shadow:0 0 3px rgba(0,0,0,0.15);-moz-box-shadow:0 0 3px rgba(0,0,0,0.15);box-shadow:0 0 3px rgba(0,0,0,0.15)}.navbar-inverse .btn-navbar{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#0e0e0e;*background-color:#040404;background-image:-moz-linear-gradient(top,#151515,#040404);background-image:-webkit-gradient(linear,0 0,0 100%,from(#151515),to(#040404));background-image:-webkit-linear-gradient(top,#151515,#040404);background-image:-o-linear-gradient(top,#151515,#040404);background-image:linear-gradient(to bottom,#151515,#040404);background-repeat:repeat-x;border-color:#040404 #040404 #000;border-color:rgba(0,0,0,0.1) rgba(0,0,0,0.1) rgba(0,0,0,0.25);filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff151515',endColorstr='#ff040404',GradientType=0);filter:progid:DXImageTransform.Microsoft.gradient(enabled=false)}.navbar-inverse .btn-navbar:hover,.navbar-inverse .btn-navbar:focus,.navbar-inverse .btn-navbar:active,.navbar-inverse .btn-navbar.active,.navbar-inverse .btn-navbar.disabled,.navbar-inverse .btn-navbar[disabled]{color:#fff;background-color:#040404;*background-color:#000}.navbar-inverse .btn-navbar:active,.navbar-inverse .btn-navbar.active{background-color:#000 \9}.breadcrumb{padding:8px 15px;margin:0 0 20px;list-style:none;background-color:#f5f5f5;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.breadcrumb>li{display:inline-block;*display:inline;text-shadow:0 1px 0 #fff;*zoom:1}.breadcrumb>li>.divider{padding:0 5px;color:#ccc}.breadcrumb>.active{color:#999}.pagination{margin:20px 0}.pagination ul{display:inline-block;*display:inline;margin-bottom:0;margin-left:0;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;*zoom:1;-webkit-box-shadow:0 1px 2px rgba(0,0,0,0.05);-moz-box-shadow:0 1px 2px rgba(0,0,0,0.05);box-shadow:0 1px 2px rgba(0,0,0,0.05)}.pagination ul>li{display:inline}.pagination ul>li>a,.pagination ul>li>span{float:left;padding:4px 12px;line-height:20px;text-decoration:none;background-color:#fff;border:1px solid #ddd;border-left-width:0}.pagination ul>li>a:hover,.pagination ul>li>a:focus,.pagination ul>.active>a,.pagination ul>.active>span{background-color:#f5f5f5}.pagination ul>.active>a,.pagination ul>.active>span{color:#999;cursor:default}.pagination ul>.disabled>span,.pagination ul>.disabled>a,.pagination ul>.disabled>a:hover,.pagination ul>.disabled>a:focus{color:#999;cursor:default;background-color:transparent}.pagination ul>li:first-child>a,.pagination ul>li:first-child>span{border-left-width:1px;-webkit-border-bottom-left-radius:4px;border-bottom-left-radius:4px;-webkit-border-top-left-radius:4px;border-top-left-radius:4px;-moz-border-radius-bottomleft:4px;-moz-border-radius-topleft:4px}.pagination ul>li:last-child>a,.pagination ul>li:last-child>span{-webkit-border-top-right-radius:4px;border-top-right-radius:4px;-webkit-border-bottom-right-radius:4px;border-bottom-right-radius:4px;-moz-border-radius-topright:4px;-moz-border-radius-bottomright:4px}.pagination-centered{text-align:center}.pagination-right{text-align:right}.pagination-large ul>li>a,.pagination-large ul>li>span{padding:11px 19px;font-size:17.5px}.pagination-large ul>li:first-child>a,.pagination-large ul>li:first-child>span{-webkit-border-bottom-left-radius:6px;border-bottom-left-radius:6px;-webkit-border-top-left-radius:6px;border-top-left-radius:6px;-moz-border-radius-bottomleft:6px;-moz-border-radius-topleft:6px}.pagination-large ul>li:last-child>a,.pagination-large ul>li:last-child>span{-webkit-border-top-right-radius:6px;border-top-right-radius:6px;-webkit-border-bottom-right-radius:6px;border-bottom-right-radius:6px;-moz-border-radius-topright:6px;-moz-border-radius-bottomright:6px}.pagination-mini ul>li:first-child>a,.pagination-small ul>li:first-child>a,.pagination-mini ul>li:first-child>span,.pagination-small ul>li:first-child>span{-webkit-border-bottom-left-radius:3px;border-bottom-left-radius:3px;-webkit-border-top-left-radius:3px;border-top-left-radius:3px;-moz-border-radius-bottomleft:3px;-moz-border-radius-topleft:3px}.pagination-mini ul>li:last-child>a,.pagination-small ul>li:last-child>a,.pagination-mini ul>li:last-child>span,.pagination-small ul>li:last-child>span{-webkit-border-top-right-radius:3px;border-top-right-radius:3px;-webkit-border-bottom-right-radius:3px;border-bottom-right-radius:3px;-moz-border-radius-topright:3px;-moz-border-radius-bottomright:3px}.pagination-small ul>li>a,.pagination-small ul>li>span{padding:2px 10px;font-size:11.9px}.pagination-mini ul>li>a,.pagination-mini ul>li>span{padding:0 6px;font-size:10.5px}.pager{margin:20px 0;text-align:center;list-style:none;*zoom:1}.pager:before,.pager:after{display:table;line-height:0;content:""}.pager:after{clear:both}.pager li{display:inline}.pager li>a,.pager li>span{display:inline-block;padding:5px 14px;background-color:#fff;border:1px solid #ddd;-webkit-border-radius:15px;-moz-border-radius:15px;border-radius:15px}.pager li>a:hover,.pager li>a:focus{text-decoration:none;background-color:#f5f5f5}.pager .next>a,.pager .next>span{float:right}.pager .previous>a,.pager .previous>span{float:left}.pager .disabled>a,.pager .disabled>a:hover,.pager .disabled>a:focus,.pager .disabled>span{color:#999;cursor:default;background-color:#fff}.modal-backdrop{position:fixed;top:0;right:0;bottom:0;left:0;z-index:1040;background-color:#000}.modal-backdrop.fade{opacity:0}.modal-backdrop,.modal-backdrop.fade.in{opacity:.8;filter:alpha(opacity=80)}.modal{position:fixed;top:10%;left:50%;z-index:1050;width:560px;margin-left:-280px;background-color:#fff;border:1px solid #999;border:1px solid rgba(0,0,0,0.3);*border:1px solid #999;-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px;outline:0;-webkit-box-shadow:0 3px 7px rgba(0,0,0,0.3);-moz-box-shadow:0 3px 7px rgba(0,0,0,0.3);box-shadow:0 3px 7px rgba(0,0,0,0.3);-webkit-background-clip:padding-box;-moz-background-clip:padding-box;background-clip:padding-box}.modal.fade{top:-25%;-webkit-transition:opacity .3s linear,top .3s ease-out;-moz-transition:opacity .3s linear,top .3s ease-out;-o-transition:opacity .3s linear,top .3s ease-out;transition:opacity .3s linear,top .3s ease-out}.modal.fade.in{top:10%}.modal-header{padding:9px 15px;border-bottom:1px solid #eee}.modal-header .close{margin-top:2px}.modal-header h3{margin:0;line-height:30px}.modal-body{position:relative;max-height:400px;padding:15px;overflow-y:auto}.modal-form{margin-bottom:0}.modal-footer{padding:14px 15px 15px;margin-bottom:0;text-align:right;background-color:#f5f5f5;border-top:1px solid #ddd;-webkit-border-radius:0 0 6px 6px;-moz-border-radius:0 0 6px 6px;border-radius:0 0 6px 6px;*zoom:1;-webkit-box-shadow:inset 0 1px 0 #fff;-moz-box-shadow:inset 0 1px 0 #fff;box-shadow:inset 0 1px 0 #fff}.modal-footer:before,.modal-footer:after{display:table;line-height:0;content:""}.modal-footer:after{clear:both}.modal-footer .btn+.btn{margin-bottom:0;margin-left:5px}.modal-footer .btn-group .btn+.btn{margin-left:-1px}.modal-footer .btn-block+.btn-block{margin-left:0}.tooltip{position:absolute;z-index:1030;display:block;font-size:11px;line-height:1.4;opacity:0;filter:alpha(opacity=0);visibility:visible}.tooltip.in{opacity:.8;filter:alpha(opacity=80)}.tooltip.top{padding:5px 0;margin-top:-3px}.tooltip.right{padding:0 5px;margin-left:3px}.tooltip.bottom{padding:5px 0;margin-top:3px}.tooltip.left{padding:0 5px;margin-left:-3px}.tooltip-inner{max-width:200px;padding:8px;color:#fff;text-align:center;text-decoration:none;background-color:#000;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.tooltip-arrow{position:absolute;width:0;height:0;border-color:transparent;border-style:solid}.tooltip.top .tooltip-arrow{bottom:0;left:50%;margin-left:-5px;border-top-color:#000;border-width:5px 5px 0}.tooltip.right .tooltip-arrow{top:50%;left:0;margin-top:-5px;border-right-color:#000;border-width:5px 5px 5px 0}.tooltip.left .tooltip-arrow{top:50%;right:0;margin-top:-5px;border-left-color:#000;border-width:5px 0 5px 5px}.tooltip.bottom .tooltip-arrow{top:0;left:50%;margin-left:-5px;border-bottom-color:#000;border-width:0 5px 5px}.popover{position:absolute;top:0;left:0;z-index:1010;display:none;max-width:276px;padding:1px;text-align:left;white-space:normal;background-color:#fff;border:1px solid #ccc;border:1px solid rgba(0,0,0,0.2);-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px;-webkit-box-shadow:0 5px 10px rgba(0,0,0,0.2);-moz-box-shadow:0 5px 10px rgba(0,0,0,0.2);box-shadow:0 5px 10px rgba(0,0,0,0.2);-webkit-background-clip:padding-box;-moz-background-clip:padding;background-clip:padding-box}.popover.top{margin-top:-10px}.popover.right{margin-left:10px}.popover.bottom{margin-top:10px}.popover.left{margin-left:-10px}.popover-title{padding:8px 14px;margin:0;font-size:14px;font-weight:normal;line-height:18px;background-color:#f7f7f7;border-bottom:1px solid #ebebeb;-webkit-border-radius:5px 5px 0 0;-moz-border-radius:5px 5px 0 0;border-radius:5px 5px 0 0}.popover-title:empty{display:none}.popover-content{padding:9px 14px}.popover .arrow,.popover .arrow:after{position:absolute;display:block;width:0;height:0;border-color:transparent;border-style:solid}.popover .arrow{border-width:11px}.popover .arrow:after{border-width:10px;content:""}.popover.top .arrow{bottom:-11px;left:50%;margin-left:-11px;border-top-color:#999;border-top-color:rgba(0,0,0,0.25);border-bottom-width:0}.popover.top .arrow:after{bottom:1px;margin-left:-10px;border-top-color:#fff;border-bottom-width:0}.popover.right .arrow{top:50%;left:-11px;margin-top:-11px;border-right-color:#999;border-right-color:rgba(0,0,0,0.25);border-left-width:0}.popover.right .arrow:after{bottom:-10px;left:1px;border-right-color:#fff;border-left-width:0}.popover.bottom .arrow{top:-11px;left:50%;margin-left:-11px;border-bottom-color:#999;border-bottom-color:rgba(0,0,0,0.25);border-top-width:0}.popover.bottom .arrow:after{top:1px;margin-left:-10px;border-bottom-color:#fff;border-top-width:0}.popover.left .arrow{top:50%;right:-11px;margin-top:-11px;border-left-color:#999;border-left-color:rgba(0,0,0,0.25);border-right-width:0}.popover.left .arrow:after{right:1px;bottom:-10px;border-left-color:#fff;border-right-width:0}.thumbnails{margin-left:-20px;list-style:none;*zoom:1}.thumbnails:before,.thumbnails:after{display:table;line-height:0;content:""}.thumbnails:after{clear:both}.row-fluid .thumbnails{margin-left:0}.thumbnails>li{float:left;margin-bottom:20px;margin-left:20px}.thumbnail{display:block;padding:4px;line-height:20px;border:1px solid #ddd;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;-webkit-box-shadow:0 1px 3px rgba(0,0,0,0.055);-moz-box-shadow:0 1px 3px rgba(0,0,0,0.055);box-shadow:0 1px 3px rgba(0,0,0,0.055);-webkit-transition:all .2s ease-in-out;-moz-transition:all .2s ease-in-out;-o-transition:all .2s ease-in-out;transition:all .2s ease-in-out}a.thumbnail:hover,a.thumbnail:focus{border-color:#08c;-webkit-box-shadow:0 1px 4px rgba(0,105,214,0.25);-moz-box-shadow:0 1px 4px rgba(0,105,214,0.25);box-shadow:0 1px 4px rgba(0,105,214,0.25)}.thumbnail>img{display:block;max-width:100%;margin-right:auto;margin-left:auto}.thumbnail .caption{padding:9px;color:#555}.media,.media-body{overflow:hidden;*overflow:visible;zoom:1}.media,.media .media{margin-top:15px}.media:first-child{margin-top:0}.media-object{display:block}.media-heading{margin:0 0 5px}.media>.pull-left{margin-right:10px}.media>.pull-right{margin-left:10px}.media-list{margin-left:0;list-style:none}.label,.badge{display:inline-block;padding:2px 4px;font-size:11.844px;font-weight:bold;line-height:14px;color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);white-space:nowrap;vertical-align:baseline;background-color:#999}.label{-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}.badge{padding-right:9px;padding-left:9px;-webkit-border-radius:9px;-moz-border-radius:9px;border-radius:9px}.label:empty,.badge:empty{display:none}a.label:hover,a.label:focus,a.badge:hover,a.badge:focus{color:#fff;text-decoration:none;cursor:pointer}.label-important,.badge-important{background-color:#b94a48}.label-important[href],.badge-important[href]{background-color:#953b39}.label-warning,.badge-warning{background-color:#f89406}.label-warning[href],.badge-warning[href]{background-color:#c67605}.label-success,.badge-success{background-color:#468847}.label-success[href],.badge-success[href]{background-color:#356635}.label-info,.badge-info{background-color:#3a87ad}.label-info[href],.badge-info[href]{background-color:#2d6987}.label-inverse,.badge-inverse{background-color:#333}.label-inverse[href],.badge-inverse[href]{background-color:#1a1a1a}.btn .label,.btn .badge{position:relative;top:-1px}.btn-mini .label,.btn-mini .badge{top:0}@-webkit-keyframes progress-bar-stripes{from{background-position:40px 0}to{background-position:0 0}}@-moz-keyframes progress-bar-stripes{from{background-position:40px 0}to{background-position:0 0}}@-ms-keyframes progress-bar-stripes{from{background-position:40px 0}to{background-position:0 0}}@-o-keyframes progress-bar-stripes{from{background-position:0 0}to{background-position:40px 0}}@keyframes progress-bar-stripes{from{background-position:40px 0}to{background-position:0 0}}.progress{height:20px;margin-bottom:20px;overflow:hidden;background-color:#f7f7f7;background-image:-moz-linear-gradient(top,#f5f5f5,#f9f9f9);background-image:-webkit-gradient(linear,0 0,0 100%,from(#f5f5f5),to(#f9f9f9));background-image:-webkit-linear-gradient(top,#f5f5f5,#f9f9f9);background-image:-o-linear-gradient(top,#f5f5f5,#f9f9f9);background-image:linear-gradient(to bottom,#f5f5f5,#f9f9f9);background-repeat:repeat-x;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#fff5f5f5',endColorstr='#fff9f9f9',GradientType=0);-webkit-box-shadow:inset 0 1px 2px rgba(0,0,0,0.1);-moz-box-shadow:inset 0 1px 2px rgba(0,0,0,0.1);box-shadow:inset 0 1px 2px rgba(0,0,0,0.1)}.progress .bar{float:left;width:0;height:100%;font-size:12px;color:#fff;text-align:center;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#0e90d2;background-image:-moz-linear-gradient(top,#149bdf,#0480be);background-image:-webkit-gradient(linear,0 0,0 100%,from(#149bdf),to(#0480be));background-image:-webkit-linear-gradient(top,#149bdf,#0480be);background-image:-o-linear-gradient(top,#149bdf,#0480be);background-image:linear-gradient(to bottom,#149bdf,#0480be);background-repeat:repeat-x;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff149bdf',endColorstr='#ff0480be',GradientType=0);-webkit-box-shadow:inset 0 -1px 0 rgba(0,0,0,0.15);-moz-box-shadow:inset 0 -1px 0 rgba(0,0,0,0.15);box-shadow:inset 0 -1px 0 rgba(0,0,0,0.15);-webkit-box-sizing:border-box;-moz-box-sizing:border-box;box-sizing:border-box;-webkit-transition:width .6s ease;-moz-transition:width .6s ease;-o-transition:width .6s ease;transition:width .6s ease}.progress .bar+.bar{-webkit-box-shadow:inset 1px 0 0 rgba(0,0,0,0.15),inset 0 -1px 0 rgba(0,0,0,0.15);-moz-box-shadow:inset 1px 0 0 rgba(0,0,0,0.15),inset 0 -1px 0 rgba(0,0,0,0.15);box-shadow:inset 1px 0 0 rgba(0,0,0,0.15),inset 0 -1px 0 rgba(0,0,0,0.15)}.progress-striped .bar{background-color:#149bdf;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);-webkit-background-size:40px 40px;-moz-background-size:40px 40px;-o-background-size:40px 40px;background-size:40px 40px}.progress.active .bar{-webkit-animation:progress-bar-stripes 2s linear infinite;-moz-animation:progress-bar-stripes 2s linear infinite;-ms-animation:progress-bar-stripes 2s linear infinite;-o-animation:progress-bar-stripes 2s linear infinite;animation:progress-bar-stripes 2s linear infinite}.progress-danger .bar,.progress .bar-danger{background-color:#dd514c;background-image:-moz-linear-gradient(top,#ee5f5b,#c43c35);background-image:-webkit-gradient(linear,0 0,0 100%,from(#ee5f5b),to(#c43c35));background-image:-webkit-linear-gradient(top,#ee5f5b,#c43c35);background-image:-o-linear-gradient(top,#ee5f5b,#c43c35);background-image:linear-gradient(to bottom,#ee5f5b,#c43c35);background-repeat:repeat-x;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ffee5f5b',endColorstr='#ffc43c35',GradientType=0)}.progress-danger.progress-striped .bar,.progress-striped .bar-danger{background-color:#ee5f5b;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent)}.progress-success .bar,.progress .bar-success{background-color:#5eb95e;background-image:-moz-linear-gradient(top,#62c462,#57a957);background-image:-webkit-gradient(linear,0 0,0 100%,from(#62c462),to(#57a957));background-image:-webkit-linear-gradient(top,#62c462,#57a957);background-image:-o-linear-gradient(top,#62c462,#57a957);background-image:linear-gradient(to bottom,#62c462,#57a957);background-repeat:repeat-x;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff62c462',endColorstr='#ff57a957',GradientType=0)}.progress-success.progress-striped .bar,.progress-striped .bar-success{background-color:#62c462;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent)}.progress-info .bar,.progress .bar-info{background-color:#4bb1cf;background-image:-moz-linear-gradient(top,#5bc0de,#339bb9);background-image:-webkit-gradient(linear,0 0,0 100%,from(#5bc0de),to(#339bb9));background-image:-webkit-linear-gradient(top,#5bc0de,#339bb9);background-image:-o-linear-gradient(top,#5bc0de,#339bb9);background-image:linear-gradient(to bottom,#5bc0de,#339bb9);background-repeat:repeat-x;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#ff5bc0de',endColorstr='#ff339bb9',GradientType=0)}.progress-info.progress-striped .bar,.progress-striped .bar-info{background-color:#5bc0de;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent)}.progress-warning .bar,.progress .bar-warning{background-color:#faa732;background-image:-moz-linear-gradient(top,#fbb450,#f89406);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fbb450),to(#f89406));background-image:-webkit-linear-gradient(top,#fbb450,#f89406);background-image:-o-linear-gradient(top,#fbb450,#f89406);background-image:linear-gradient(to bottom,#fbb450,#f89406);background-repeat:repeat-x;filter:progid:DXImageTransform.Microsoft.gradient(startColorstr='#fffbb450',endColorstr='#fff89406',GradientType=0)}.progress-warning.progress-striped .bar,.progress-striped .bar-warning{background-color:#fbb450;background-image:-webkit-gradient(linear,0 100%,100% 0,color-stop(0.25,rgba(255,255,255,0.15)),color-stop(0.25,transparent),color-stop(0.5,transparent),color-stop(0.5,rgba(255,255,255,0.15)),color-stop(0.75,rgba(255,255,255,0.15)),color-stop(0.75,transparent),to(transparent));background-image:-webkit-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-moz-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:-o-linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent);background-image:linear-gradient(45deg,rgba(255,255,255,0.15) 25%,transparent 25%,transparent 50%,rgba(255,255,255,0.15) 50%,rgba(255,255,255,0.15) 75%,transparent 75%,transparent)}.accordion{margin-bottom:20px}.accordion-group{margin-bottom:2px;border:1px solid #e5e5e5;-webkit-border-radius:4px;-moz-border-radius:4px;border-radius:4px}.accordion-heading{border-bottom:0}.accordion-heading .accordion-toggle{display:block;padding:8px 15px}.accordion-toggle{cursor:pointer}.accordion-inner{padding:9px 15px;border-top:1px solid #e5e5e5}.carousel{position:relative;margin-bottom:20px;line-height:1}.carousel-inner{position:relative;width:100%;overflow:hidden}.carousel-inner>.item{position:relative;display:none;-webkit-transition:.6s ease-in-out left;-moz-transition:.6s ease-in-out left;-o-transition:.6s ease-in-out left;transition:.6s ease-in-out left}.carousel-inner>.item>img,.carousel-inner>.item>a>img{display:block;line-height:1}.carousel-inner>.active,.carousel-inner>.next,.carousel-inner>.prev{display:block}.carousel-inner>.active{left:0}.carousel-inner>.next,.carousel-inner>.prev{position:absolute;top:0;width:100%}.carousel-inner>.next{left:100%}.carousel-inner>.prev{left:-100%}.carousel-inner>.next.left,.carousel-inner>.prev.right{left:0}.carousel-inner>.active.left{left:-100%}.carousel-inner>.active.right{left:100%}.carousel-control{position:absolute;top:40%;left:15px;width:40px;height:40px;margin-top:-20px;font-size:60px;font-weight:100;line-height:30px;color:#fff;text-align:center;background:#222;border:3px solid #fff;-webkit-border-radius:23px;-moz-border-radius:23px;border-radius:23px;opacity:.5;filter:alpha(opacity=50)}.carousel-control.right{right:15px;left:auto}.carousel-control:hover,.carousel-control:focus{color:#fff;text-decoration:none;opacity:.9;filter:alpha(opacity=90)}.carousel-indicators{position:absolute;top:15px;right:15px;z-index:5;margin:0;list-style:none}.carousel-indicators li{display:block;float:left;width:10px;height:10px;margin-left:5px;text-indent:-999px;background-color:#ccc;background-color:rgba(255,255,255,0.25);border-radius:5px}.carousel-indicators .active{background-color:#fff}.carousel-caption{position:absolute;right:0;bottom:0;left:0;padding:15px;background:#333;background:rgba(0,0,0,0.75)}.carousel-caption h4,.carousel-caption p{line-height:20px;color:#fff}.carousel-caption h4{margin:0 0 5px}.carousel-caption p{margin-bottom:0}.hero-unit{padding:60px;margin-bottom:30px;font-size:18px;font-weight:200;line-height:30px;color:inherit;background-color:#eee;-webkit-border-radius:6px;-moz-border-radius:6px;border-radius:6px}.hero-unit h1{margin-bottom:0;font-size:60px;line-height:1;letter-spacing:-1px;color:inherit}.hero-unit li{line-height:30px}.pull-right{float:right}.pull-left{float:left}.hide{display:none}.show{display:block}.invisible{visibility:hidden}.affix{position:fixed} diff --git a/core/src/main/resources/spark/deploy/static/sorttable.js b/core/src/main/resources/spark/ui/static/sorttable.js index 7abb9011cc..7abb9011cc 100644 --- a/core/src/main/resources/spark/deploy/static/sorttable.js +++ b/core/src/main/resources/spark/ui/static/sorttable.js diff --git a/core/src/main/resources/spark/ui/static/spark-logo-77x50px-hd.png b/core/src/main/resources/spark/ui/static/spark-logo-77x50px-hd.png Binary files differnew file mode 100644 index 0000000000..6c5f0993c4 --- /dev/null +++ b/core/src/main/resources/spark/ui/static/spark-logo-77x50px-hd.png diff --git a/core/src/main/resources/spark/deploy/static/spark_logo.png b/core/src/main/resources/spark/ui/static/spark_logo.png Binary files differindex 4b18734779..4b18734779 100644 --- a/core/src/main/resources/spark/deploy/static/spark_logo.png +++ b/core/src/main/resources/spark/ui/static/spark_logo.png diff --git a/core/src/main/resources/spark/ui/static/webui.css b/core/src/main/resources/spark/ui/static/webui.css new file mode 100644 index 0000000000..f7537bb766 --- /dev/null +++ b/core/src/main/resources/spark/ui/static/webui.css @@ -0,0 +1,49 @@ +.navbar .brand { + height: 50px; + width: 110px; + margin-left: 1px; + padding: 0; +} + +.version { + line-height: 30px; + vertical-align: bottom; + font-size: 12px; + padding: 0; + margin: 0; + font-weight: bold; + color: #777; +} + +.navbar-inner { + padding-top: 2px; + height: 50px; +} + +.navbar-inner .nav { + margin-top: 5px; + font-size: 15px; +} + + +#infolist { + margin-left: 400px; + margin-top: 14px; +} + +#infolist li { + display: inline; + list-style-type: none; + list-style-position: outside; + padding-right: 20px; + padding-top: 10px; + padding-bottom: 10px; +} + +.progress-cell { + width: 134px; + border-right: 0; + padding: 0; + padding-top: 7px; + padding-left: 4px; +} diff --git a/core/src/main/scala/spark/Accumulators.scala b/core/src/main/scala/spark/Accumulators.scala index 57c6df35be..6ff92ce833 100644 --- a/core/src/main/scala/spark/Accumulators.scala +++ b/core/src/main/scala/spark/Accumulators.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io._ diff --git a/core/src/main/scala/spark/Aggregator.scala b/core/src/main/scala/spark/Aggregator.scala index df8ce9c054..136b4da61e 100644 --- a/core/src/main/scala/spark/Aggregator.scala +++ b/core/src/main/scala/spark/Aggregator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.util.{HashMap => JHashMap} diff --git a/core/src/main/scala/spark/BlockStoreShuffleFetcher.scala b/core/src/main/scala/spark/BlockStoreShuffleFetcher.scala index 2987dbbe58..8f6953b1f5 100644 --- a/core/src/main/scala/spark/BlockStoreShuffleFetcher.scala +++ b/core/src/main/scala/spark/BlockStoreShuffleFetcher.scala @@ -1,14 +1,36 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark -import executor.{ShuffleReadMetrics, TaskMetrics} import scala.collection.mutable.ArrayBuffer import scala.collection.mutable.HashMap +import spark.executor.{ShuffleReadMetrics, TaskMetrics} +import spark.serializer.Serializer import spark.storage.BlockManagerId import spark.util.CompletionIterator + private[spark] class BlockStoreShuffleFetcher extends ShuffleFetcher with Logging { - override def fetch[K, V](shuffleId: Int, reduceId: Int, metrics: TaskMetrics) = { + + override def fetch[K, V]( + shuffleId: Int, reduceId: Int, metrics: TaskMetrics, serializer: Serializer) = { + logDebug("Fetching outputs for shuffle %d, reduce %d".format(shuffleId, reduceId)) val blockManager = SparkEnv.get.blockManager @@ -48,10 +70,12 @@ private[spark] class BlockStoreShuffleFetcher extends ShuffleFetcher with Loggin } } - val blockFetcherItr = blockManager.getMultiple(blocksByAddress) + val blockFetcherItr = blockManager.getMultiple(blocksByAddress, serializer) val itr = blockFetcherItr.flatMap(unpackBlock) + CompletionIterator[(K,V), Iterator[(K,V)]](itr, { val shuffleMetrics = new ShuffleReadMetrics + shuffleMetrics.shuffleFinishTime = System.currentTimeMillis shuffleMetrics.remoteFetchTime = blockFetcherItr.remoteFetchTime shuffleMetrics.fetchWaitTime = blockFetcherItr.fetchWaitTime shuffleMetrics.remoteBytesRead = blockFetcherItr.remoteBytesRead diff --git a/core/src/main/scala/spark/Cache.scala b/core/src/main/scala/spark/Cache.scala index 20d677a854..b0c83ce59d 100644 --- a/core/src/main/scala/spark/Cache.scala +++ b/core/src/main/scala/spark/Cache.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.util.concurrent.atomic.AtomicInteger diff --git a/core/src/main/scala/spark/CacheManager.scala b/core/src/main/scala/spark/CacheManager.scala index f7a2b7e802..81314805a9 100644 --- a/core/src/main/scala/spark/CacheManager.scala +++ b/core/src/main/scala/spark/CacheManager.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import scala.collection.mutable.{ArrayBuffer, HashSet} diff --git a/core/src/main/scala/spark/ClosureCleaner.scala b/core/src/main/scala/spark/ClosureCleaner.scala index 50d6a1c5c9..8b39241095 100644 --- a/core/src/main/scala/spark/ClosureCleaner.scala +++ b/core/src/main/scala/spark/ClosureCleaner.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.lang.reflect.Field @@ -5,8 +22,7 @@ import java.lang.reflect.Field import scala.collection.mutable.Map import scala.collection.mutable.Set -import org.objectweb.asm.{ClassReader, MethodVisitor, Type} -import org.objectweb.asm.commons.EmptyVisitor +import org.objectweb.asm.{ClassReader, ClassVisitor, MethodVisitor, Type} import org.objectweb.asm.Opcodes._ import java.io.{InputStream, IOException, ByteArrayOutputStream, ByteArrayInputStream, BufferedInputStream} @@ -162,10 +178,10 @@ private[spark] object ClosureCleaner extends Logging { } } -private[spark] class FieldAccessFinder(output: Map[Class[_], Set[String]]) extends EmptyVisitor { +private[spark] class FieldAccessFinder(output: Map[Class[_], Set[String]]) extends ClassVisitor(ASM4) { override def visitMethod(access: Int, name: String, desc: String, sig: String, exceptions: Array[String]): MethodVisitor = { - return new EmptyVisitor { + return new MethodVisitor(ASM4) { override def visitFieldInsn(op: Int, owner: String, name: String, desc: String) { if (op == GETFIELD) { for (cl <- output.keys if cl.getName == owner.replace('/', '.')) { @@ -188,7 +204,7 @@ private[spark] class FieldAccessFinder(output: Map[Class[_], Set[String]]) exten } } -private[spark] class InnerClosureFinder(output: Set[Class[_]]) extends EmptyVisitor { +private[spark] class InnerClosureFinder(output: Set[Class[_]]) extends ClassVisitor(ASM4) { var myName: String = null override def visit(version: Int, access: Int, name: String, sig: String, @@ -198,7 +214,7 @@ private[spark] class InnerClosureFinder(output: Set[Class[_]]) extends EmptyVisi override def visitMethod(access: Int, name: String, desc: String, sig: String, exceptions: Array[String]): MethodVisitor = { - return new EmptyVisitor { + return new MethodVisitor(ASM4) { override def visitMethodInsn(op: Int, owner: String, name: String, desc: String) { val argTypes = Type.getArgumentTypes(desc) diff --git a/core/src/main/scala/spark/Dependency.scala b/core/src/main/scala/spark/Dependency.scala index 5eea907322..d17e70a4fa 100644 --- a/core/src/main/scala/spark/Dependency.scala +++ b/core/src/main/scala/spark/Dependency.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark /** @@ -25,10 +42,12 @@ abstract class NarrowDependency[T](rdd: RDD[T]) extends Dependency(rdd) { * @param shuffleId the shuffle id * @param rdd the parent RDD * @param partitioner partitioner used to partition the shuffle output + * @param serializerClass class name of the serializer to use */ class ShuffleDependency[K, V]( @transient rdd: RDD[(K, V)], - val partitioner: Partitioner) + val partitioner: Partitioner, + val serializerClass: String = null) extends Dependency(rdd) { val shuffleId: Int = rdd.context.newShuffleId() diff --git a/core/src/main/scala/spark/DoubleRDDFunctions.scala b/core/src/main/scala/spark/DoubleRDDFunctions.scala index 178d31a73b..104168e61c 100644 --- a/core/src/main/scala/spark/DoubleRDDFunctions.scala +++ b/core/src/main/scala/spark/DoubleRDDFunctions.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import spark.partial.BoundedDouble @@ -37,7 +54,13 @@ class DoubleRDDFunctions(self: RDD[Double]) extends Logging with Serializable { * Compute the sample standard deviation of this RDD's elements (which corrects for bias in * estimating the standard deviation by dividing by N-1 instead of N). */ - def sampleStdev(): Double = stats().stdev + def sampleStdev(): Double = stats().sampleStdev + + /** + * Compute the sample variance of this RDD's elements (which corrects for bias in + * estimating the variance by dividing by N-1 instead of N). + */ + def sampleVariance(): Double = stats().sampleVariance /** (Experimental) Approximate operation to return the mean within a timeout. */ def meanApprox(timeout: Long, confidence: Double = 0.95): PartialResult[BoundedDouble] = { diff --git a/core/src/main/scala/spark/FetchFailedException.scala b/core/src/main/scala/spark/FetchFailedException.scala index 40b0193f19..a2dae6cae9 100644 --- a/core/src/main/scala/spark/FetchFailedException.scala +++ b/core/src/main/scala/spark/FetchFailedException.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import spark.storage.BlockManagerId diff --git a/core/src/main/scala/spark/HadoopWriter.scala b/core/src/main/scala/spark/HadoopWriter.scala index 5e8396edb9..b1fe0075a3 100644 --- a/core/src/main/scala/spark/HadoopWriter.scala +++ b/core/src/main/scala/spark/HadoopWriter.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package org.apache.hadoop.mapred import org.apache.hadoop.fs.FileSystem diff --git a/core/src/main/scala/spark/HttpFileServer.scala b/core/src/main/scala/spark/HttpFileServer.scala index 00901d95e2..a13a7a2859 100644 --- a/core/src/main/scala/spark/HttpFileServer.scala +++ b/core/src/main/scala/spark/HttpFileServer.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io.{File} @@ -42,4 +59,4 @@ private[spark] class HttpFileServer extends Logging { return dir + "/" + file.getName } -}
\ No newline at end of file +} diff --git a/core/src/main/scala/spark/HttpServer.scala b/core/src/main/scala/spark/HttpServer.scala index 4e0507c080..c9dffbc631 100644 --- a/core/src/main/scala/spark/HttpServer.scala +++ b/core/src/main/scala/spark/HttpServer.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io.File diff --git a/core/src/main/scala/spark/JavaSerializer.scala b/core/src/main/scala/spark/JavaSerializer.scala index b04a27d073..04c5f44e6b 100644 --- a/core/src/main/scala/spark/JavaSerializer.scala +++ b/core/src/main/scala/spark/JavaSerializer.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io._ diff --git a/core/src/main/scala/spark/KryoSerializer.scala b/core/src/main/scala/spark/KryoSerializer.scala index d723ab7b1e..eeb2993d8a 100644 --- a/core/src/main/scala/spark/KryoSerializer.scala +++ b/core/src/main/scala/spark/KryoSerializer.scala @@ -1,25 +1,34 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io._ import java.nio.ByteBuffer -import java.nio.channels.Channels - -import scala.collection.immutable -import scala.collection.mutable - -import com.esotericsoftware.kryo._ -import com.esotericsoftware.kryo.{Serializer => KSerializer} +import com.esotericsoftware.kryo.{Kryo, KryoException} import com.esotericsoftware.kryo.io.{Input => KryoInput, Output => KryoOutput} import com.esotericsoftware.kryo.serializers.{JavaSerializer => KryoJavaSerializer} -import de.javakaffee.kryoserializers.KryoReflectionFactorySupport - +import com.twitter.chill.ScalaKryoInstantiator import serializer.{SerializerInstance, DeserializationStream, SerializationStream} import spark.broadcast._ import spark.storage._ private[spark] class KryoSerializationStream(kryo: Kryo, outStream: OutputStream) extends SerializationStream { - val output = new KryoOutput(outStream) def writeObject[T](t: T): SerializationStream = { @@ -33,7 +42,6 @@ class KryoSerializationStream(kryo: Kryo, outStream: OutputStream) extends Seria private[spark] class KryoDeserializationStream(kryo: Kryo, inStream: InputStream) extends DeserializationStream { - val input = new KryoInput(inStream) def readObject[T](): T = { @@ -41,7 +49,7 @@ class KryoDeserializationStream(kryo: Kryo, inStream: InputStream) extends Deser kryo.readClassAndObject(input).asInstanceOf[T] } catch { // DeserializationStream uses the EOF exception to indicate stopping condition. - case e: com.esotericsoftware.kryo.KryoException => throw new java.io.EOFException + case _: KryoException => throw new EOFException } } @@ -52,10 +60,9 @@ class KryoDeserializationStream(kryo: Kryo, inStream: InputStream) extends Deser } private[spark] class KryoSerializerInstance(ks: KryoSerializer) extends SerializerInstance { - - val kryo = ks.kryo.get() - val output = ks.output.get() - val input = ks.input.get() + val kryo = ks.newKryo() + val output = ks.newKryoOutput() + val input = ks.newKryoInput() def serialize[T](t: T): ByteBuffer = { output.clear() @@ -91,130 +98,59 @@ private[spark] class KryoSerializerInstance(ks: KryoSerializer) extends Serializ * serialization. */ trait KryoRegistrator { - def registerClasses(kryo: Kryo): Unit + def registerClasses(kryo: Kryo) } /** * A Spark serializer that uses the [[http://code.google.com/p/kryo/wiki/V1Documentation Kryo 1.x library]]. */ class KryoSerializer extends spark.serializer.Serializer with Logging { + private val bufferSize = System.getProperty("spark.kryoserializer.buffer.mb", "2").toInt * 1024 * 1024 - val bufferSize = System.getProperty("spark.kryoserializer.buffer.mb", "2").toInt * 1024 * 1024 - - val kryo = new ThreadLocal[Kryo] { - override def initialValue = createKryo() - } - - val output = new ThreadLocal[KryoOutput] { - override def initialValue = new KryoOutput(bufferSize) - } + def newKryoOutput() = new KryoOutput(bufferSize) - val input = new ThreadLocal[KryoInput] { - override def initialValue = new KryoInput(bufferSize) - } + def newKryoInput() = new KryoInput(bufferSize) - def createKryo(): Kryo = { - val kryo = new KryoReflectionFactorySupport() + def newKryo(): Kryo = { + val instantiator = new ScalaKryoInstantiator + val kryo = instantiator.newKryo() + val classLoader = Thread.currentThread.getContextClassLoader // Register some commonly used classes val toRegister: Seq[AnyRef] = Seq( - // Arrays - Array(1), Array(1.0), Array(1.0f), Array(1L), Array(""), Array(("", "")), - Array(new java.lang.Object), Array(1.toByte), Array(true), Array('c'), - // Specialized Tuple2s - ("", ""), ("", 1), (1, 1), (1.0, 1.0), (1L, 1L), - (1, 1.0), (1.0, 1), (1L, 1.0), (1.0, 1L), (1, 1L), (1L, 1), - // Scala collections - List(1), mutable.ArrayBuffer(1), - // Options and Either - Some(1), Left(1), Right(1), - // Higher-dimensional tuples - (1, 1, 1), (1, 1, 1, 1), (1, 1, 1, 1, 1), - None, ByteBuffer.allocate(1), StorageLevel.MEMORY_ONLY, PutBlock("1", ByteBuffer.allocate(1), StorageLevel.MEMORY_ONLY), GotBlock("1", ByteBuffer.allocate(1)), GetBlock("1") ) - for (obj <- toRegister) { - kryo.register(obj.getClass) - } + + for (obj <- toRegister) kryo.register(obj.getClass) // Allow sending SerializableWritable kryo.register(classOf[SerializableWritable[_]], new KryoJavaSerializer()) kryo.register(classOf[HttpBroadcast[_]], new KryoJavaSerializer()) - // Register some commonly used Scala singleton objects. Because these - // are singletons, we must return the exact same local object when we - // deserialize rather than returning a clone as FieldSerializer would. - class SingletonSerializer[T](obj: T) extends KSerializer[T] { - override def write(kryo: Kryo, output: KryoOutput, obj: T) {} - override def read(kryo: Kryo, input: KryoInput, cls: java.lang.Class[T]): T = obj - } - kryo.register(None.getClass, new SingletonSerializer[AnyRef](None)) - kryo.register(Nil.getClass, new SingletonSerializer[AnyRef](Nil)) - - // Register maps with a special serializer since they have complex internal structure - class ScalaMapSerializer(buildMap: Array[(Any, Any)] => scala.collection.Map[Any, Any]) - extends KSerializer[Array[(Any, Any)] => scala.collection.Map[Any, Any]] { - - //hack, look at https://groups.google.com/forum/#!msg/kryo-users/Eu5V4bxCfws/k-8UQ22y59AJ - private final val FAKE_REFERENCE = new Object() - override def write( - kryo: Kryo, - output: KryoOutput, - obj: Array[(Any, Any)] => scala.collection.Map[Any, Any]) { - val map = obj.asInstanceOf[scala.collection.Map[Any, Any]] - output.writeInt(map.size) - for ((k, v) <- map) { - kryo.writeClassAndObject(output, k) - kryo.writeClassAndObject(output, v) - } - } - override def read ( - kryo: Kryo, - input: KryoInput, - cls: Class[Array[(Any, Any)] => scala.collection.Map[Any, Any]]) - : Array[(Any, Any)] => scala.collection.Map[Any, Any] = { - kryo.reference(FAKE_REFERENCE) - val size = input.readInt() - val elems = new Array[(Any, Any)](size) - for (i <- 0 until size) { - val k = kryo.readClassAndObject(input) - val v = kryo.readClassAndObject(input) - elems(i)=(k,v) - } - buildMap(elems).asInstanceOf[Array[(Any, Any)] => scala.collection.Map[Any, Any]] + // Allow the user to register their own classes by setting spark.kryo.registrator + try { + Option(System.getProperty("spark.kryo.registrator")).foreach { regCls => + logDebug("Running user registrator: " + regCls) + val reg = Class.forName(regCls, true, classLoader).newInstance().asInstanceOf[KryoRegistrator] + reg.registerClasses(kryo) } + } catch { + case _: Exception => println("Failed to register spark.kryo.registrator") } - kryo.register(mutable.HashMap().getClass, new ScalaMapSerializer(mutable.HashMap() ++ _)) - // TODO: add support for immutable maps too; this is more annoying because there are many - // subclasses of immutable.Map for small maps (with <= 4 entries) - val map1 = Map[Any, Any](1 -> 1) - val map2 = Map[Any, Any](1 -> 1, 2 -> 2) - val map3 = Map[Any, Any](1 -> 1, 2 -> 2, 3 -> 3) - val map4 = Map[Any, Any](1 -> 1, 2 -> 2, 3 -> 3, 4 -> 4) - val map5 = Map[Any, Any](1 -> 1, 2 -> 2, 3 -> 3, 4 -> 4, 5 -> 5) - kryo.register(map1.getClass, new ScalaMapSerializer(mutable.HashMap() ++ _ toMap)) - kryo.register(map2.getClass, new ScalaMapSerializer(mutable.HashMap() ++ _ toMap)) - kryo.register(map3.getClass, new ScalaMapSerializer(mutable.HashMap() ++ _ toMap)) - kryo.register(map4.getClass, new ScalaMapSerializer(mutable.HashMap() ++ _ toMap)) - kryo.register(map5.getClass, new ScalaMapSerializer(mutable.HashMap() ++ _ toMap)) - // Allow the user to register their own classes by setting spark.kryo.registrator - val regCls = System.getProperty("spark.kryo.registrator") - if (regCls != null) { - logInfo("Running user registrator: " + regCls) - val classLoader = Thread.currentThread.getContextClassLoader - val reg = Class.forName(regCls, true, classLoader).newInstance().asInstanceOf[KryoRegistrator] - reg.registerClasses(kryo) - } + kryo.setClassLoader(classLoader) + + // Allow disabling Kryo reference tracking if user knows their object graphs don't have loops + kryo.setReferences(System.getProperty("spark.kryo.referenceTracking", "true").toBoolean) + kryo } def newInstance(): SerializerInstance = { - this.kryo.get().setClassLoader(Thread.currentThread().getContextClassLoader) new KryoSerializerInstance(this) } -} +}
\ No newline at end of file diff --git a/core/src/main/scala/spark/Logging.scala b/core/src/main/scala/spark/Logging.scala index 0fc8c31463..79b0362830 100644 --- a/core/src/main/scala/spark/Logging.scala +++ b/core/src/main/scala/spark/Logging.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.slf4j.Logger diff --git a/core/src/main/scala/spark/MapOutputTracker.scala b/core/src/main/scala/spark/MapOutputTracker.scala index fde597ffd1..2c417e31db 100644 --- a/core/src/main/scala/spark/MapOutputTracker.scala +++ b/core/src/main/scala/spark/MapOutputTracker.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io._ diff --git a/core/src/main/scala/spark/PairRDDFunctions.scala b/core/src/main/scala/spark/PairRDDFunctions.scala index 67fd1c1a8f..6b0cc2fbf1 100644 --- a/core/src/main/scala/spark/PairRDDFunctions.scala +++ b/core/src/main/scala/spark/PairRDDFunctions.scala @@ -1,5 +1,23 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark +import java.nio.ByteBuffer import java.util.{Date, HashMap => JHashMap} import java.text.SimpleDateFormat @@ -10,6 +28,8 @@ import scala.collection.JavaConversions._ import org.apache.hadoop.conf.Configuration import org.apache.hadoop.fs.Path +import org.apache.hadoop.io.compress.CompressionCodec +import org.apache.hadoop.io.SequenceFile.CompressionType import org.apache.hadoop.mapred.FileOutputCommitter import org.apache.hadoop.mapred.FileOutputFormat import org.apache.hadoop.mapred.HadoopWriter @@ -17,7 +37,8 @@ import org.apache.hadoop.mapred.JobConf import org.apache.hadoop.mapred.OutputFormat import org.apache.hadoop.mapreduce.lib.output.{FileOutputFormat => NewFileOutputFormat} -import org.apache.hadoop.mapreduce.{OutputFormat => NewOutputFormat, RecordWriter => NewRecordWriter, Job => NewAPIHadoopJob, HadoopMapReduceUtil, TaskAttemptID, TaskAttemptContext} +import org.apache.hadoop.mapreduce.{OutputFormat => NewOutputFormat, RecordWriter => NewRecordWriter, Job => NewAPIHadoopJob, HadoopMapReduceUtil} +import org.apache.hadoop.security.UserGroupInformation import spark.partial.BoundedDouble import spark.partial.PartialResult @@ -52,7 +73,8 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( mergeValue: (C, V) => C, mergeCombiners: (C, C) => C, partitioner: Partitioner, - mapSideCombine: Boolean = true): RDD[(K, C)] = { + mapSideCombine: Boolean = true, + serializerClass: String = null): RDD[(K, C)] = { if (getKeyClass().isArray) { if (mapSideCombine) { throw new SparkException("Cannot use map-side combining with array keys.") @@ -61,19 +83,18 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( throw new SparkException("Default partitioner cannot partition array keys.") } } - val aggregator = - new Aggregator[K, V, C](createCombiner, mergeValue, mergeCombiners) + val aggregator = new Aggregator[K, V, C](createCombiner, mergeValue, mergeCombiners) if (self.partitioner == Some(partitioner)) { self.mapPartitions(aggregator.combineValuesByKey(_), true) } else if (mapSideCombine) { val mapSideCombined = self.mapPartitions(aggregator.combineValuesByKey(_), true) - val partitioned = new ShuffledRDD[K, C](mapSideCombined, partitioner) + val partitioned = new ShuffledRDD[K, C](mapSideCombined, partitioner, serializerClass) partitioned.mapPartitions(aggregator.combineCombinersByKey(_), true) } else { // Don't apply map-side combiner. // A sanity check to make sure mergeCombiners is not defined. assert(mergeCombiners == null) - val values = new ShuffledRDD[K, V](self, partitioner) + val values = new ShuffledRDD[K, V](self, partitioner, serializerClass) values.mapPartitions(aggregator.combineValuesByKey(_), true) } } @@ -94,7 +115,16 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( * list concatenation, 0 for addition, or 1 for multiplication.). */ def foldByKey(zeroValue: V, partitioner: Partitioner)(func: (V, V) => V): RDD[(K, V)] = { - combineByKey[V]({v: V => func(zeroValue, v)}, func, func, partitioner) + // Serialize the zero value to a byte array so that we can get a new clone of it on each key + val zeroBuffer = SparkEnv.get.closureSerializer.newInstance().serialize(zeroValue) + val zeroArray = new Array[Byte](zeroBuffer.limit) + zeroBuffer.get(zeroArray) + + // When deserializing, use a lazy val to create just one instance of the serializer per task + lazy val cachedSerializer = SparkEnv.get.closureSerializer.newInstance() + def createZero() = cachedSerializer.deserialize[V](ByteBuffer.wrap(zeroArray)) + + combineByKey[V]((v: V) => func(createZero(), v), func, func, partitioner) } /** @@ -184,11 +214,13 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( * partitioning of the resulting key-value pair RDD by passing a Partitioner. */ def groupByKey(partitioner: Partitioner): RDD[(K, Seq[V])] = { + // groupByKey shouldn't use map side combine because map side combine does not + // reduce the amount of data shuffled and requires all map side data be inserted + // into a hash table, leading to more objects in the old gen. def createCombiner(v: V) = ArrayBuffer(v) def mergeValue(buf: ArrayBuffer[V], v: V) = buf += v - def mergeCombiners(b1: ArrayBuffer[V], b2: ArrayBuffer[V]) = b1 ++= b2 val bufs = combineByKey[ArrayBuffer[V]]( - createCombiner _, mergeValue _, mergeCombiners _, partitioner) + createCombiner _, mergeValue _, null, partitioner, mapSideCombine=false) bufs.asInstanceOf[RDD[(K, Seq[V])]] } @@ -469,7 +501,7 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( /** * Return an RDD with the pairs from `this` whose keys are not in `other`. - * + * * Uses `this` partitioner/partition size, because even if `other` is huge, the resulting * RDD will be <= us. */ @@ -515,6 +547,16 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( } /** + * Output the RDD to any Hadoop-supported file system, using a Hadoop `OutputFormat` class + * supporting the key and value types K and V in this RDD. Compress the result with the + * supplied codec. + */ + def saveAsHadoopFile[F <: OutputFormat[K, V]]( + path: String, codec: Class[_ <: CompressionCodec]) (implicit fm: ClassManifest[F]) { + saveAsHadoopFile(path, getKeyClass, getValueClass, fm.erasure.asInstanceOf[Class[F]], codec) + } + + /** * Output the RDD to any Hadoop-supported file system, using a new Hadoop API `OutputFormat` * (mapreduce.OutputFormat) object supporting the key and value types K and V in this RDD. */ @@ -575,6 +617,20 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( /** * Output the RDD to any Hadoop-supported file system, using a Hadoop `OutputFormat` class + * supporting the key and value types K and V in this RDD. Compress with the supplied codec. + */ + def saveAsHadoopFile( + path: String, + keyClass: Class[_], + valueClass: Class[_], + outputFormatClass: Class[_ <: OutputFormat[_, _]], + codec: Class[_ <: CompressionCodec]) { + saveAsHadoopFile(path, keyClass, valueClass, outputFormatClass, + new JobConf(self.context.hadoopConfiguration), Some(codec)) + } + + /** + * Output the RDD to any Hadoop-supported file system, using a Hadoop `OutputFormat` class * supporting the key and value types K and V in this RDD. */ def saveAsHadoopFile( @@ -582,11 +638,19 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( keyClass: Class[_], valueClass: Class[_], outputFormatClass: Class[_ <: OutputFormat[_, _]], - conf: JobConf = new JobConf(self.context.hadoopConfiguration)) { + conf: JobConf = new JobConf(self.context.hadoopConfiguration), + codec: Option[Class[_ <: CompressionCodec]] = None) { conf.setOutputKeyClass(keyClass) conf.setOutputValueClass(valueClass) // conf.setOutputFormat(outputFormatClass) // Doesn't work in Scala 2.9 due to what may be a generics bug conf.set("mapred.output.format.class", outputFormatClass.getName) + for (c <- codec) { + conf.setCompressMapOutput(true) + conf.set("mapred.output.compress", "true") + conf.setMapOutputCompressorClass(c) + conf.set("mapred.output.compression.codec", c.getCanonicalName) + conf.set("mapred.output.compression.type", CompressionType.BLOCK.toString) + } conf.setOutputCommitter(classOf[FileOutputCommitter]) FileOutputFormat.setOutputPath(conf, HadoopWriter.createPathFromString(path, conf)) saveAsHadoopDataset(conf) @@ -645,7 +709,7 @@ class PairRDDFunctions[K: ClassManifest, V: ClassManifest]( * Return an RDD with the keys of each tuple. */ def keys: RDD[K] = self.map(_._1) - + /** * Return an RDD with the values of each tuple. */ diff --git a/core/src/main/scala/spark/Partition.scala b/core/src/main/scala/spark/Partition.scala index e384308ef6..2a4edcec98 100644 --- a/core/src/main/scala/spark/Partition.scala +++ b/core/src/main/scala/spark/Partition.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark /** diff --git a/core/src/main/scala/spark/Partitioner.scala b/core/src/main/scala/spark/Partitioner.scala index 6f8cd17c88..660af70d52 100644 --- a/core/src/main/scala/spark/Partitioner.scala +++ b/core/src/main/scala/spark/Partitioner.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark /** diff --git a/core/src/main/scala/spark/RDD.scala b/core/src/main/scala/spark/RDD.scala index 09e52ebf3e..ca7cdd622a 100644 --- a/core/src/main/scala/spark/RDD.scala +++ b/core/src/main/scala/spark/RDD.scala @@ -1,21 +1,37 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark -import java.net.URL -import java.util.{Date, Random} -import java.util.{HashMap => JHashMap} +import java.util.Random import scala.collection.Map import scala.collection.JavaConversions.mapAsScalaMap import scala.collection.mutable.ArrayBuffer -import scala.collection.mutable.HashMap import org.apache.hadoop.io.BytesWritable +import org.apache.hadoop.io.compress.CompressionCodec import org.apache.hadoop.io.NullWritable import org.apache.hadoop.io.Text import org.apache.hadoop.mapred.TextOutputFormat import it.unimi.dsi.fastutil.objects.{Object2LongOpenHashMap => OLMap} +import spark.broadcast.Broadcast import spark.Partitioner._ import spark.partial.BoundedDouble import spark.partial.CountEvaluator @@ -32,13 +48,13 @@ import spark.rdd.MapPartitionsWithIndexRDD import spark.rdd.PipedRDD import spark.rdd.SampledRDD import spark.rdd.ShuffledRDD -import spark.rdd.SubtractedRDD import spark.rdd.UnionRDD import spark.rdd.ZippedRDD import spark.rdd.ZippedPartitionsRDD2 import spark.rdd.ZippedPartitionsRDD3 import spark.rdd.ZippedPartitionsRDD4 import spark.storage.StorageLevel +import spark.util.BoundedPriorityQueue import SparkContext._ @@ -106,8 +122,11 @@ abstract class RDD[T: ClassManifest]( // Methods and fields available on all RDDs // ======================================================================= + /** The SparkContext that created this RDD. */ + def sparkContext: SparkContext = sc + /** A unique ID for this RDD (within its SparkContext). */ - val id = sc.newRddId() + val id: Int = sc.newRddId() /** A friendly name for this RDD */ var name: String = null @@ -118,9 +137,18 @@ abstract class RDD[T: ClassManifest]( this } + /** User-defined generator of this RDD*/ + var generator = Utils.getCallSiteInfo.firstUserClass + + /** Reset generator*/ + def setGenerator(_generator: String) = { + generator = _generator + } + /** * Set this RDD's storage level to persist its values across operations after the first time - * it is computed. Can only be called once on each RDD. + * it is computed. This can only be used to assign a new storage level if the RDD does not + * have a storage level set yet.. */ def persist(newLevel: StorageLevel): RDD[T] = { // TODO: Handle changes of StorageLevel @@ -140,6 +168,20 @@ abstract class RDD[T: ClassManifest]( /** Persist this RDD with the default storage level (`MEMORY_ONLY`). */ def cache(): RDD[T] = persist() + /** + * Mark the RDD as non-persistent, and remove all blocks for it from memory and disk. + * + * @param blocking Whether to block until all blocks are deleted. + * @return This RDD. + */ + def unpersist(blocking: Boolean = true): RDD[T] = { + logInfo("Removing RDD " + id + " from persistence list") + sc.env.blockManager.master.removeRdd(id, blocking) + sc.persistentRdds.remove(id) + storageLevel = StorageLevel.NONE + this + } + /** Get the RDD's current storage level, or StorageLevel.NONE if none is set. */ def getStorageLevel = storageLevel @@ -260,30 +302,34 @@ abstract class RDD[T: ClassManifest]( var fraction = 0.0 var total = 0 var multiplier = 3.0 - var initialCount = count() + var initialCount = this.count() var maxSelected = 0 + if (num < 0) { + throw new IllegalArgumentException("Negative number of elements requested") + } + if (initialCount > Integer.MAX_VALUE - 1) { maxSelected = Integer.MAX_VALUE - 1 } else { maxSelected = initialCount.toInt } - if (num > initialCount) { + if (num > initialCount && !withReplacement) { total = maxSelected - fraction = math.min(multiplier * (maxSelected + 1) / initialCount, 1.0) - } else if (num < 0) { - throw(new IllegalArgumentException("Negative number of elements requested")) + fraction = multiplier * (maxSelected + 1) / initialCount } else { - fraction = math.min(multiplier * (num + 1) / initialCount, 1.0) + fraction = multiplier * (num + 1) / initialCount total = num } val rand = new Random(seed) - var samples = this.sample(withReplacement, fraction, rand.nextInt).collect() + var samples = this.sample(withReplacement, fraction, rand.nextInt()).collect() + // If the first sample didn't turn out large enough, keep trying to take samples; + // this shouldn't happen often because we use a big multiplier for thei initial size while (samples.length < total) { - samples = this.sample(withReplacement, fraction, rand.nextInt).collect() + samples = this.sample(withReplacement, fraction, rand.nextInt()).collect() } Utils.randomizeInPlace(samples, rand).take(total) @@ -341,13 +387,36 @@ abstract class RDD[T: ClassManifest]( /** * Return an RDD created by piping elements to a forked external process. */ - def pipe(command: Seq[String]): RDD[String] = new PipedRDD(this, command) + def pipe(command: String, env: Map[String, String]): RDD[String] = + new PipedRDD(this, command, env) + /** * Return an RDD created by piping elements to a forked external process. - */ - def pipe(command: Seq[String], env: Map[String, String]): RDD[String] = - new PipedRDD(this, command, env) + * The print behavior can be customized by providing two functions. + * + * @param command command to run in forked process. + * @param env environment variables to set. + * @param printPipeContext Before piping elements, this function is called as an oppotunity + * to pipe context data. Print line function (like out.println) will be + * passed as printPipeContext's parameter. + * @param printRDDElement Use this function to customize how to pipe elements. This function + * will be called with each RDD element as the 1st parameter, and the + * print line function (like out.println()) as the 2nd parameter. + * An example of pipe the RDD data of groupBy() in a streaming way, + * instead of constructing a huge String to concat all the elements: + * def printRDDElement(record:(String, Seq[String]), f:String=>Unit) = + * for (e <- record._2){f(e)} + * @return the result RDD + */ + def pipe( + command: Seq[String], + env: Map[String, String] = Map(), + printPipeContext: (String => Unit) => Unit = null, + printRDDElement: (T, String => Unit) => Unit = null): RDD[String] = + new PipedRDD(this, command, env, + if (printPipeContext ne null) sc.clean(printPipeContext) else null, + if (printRDDElement ne null) sc.clean(printRDDElement) else null) /** * Return a new RDD by applying a function to each partition of this RDD. @@ -479,7 +548,7 @@ abstract class RDD[T: ClassManifest]( */ def foreachPartition(f: Iterator[T] => Unit) { val cleanF = sc.clean(f) - sc.runJob(this, (iter: Iterator[T]) => f(iter)) + sc.runJob(this, (iter: Iterator[T]) => cleanF(iter)) } /** @@ -712,6 +781,34 @@ abstract class RDD[T: ClassManifest]( } /** + * Returns the top K elements from this RDD as defined by + * the specified implicit Ordering[T]. + * @param num the number of top elements to return + * @param ord the implicit ordering for T + * @return an array of top elements + */ + def top(num: Int)(implicit ord: Ordering[T]): Array[T] = { + mapPartitions { items => + val queue = new BoundedPriorityQueue[T](num) + queue ++= items + Iterator.single(queue) + }.reduce { (queue1, queue2) => + queue1 ++= queue2 + queue1 + }.toArray.sorted(ord.reverse) + } + + /** + * Returns the first K elements from this RDD as defined by + * the specified implicit Ordering[T] and maintains the + * ordering. + * @param num the number of top elements to return + * @param ord the implicit ordering for T + * @return an array of top elements + */ + def takeOrdered(num: Int)(implicit ord: Ordering[T]): Array[T] = top(num)(ord.reverse) + + /** * Save this RDD as a text file, using string representations of elements. */ def saveAsTextFile(path: String) { @@ -720,6 +817,14 @@ abstract class RDD[T: ClassManifest]( } /** + * Save this RDD as a compressed text file, using string representations of elements. + */ + def saveAsTextFile(path: String, codec: Class[_ <: CompressionCodec]) { + this.map(x => (NullWritable.get(), new Text(x.toString))) + .saveAsHadoopFile[TextOutputFormat[NullWritable, Text]](path, codec) + } + + /** * Save this RDD as a SequenceFile of serialized objects. */ def saveAsObjectFile(path: String) { @@ -777,7 +882,7 @@ abstract class RDD[T: ClassManifest]( private var storageLevel: StorageLevel = StorageLevel.NONE /** Record user function generating this RDD. */ - private[spark] val origin = Utils.getSparkCallSite + private[spark] val origin = Utils.formatSparkCallSite private[spark] def elementClassManifest: ClassManifest[T] = classManifest[T] diff --git a/core/src/main/scala/spark/RDDCheckpointData.scala b/core/src/main/scala/spark/RDDCheckpointData.scala index 57e0405fb4..b615f820eb 100644 --- a/core/src/main/scala/spark/RDDCheckpointData.scala +++ b/core/src/main/scala/spark/RDDCheckpointData.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.apache.hadoop.fs.Path diff --git a/core/src/main/scala/spark/SequenceFileRDDFunctions.scala b/core/src/main/scala/spark/SequenceFileRDDFunctions.scala index 518034e07b..9f30b7f22f 100644 --- a/core/src/main/scala/spark/SequenceFileRDDFunctions.scala +++ b/core/src/main/scala/spark/SequenceFileRDDFunctions.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io.EOFException @@ -18,6 +35,7 @@ import org.apache.hadoop.mapred.TextOutputFormat import org.apache.hadoop.mapred.SequenceFileOutputFormat import org.apache.hadoop.mapred.OutputCommitter import org.apache.hadoop.mapred.FileOutputCommitter +import org.apache.hadoop.io.compress.CompressionCodec import org.apache.hadoop.io.Writable import org.apache.hadoop.io.NullWritable import org.apache.hadoop.io.BytesWritable @@ -62,7 +80,7 @@ class SequenceFileRDDFunctions[K <% Writable: ClassManifest, V <% Writable : Cla * byte arrays to BytesWritable, and Strings to Text. The `path` can be on any Hadoop-supported * file system. */ - def saveAsSequenceFile(path: String) { + def saveAsSequenceFile(path: String, codec: Option[Class[_ <: CompressionCodec]] = None) { def anyToWritable[U <% Writable](u: U): Writable = u val keyClass = getWritableClass[K] @@ -72,14 +90,18 @@ class SequenceFileRDDFunctions[K <% Writable: ClassManifest, V <% Writable : Cla logInfo("Saving as sequence file of type (" + keyClass.getSimpleName + "," + valueClass.getSimpleName + ")" ) val format = classOf[SequenceFileOutputFormat[Writable, Writable]] + val jobConf = new JobConf(self.context.hadoopConfiguration) if (!convertKey && !convertValue) { - self.saveAsHadoopFile(path, keyClass, valueClass, format) + self.saveAsHadoopFile(path, keyClass, valueClass, format, jobConf, codec) } else if (!convertKey && convertValue) { - self.map(x => (x._1,anyToWritable(x._2))).saveAsHadoopFile(path, keyClass, valueClass, format) + self.map(x => (x._1,anyToWritable(x._2))).saveAsHadoopFile( + path, keyClass, valueClass, format, jobConf, codec) } else if (convertKey && !convertValue) { - self.map(x => (anyToWritable(x._1),x._2)).saveAsHadoopFile(path, keyClass, valueClass, format) + self.map(x => (anyToWritable(x._1),x._2)).saveAsHadoopFile( + path, keyClass, valueClass, format, jobConf, codec) } else if (convertKey && convertValue) { - self.map(x => (anyToWritable(x._1),anyToWritable(x._2))).saveAsHadoopFile(path, keyClass, valueClass, format) + self.map(x => (anyToWritable(x._1),anyToWritable(x._2))).saveAsHadoopFile( + path, keyClass, valueClass, format, jobConf, codec) } } } diff --git a/core/src/main/scala/spark/SerializableWritable.scala b/core/src/main/scala/spark/SerializableWritable.scala index 8306fbf570..0236611ef9 100644 --- a/core/src/main/scala/spark/SerializableWritable.scala +++ b/core/src/main/scala/spark/SerializableWritable.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io._ diff --git a/core/src/main/scala/spark/ShuffleFetcher.scala b/core/src/main/scala/spark/ShuffleFetcher.scala index 442e9f0269..dcced035e7 100644 --- a/core/src/main/scala/spark/ShuffleFetcher.scala +++ b/core/src/main/scala/spark/ShuffleFetcher.scala @@ -1,13 +1,33 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark -import executor.TaskMetrics +import spark.executor.TaskMetrics +import spark.serializer.Serializer + private[spark] abstract class ShuffleFetcher { /** * Fetch the shuffle outputs for a given ShuffleDependency. * @return An iterator over the elements of the fetched shuffle outputs. */ - def fetch[K, V](shuffleId: Int, reduceId: Int, metrics: TaskMetrics) : Iterator[(K,V)] + def fetch[K, V](shuffleId: Int, reduceId: Int, metrics: TaskMetrics, + serializer: Serializer = SparkEnv.get.serializerManager.default): Iterator[(K,V)] /** Stop the fetcher */ def stop() {} diff --git a/core/src/main/scala/spark/SizeEstimator.scala b/core/src/main/scala/spark/SizeEstimator.scala index d4e1157250..6cc57566d7 100644 --- a/core/src/main/scala/spark/SizeEstimator.scala +++ b/core/src/main/scala/spark/SizeEstimator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.lang.reflect.Field @@ -198,7 +215,7 @@ private[spark] object SizeEstimator extends Logging { val elem = JArray.get(array, index) size += SizeEstimator.estimate(elem, state.visited) } - state.size += ((length / 100.0) * size).toLong + state.size += ((length / (ARRAY_SAMPLE_SIZE * 1.0)) * size).toLong } } } diff --git a/core/src/main/scala/spark/SoftReferenceCache.scala b/core/src/main/scala/spark/SoftReferenceCache.scala index 3dd0a4b1f9..f41a379582 100644 --- a/core/src/main/scala/spark/SoftReferenceCache.scala +++ b/core/src/main/scala/spark/SoftReferenceCache.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import com.google.common.collect.MapMaker diff --git a/core/src/main/scala/spark/SparkContext.scala b/core/src/main/scala/spark/SparkContext.scala index 5f5ec0b0f4..77cb0ee0cd 100644 --- a/core/src/main/scala/spark/SparkContext.scala +++ b/core/src/main/scala/spark/SparkContext.scala @@ -1,47 +1,73 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io._ -import java.util.concurrent.atomic.AtomicInteger import java.net.URI +import java.util.Properties +import java.util.concurrent.ConcurrentHashMap +import java.util.concurrent.atomic.AtomicInteger +import scala.collection.JavaConversions._ import scala.collection.Map import scala.collection.generic.Growable import scala.collection.mutable.HashMap import scala.collection.JavaConversions._ +import scala.util.DynamicVariable +import scala.collection.mutable.{ConcurrentMap, HashMap} + +import akka.actor.Actor._ -import org.apache.hadoop.fs.Path import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.mapred.InputFormat -import org.apache.hadoop.mapred.SequenceFileInputFormat -import org.apache.hadoop.io.Writable -import org.apache.hadoop.io.IntWritable -import org.apache.hadoop.io.LongWritable -import org.apache.hadoop.io.FloatWritable -import org.apache.hadoop.io.DoubleWritable +import org.apache.hadoop.fs.Path +import org.apache.hadoop.io.ArrayWritable import org.apache.hadoop.io.BooleanWritable import org.apache.hadoop.io.BytesWritable -import org.apache.hadoop.io.ArrayWritable +import org.apache.hadoop.io.DoubleWritable +import org.apache.hadoop.io.FloatWritable +import org.apache.hadoop.io.IntWritable +import org.apache.hadoop.io.LongWritable import org.apache.hadoop.io.NullWritable import org.apache.hadoop.io.Text +import org.apache.hadoop.io.Writable import org.apache.hadoop.mapred.FileInputFormat +import org.apache.hadoop.mapred.InputFormat import org.apache.hadoop.mapred.JobConf +import org.apache.hadoop.mapred.SequenceFileInputFormat import org.apache.hadoop.mapred.TextInputFormat import org.apache.hadoop.mapreduce.{InputFormat => NewInputFormat} -import org.apache.hadoop.mapreduce.lib.input.{FileInputFormat => NewFileInputFormat} import org.apache.hadoop.mapreduce.{Job => NewHadoopJob} +import org.apache.hadoop.mapreduce.lib.input.{FileInputFormat => NewFileInputFormat} +import org.apache.hadoop.security.UserGroupInformation + import org.apache.mesos.MesosNativeLibrary -import spark.deploy.{SparkHadoopUtil, LocalSparkCluster} -import spark.partial.ApproximateEvaluator -import spark.partial.PartialResult +import spark.deploy.{LocalSparkCluster, SparkHadoopUtil} +import spark.partial.{ApproximateEvaluator, PartialResult} import spark.rdd.{CheckpointRDD, HadoopRDD, NewHadoopRDD, UnionRDD, ParallelCollectionRDD} -import spark.scheduler._ +import spark.scheduler.{DAGScheduler, DAGSchedulerSource, ResultTask, ShuffleMapTask, SparkListener, SplitInfo, Stage, StageInfo, TaskScheduler} +import spark.scheduler.cluster.{StandaloneSchedulerBackend, SparkDeploySchedulerBackend, ClusterScheduler} import spark.scheduler.local.LocalScheduler -import spark.scheduler.cluster.{StandaloneSchedulerBackend, SparkDeploySchedulerBackend, SchedulerBackend, ClusterScheduler} import spark.scheduler.mesos.{CoarseMesosSchedulerBackend, MesosSchedulerBackend} -import spark.storage.BlockManagerUI +import spark.storage.{StorageStatus, StorageUtils, RDDInfo, BlockManagerSource} import spark.util.{MetadataCleaner, TimeStampedHashMap} -import spark.storage.{StorageStatus, StorageUtils, RDDInfo} +import ui.{SparkUI} +import spark.metrics._ /** * Main entry point for Spark functionality. A SparkContext represents the connection to a Spark @@ -87,19 +113,17 @@ class SparkContext( isLocal) SparkEnv.set(env) - // Start the BlockManager UI - private[spark] val ui = new BlockManagerUI( - env.actorSystem, env.blockManager.master.driverActor, this) - ui.start() - // Used to store a URL for each static file/jar together with the file's local timestamp private[spark] val addedFiles = HashMap[String, Long]() private[spark] val addedJars = HashMap[String, Long]() // Keeps track of all persisted RDDs - private[spark] val persistentRdds = new TimeStampedHashMap[Int, RDD[_]]() + private[spark] val persistentRdds = new TimeStampedHashMap[Int, RDD[_]] private[spark] val metadataCleaner = new MetadataCleaner("SparkContext", this.cleanup) + // Initalize the Spark UI + private[spark] val ui = new SparkUI(this) + ui.bind() // Add each JAR given through the constructor if (jars != null) { @@ -109,13 +133,14 @@ class SparkContext( // Environment variables to pass to our executors private[spark] val executorEnvs = HashMap[String, String]() // Note: SPARK_MEM is included for Mesos, but overwritten for standalone mode in ExecutorRunner - for (key <- Seq("SPARK_MEM", "SPARK_CLASSPATH", "SPARK_LIBRARY_PATH", "SPARK_JAVA_OPTS", - "SPARK_TESTING")) { + for (key <- Seq("SPARK_CLASSPATH", "SPARK_LIBRARY_PATH", "SPARK_JAVA_OPTS", "SPARK_TESTING")) { val value = System.getenv(key) if (value != null) { executorEnvs(key) = value } } + // Since memory can be set with a system property too, use that + executorEnvs("SPARK_MEM") = SparkContext.executorMemoryRequested + "m" if (environment != null) { executorEnvs ++= environment } @@ -150,14 +175,12 @@ class SparkContext( scheduler case LOCAL_CLUSTER_REGEX(numSlaves, coresPerSlave, memoryPerSlave) => - // Check to make sure SPARK_MEM <= memoryPerSlave. Otherwise Spark will just hang. + // Check to make sure memory requested <= memoryPerSlave. Otherwise Spark will just hang. val memoryPerSlaveInt = memoryPerSlave.toInt - val sparkMemEnv = System.getenv("SPARK_MEM") - val sparkMemEnvInt = if (sparkMemEnv != null) Utils.memoryStringToMb(sparkMemEnv) else 512 - if (sparkMemEnvInt > memoryPerSlaveInt) { + if (SparkContext.executorMemoryRequested > memoryPerSlaveInt) { throw new SparkException( - "Slave memory (%d MB) cannot be smaller than SPARK_MEM (%d MB)".format( - memoryPerSlaveInt, sparkMemEnvInt)) + "Asked to launch cluster with %d MB RAM / worker but requested %d MB/worker".format( + memoryPerSlaveInt, SparkContext.executorMemoryRequested)) } val scheduler = new ClusterScheduler(this) @@ -209,6 +232,8 @@ class SparkContext( @volatile private var dagScheduler = new DAGScheduler(taskScheduler) dagScheduler.start() + ui.start() + /** A default Hadoop Configuration for the Hadoop code (e.g. file systems) that we reuse. */ val hadoopConfiguration = { val conf = SparkHadoopUtil.newConfiguration() @@ -230,9 +255,32 @@ class SparkContext( private[spark] var checkpointDir: Option[String] = None + // Thread Local variable that can be used by users to pass information down the stack + private val localProperties = new DynamicVariable[Properties](null) + + def initLocalProperties() { + localProperties.value = new Properties() + } + + def addLocalProperties(key: String, value: String) { + if(localProperties.value == null) { + localProperties.value = new Properties() + } + localProperties.value.setProperty(key,value) + } // Post init taskScheduler.postStartHook() + val dagSchedulerSource = new DAGSchedulerSource(this.dagScheduler) + val blockManagerSource = new BlockManagerSource(SparkEnv.get.blockManager) + + def initDriverMetrics() { + SparkEnv.get.metricsSystem.registerSource(dagSchedulerSource) + SparkEnv.get.metricsSystem.registerSource(blockManagerSource) + } + + initDriverMetrics() + // Methods for creating RDDs /** Distribute a local Scala collection to form an RDD. */ @@ -505,10 +553,16 @@ class SparkContext( * Return information about what RDDs are cached, if they are in mem or on disk, how much space * they take, etc. */ - def getRDDStorageInfo : Array[RDDInfo] = { + def getRDDStorageInfo: Array[RDDInfo] = { StorageUtils.rddInfoFromStorageStatus(getExecutorStorageStatus, this) } + /** + * Returns an immutable map of RDDs that have marked themselves as persistent via cache() call. + * Note that this does not necessarily mean the caching or computation was successful. + */ + def getPersistentRDDs: Map[Int, RDD[_]] = persistentRdds.toMap + def getStageInfo: Map[Stage,StageInfo] = { dagScheduler.stageToInfos } @@ -516,7 +570,7 @@ class SparkContext( /** * Return information about blocks stored in all of the slaves */ - def getExecutorStorageStatus : Array[StorageStatus] = { + def getExecutorStorageStatus: Array[StorageStatus] = { env.blockManager.master.getStorageStatus } @@ -534,13 +588,23 @@ class SparkContext( * filesystems), or an HTTP, HTTPS or FTP URI. */ def addJar(path: String) { - val uri = new URI(path) - val key = uri.getScheme match { - case null | "file" => env.httpFileServer.addJar(new File(uri.getPath)) - case _ => path + if (null == path) { + logWarning("null specified as parameter to addJar", + new SparkException("null specified as parameter to addJar")) + } else { + val uri = new URI(path) + val key = uri.getScheme match { + case null | "file" => + if (SparkHadoopUtil.isYarnMode()) { + logWarning("local jar specified as parameter to addJar under Yarn mode") + return + } + env.httpFileServer.addJar(new File(uri.getPath)) + case _ => path + } + addedJars(key) = System.currentTimeMillis + logInfo("Added JAR " + path + " at " + key + " with timestamp " + addedJars(key)) } - addedJars(key) = System.currentTimeMillis - logInfo("Added JAR " + path + " at " + key + " with timestamp " + addedJars(key)) } /** @@ -553,6 +617,7 @@ class SparkContext( /** Shut down the SparkContext. */ def stop() { + ui.stop() // Do this only if not stopped already - best case effort. // prevent NPE if stopped more than once. val dagSchedulerCopy = dagScheduler @@ -605,10 +670,10 @@ class SparkContext( partitions: Seq[Int], allowLocal: Boolean, resultHandler: (Int, U) => Unit) { - val callSite = Utils.getSparkCallSite + val callSite = Utils.formatSparkCallSite logInfo("Starting job: " + callSite) val start = System.nanoTime - val result = dagScheduler.runJob(rdd, func, partitions, callSite, allowLocal, resultHandler) + val result = dagScheduler.runJob(rdd, func, partitions, callSite, allowLocal, resultHandler, localProperties.value) logInfo("Job finished: " + callSite + ", took " + (System.nanoTime - start) / 1e9 + " s") rdd.doCheckpoint() result @@ -687,12 +752,11 @@ class SparkContext( rdd: RDD[T], func: (TaskContext, Iterator[T]) => U, evaluator: ApproximateEvaluator[U, R], - timeout: Long - ): PartialResult[R] = { - val callSite = Utils.getSparkCallSite + timeout: Long): PartialResult[R] = { + val callSite = Utils.formatSparkCallSite logInfo("Starting job: " + callSite) val start = System.nanoTime - val result = dagScheduler.runApproximateJob(rdd, func, evaluator, callSite, timeout) + val result = dagScheduler.runApproximateJob(rdd, func, evaluator, callSite, timeout, localProperties.value) logInfo("Job finished: " + callSite + ", took " + (System.nanoTime - start) / 1e9 + " s") result } @@ -858,6 +922,15 @@ object SparkContext { /** Find the JAR that contains the class of a particular object */ def jarOfObject(obj: AnyRef): Seq[String] = jarOfClass(obj.getClass) + + /** Get the amount of memory per executor requested through system properties or SPARK_MEM */ + private[spark] val executorMemoryRequested = { + // TODO: Might need to add some extra memory for the non-heap parts of the JVM + Option(System.getProperty("spark.executor.memory")) + .orElse(Option(System.getenv("SPARK_MEM"))) + .map(Utils.memoryStringToMb) + .getOrElse(512) + } } diff --git a/core/src/main/scala/spark/SparkEnv.scala b/core/src/main/scala/spark/SparkEnv.scala index ffb40bab3a..4a1d341f5d 100644 --- a/core/src/main/scala/spark/SparkEnv.scala +++ b/core/src/main/scala/spark/SparkEnv.scala @@ -1,14 +1,37 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark +import collection.mutable +import serializer.Serializer + import akka.actor.{Actor, ActorRef, Props, ActorSystemImpl, ActorSystem} import akka.remote.RemoteActorRefProvider -import serializer.Serializer import spark.broadcast.BroadcastManager +import spark.metrics.MetricsSystem import spark.storage.BlockManager import spark.storage.BlockManagerMaster import spark.network.ConnectionManager +import spark.serializer.{Serializer, SerializerManager} import spark.util.AkkaUtils +import spark.api.python.PythonWorkerFactory + /** * Holds all the runtime environment objects for a running Spark instance (either master or worker), @@ -20,6 +43,7 @@ import spark.util.AkkaUtils class SparkEnv ( val executorId: String, val actorSystem: ActorSystem, + val serializerManager: SerializerManager, val serializer: Serializer, val closureSerializer: Serializer, val cacheManager: CacheManager, @@ -29,21 +53,46 @@ class SparkEnv ( val blockManager: BlockManager, val connectionManager: ConnectionManager, val httpFileServer: HttpFileServer, - val sparkFilesDir: String - ) { + val sparkFilesDir: String, + val metricsSystem: MetricsSystem, + // To be set only as part of initialization of SparkContext. + // (executorId, defaultHostPort) => executorHostPort + // If executorId is NOT found, return defaultHostPort + var executorIdToHostPort: Option[(String, String) => String]) { + + private val pythonWorkers = mutable.HashMap[(String, Map[String, String]), PythonWorkerFactory]() def stop() { + pythonWorkers.foreach { case(key, worker) => worker.stop() } httpFileServer.stop() mapOutputTracker.stop() shuffleFetcher.stop() broadcastManager.stop() blockManager.stop() blockManager.master.stop() + metricsSystem.stop() actorSystem.shutdown() // Unfortunately Akka's awaitTermination doesn't actually wait for the Netty server to shut // down, but let's call it anyway in case it gets fixed in a later release actorSystem.awaitTermination() } + + def createPythonWorker(pythonExec: String, envVars: Map[String, String]): java.net.Socket = { + synchronized { + val key = (pythonExec, envVars) + pythonWorkers.getOrElseUpdate(key, new PythonWorkerFactory(pythonExec, envVars)).create() + } + } + + def resolveExecutorIdToHostPort(executorId: String, defaultHostPort: String): String = { + val env = SparkEnv.get + if (env.executorIdToHostPort.isEmpty) { + // default to using host, not host port. Relevant to non cluster modes. + return defaultHostPort + } + + env.executorIdToHostPort.get(executorId, defaultHostPort) + } } object SparkEnv extends Logging { @@ -91,8 +140,14 @@ object SparkEnv extends Logging { Class.forName(name, true, classLoader).newInstance().asInstanceOf[T] } - val serializer = instantiateClass[Serializer]("spark.serializer", "spark.JavaSerializer") - + val serializerManager = new SerializerManager + + val serializer = serializerManager.setDefault( + System.getProperty("spark.serializer", "spark.JavaSerializer")) + + val closureSerializer = serializerManager.get( + System.getProperty("spark.closure.serializer", "spark.JavaSerializer")) + def registerOrLookup(name: String, newActor: => Actor): ActorRef = { if (isDriver) { logInfo("Registering " + name) @@ -116,9 +171,6 @@ object SparkEnv extends Logging { val broadcastManager = new BroadcastManager(isDriver) - val closureSerializer = instantiateClass[Serializer]( - "spark.closure.serializer", "spark.JavaSerializer") - val cacheManager = new CacheManager(blockManager) // Have to assign trackerActor after initialization as MapOutputTrackerActor @@ -135,6 +187,13 @@ object SparkEnv extends Logging { httpFileServer.initialize() System.setProperty("spark.fileserver.uri", httpFileServer.serverUri) + val metricsSystem = if (isDriver) { + MetricsSystem.createMetricsSystem("driver") + } else { + MetricsSystem.createMetricsSystem("executor") + } + metricsSystem.start() + // Set the sparkFiles directory, used when downloading dependencies. In local mode, // this is a temporary directory; in distributed mode, this is the executor's current working // directory. @@ -153,6 +212,7 @@ object SparkEnv extends Logging { new SparkEnv( executorId, actorSystem, + serializerManager, serializer, closureSerializer, cacheManager, @@ -162,7 +222,8 @@ object SparkEnv extends Logging { blockManager, connectionManager, httpFileServer, - sparkFilesDir) + sparkFilesDir, + metricsSystem, + None) } - } diff --git a/core/src/main/scala/spark/SparkException.scala b/core/src/main/scala/spark/SparkException.scala index aa7a16d7dd..b7045eea63 100644 --- a/core/src/main/scala/spark/SparkException.scala +++ b/core/src/main/scala/spark/SparkException.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark class SparkException(message: String, cause: Throwable) diff --git a/core/src/main/scala/spark/SparkFiles.java b/core/src/main/scala/spark/SparkFiles.java index 566aec622c..f9b3f7965e 100644 --- a/core/src/main/scala/spark/SparkFiles.java +++ b/core/src/main/scala/spark/SparkFiles.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark; import java.io.File; diff --git a/core/src/main/scala/spark/TaskContext.scala b/core/src/main/scala/spark/TaskContext.scala index dd0609026a..b79f4ca813 100644 --- a/core/src/main/scala/spark/TaskContext.scala +++ b/core/src/main/scala/spark/TaskContext.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import executor.TaskMetrics diff --git a/core/src/main/scala/spark/TaskEndReason.scala b/core/src/main/scala/spark/TaskEndReason.scala index 420c54bc9a..3ad665da34 100644 --- a/core/src/main/scala/spark/TaskEndReason.scala +++ b/core/src/main/scala/spark/TaskEndReason.scala @@ -1,5 +1,23 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark +import spark.executor.TaskMetrics import spark.storage.BlockManagerId /** @@ -14,9 +32,20 @@ private[spark] case object Success extends TaskEndReason private[spark] case object Resubmitted extends TaskEndReason // Task was finished earlier but we've now lost it -private[spark] -case class FetchFailed(bmAddress: BlockManagerId, shuffleId: Int, mapId: Int, reduceId: Int) extends TaskEndReason +private[spark] case class FetchFailed( + bmAddress: BlockManagerId, + shuffleId: Int, + mapId: Int, + reduceId: Int) + extends TaskEndReason -private[spark] case class ExceptionFailure(exception: Throwable) extends TaskEndReason +private[spark] case class ExceptionFailure( + className: String, + description: String, + stackTrace: Array[StackTraceElement], + metrics: Option[TaskMetrics]) + extends TaskEndReason private[spark] case class OtherFailure(message: String) extends TaskEndReason + +private[spark] case class TaskResultTooBigFailure() extends TaskEndReason diff --git a/core/src/main/scala/spark/TaskState.scala b/core/src/main/scala/spark/TaskState.scala index 78eb33a628..9df7d8277b 100644 --- a/core/src/main/scala/spark/TaskState.scala +++ b/core/src/main/scala/spark/TaskState.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.apache.mesos.Protos.{TaskState => MesosTaskState} diff --git a/core/src/main/scala/spark/Utils.scala b/core/src/main/scala/spark/Utils.scala index 9f48cbe490..e6a96a5ec1 100644 --- a/core/src/main/scala/spark/Utils.scala +++ b/core/src/main/scala/spark/Utils.scala @@ -1,23 +1,47 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io._ import java.net.{InetAddress, URL, URI, NetworkInterface, Inet4Address, ServerSocket} import java.util.{Locale, Random, UUID} import java.util.concurrent.{ConcurrentHashMap, Executors, ThreadFactory, ThreadPoolExecutor} -import org.apache.hadoop.fs.{Path, FileSystem, FileUtil} +import java.util.regex.Pattern + +import scala.collection.Map import scala.collection.mutable.{ArrayBuffer, HashMap} import scala.collection.JavaConversions._ import scala.io.Source + import com.google.common.io.Files import com.google.common.util.concurrent.ThreadFactoryBuilder + +import org.apache.hadoop.fs.{Path, FileSystem, FileUtil} + import spark.serializer.SerializerInstance import spark.deploy.SparkHadoopUtil -import java.util.regex.Pattern + /** * Various utility methods used by Spark. */ private object Utils extends Logging { + /** Serialize an object using Java serialization */ def serialize[T](o: T): Array[Byte] = { val bos = new ByteArrayOutputStream() @@ -68,7 +92,6 @@ private object Utils extends Logging { return buf } - private val shutdownDeletePaths = new collection.mutable.HashSet[String]() // Register the path to be deleted via shutdown hook @@ -87,19 +110,19 @@ private object Utils extends Logging { } } - // Note: if file is child of some registered path, while not equal to it, then return true; else false - // This is to ensure that two shutdown hooks do not try to delete each others paths - resulting in IOException - // and incomplete cleanup + // Note: if file is child of some registered path, while not equal to it, then return true; + // else false. This is to ensure that two shutdown hooks do not try to delete each others + // paths - resulting in IOException and incomplete cleanup. def hasRootAsShutdownDeleteDir(file: File): Boolean = { - val absolutePath = file.getAbsolutePath() - val retval = shutdownDeletePaths.synchronized { - shutdownDeletePaths.find(path => ! absolutePath.equals(path) && absolutePath.startsWith(path) ).isDefined + shutdownDeletePaths.find { path => + !absolutePath.equals(path) && absolutePath.startsWith(path) + }.isDefined + } + if (retval) { + logInfo("path = " + file + ", already present as root for deletion.") } - - if (retval) logInfo("path = " + file + ", already present as root for deletion.") - retval } @@ -111,8 +134,8 @@ private object Utils extends Logging { while (dir == null) { attempts += 1 if (attempts > maxAttempts) { - throw new IOException("Failed to create a temp directory after " + maxAttempts + - " attempts!") + throw new IOException("Failed to create a temp directory (under " + root + ") after " + + maxAttempts + " attempts!") } try { dir = new File(root, "spark-" + UUID.randomUUID.toString) @@ -131,7 +154,7 @@ private object Utils extends Logging { if (! hasRootAsShutdownDeleteDir(dir)) Utils.deleteRecursively(dir) } }) - return dir + dir } /** Copy all data from an InputStream to an OutputStream */ @@ -174,35 +197,30 @@ private object Utils extends Logging { Utils.copyStream(in, out, true) if (targetFile.exists && !Files.equal(tempFile, targetFile)) { tempFile.delete() - throw new SparkException("File " + targetFile + " exists and does not match contents of" + - " " + url) + throw new SparkException( + "File " + targetFile + " exists and does not match contents of" + " " + url) } else { Files.move(tempFile, targetFile) } case "file" | null => - val sourceFile = if (uri.isAbsolute) { - new File(uri) - } else { - new File(url) - } - if (targetFile.exists && !Files.equal(sourceFile, targetFile)) { - throw new SparkException("File " + targetFile + " exists and does not match contents of" + - " " + url) - } else { - // Remove the file if it already exists - targetFile.delete() - // Symlink the file locally. - if (uri.isAbsolute) { - // url is absolute, i.e. it starts with "file:///". Extract the source - // file's absolute path from the url. - val sourceFile = new File(uri) - logInfo("Symlinking " + sourceFile.getAbsolutePath + " to " + targetFile.getAbsolutePath) - FileUtil.symLink(sourceFile.getAbsolutePath, targetFile.getAbsolutePath) + // In the case of a local file, copy the local file to the target directory. + // Note the difference between uri vs url. + val sourceFile = if (uri.isAbsolute) new File(uri) else new File(url) + if (targetFile.exists) { + // If the target file already exists, warn the user if + if (!Files.equal(sourceFile, targetFile)) { + throw new SparkException( + "File " + targetFile + " exists and does not match contents of" + " " + url) } else { - // url is not absolute, i.e. itself is the path to the source file. - logInfo("Symlinking " + url + " to " + targetFile.getAbsolutePath) - FileUtil.symLink(url, targetFile.getAbsolutePath) + // Do nothing if the file contents are the same, i.e. this file has been copied + // previously. + logInfo(sourceFile.getAbsolutePath + " has been previously copied to " + + targetFile.getAbsolutePath) } + } else { + // The file does not exist in the target directory. Copy it there. + logInfo("Copying " + sourceFile.getAbsolutePath + " to " + targetFile.getAbsolutePath) + Files.copy(sourceFile, targetFile) } case _ => // Use the Hadoop filesystem library, which supports file://, hdfs://, s3://, and others @@ -323,8 +341,6 @@ private object Utils extends Logging { InetAddress.getByName(address).getHostName } - - def localHostPort(): String = { val retval = System.getProperty("spark.hostPort", null) if (retval == null) { @@ -335,7 +351,7 @@ private object Utils extends Logging { retval } - /* +/* // Used by DEBUG code : remove when all testing done private val ipPattern = Pattern.compile("^[0-9]+(\\.[0-9]+)*$") def checkHost(host: String, message: String = "") { @@ -357,26 +373,32 @@ private object Utils extends Logging { Utils.logErrorWithStack("Unexpected to have port " + port + " which is not valid in " + hostPort + ". Message " + message) } } - */ + + // Used by DEBUG code : remove when all testing done + def logErrorWithStack(msg: String) { + try { throw new Exception } catch { case ex: Exception => { logError(msg, ex) } } + // temp code for debug + System.exit(-1) + } +*/ // Once testing is complete in various modes, replace with this ? def checkHost(host: String, message: String = "") {} def checkHostPort(hostPort: String, message: String = "") {} - def getUserNameFromEnvironment(): String = { - SparkHadoopUtil.getUserNameFromEnvironment - } - // Used by DEBUG code : remove when all testing done def logErrorWithStack(msg: String) { try { throw new Exception } catch { case ex: Exception => { logError(msg, ex) } } - // temp code for debug - System.exit(-1) + } + + def getUserNameFromEnvironment(): String = { + SparkHadoopUtil.getUserNameFromEnvironment } // Typically, this will be of order of number of nodes in cluster // If not, we should change it to LRUCache or something. private val hostPortParseResults = new ConcurrentHashMap[String, (String, Int)]() + def parseHostPort(hostPort: String): (String, Int) = { { // Check cache first. @@ -385,7 +407,8 @@ private object Utils extends Logging { } val indx: Int = hostPort.lastIndexOf(':') - // This is potentially broken - when dealing with ipv6 addresses for example, sigh ... but then hadoop does not support ipv6 right now. + // This is potentially broken - when dealing with ipv6 addresses for example, sigh ... + // but then hadoop does not support ipv6 right now. // For now, we assume that if port exists, then it is valid - not check if it is an int > 0 if (-1 == indx) { val retval = (hostPort, 0) @@ -398,17 +421,6 @@ private object Utils extends Logging { hostPortParseResults.get(hostPort) } - def addIfNoPort(hostPort: String, port: Int): String = { - if (port <= 0) throw new IllegalArgumentException("Invalid port specified " + port) - - // This is potentially broken - when dealing with ipv6 addresses for example, sigh ... but then hadoop does not support ipv6 right now. - // For now, we assume that if port exists, then it is valid - not check if it is an int > 0 - val indx: Int = hostPort.lastIndexOf(':') - if (-1 != indx) return hostPort - - hostPort + ":" + port - } - private[spark] val daemonThreadFactory: ThreadFactory = new ThreadFactoryBuilder().setDaemon(true).build() @@ -492,6 +504,26 @@ private object Utils extends Logging { } /** + * Returns a human-readable string representing a duration such as "35ms" + */ + def msDurationToString(ms: Long): String = { + val second = 1000 + val minute = 60 * second + val hour = 60 * minute + + ms match { + case t if t < second => + "%d ms".format(t) + case t if t < minute => + "%.1f s".format(t.toFloat / second) + case t if t < hour => + "%.1f m".format(t.toFloat / minute) + case t => + "%.2f h".format(t.toFloat / hour) + } + } + + /** * Convert a memory quantity in megabytes to a human-readable string such as "4.0 MB". */ def memoryMegabytesToString(megabytes: Long): String = { @@ -528,13 +560,57 @@ private object Utils extends Logging { execute(command, new File(".")) } + /** + * Execute a command and get its output, throwing an exception if it yields a code other than 0. + */ + def executeAndGetOutput(command: Seq[String], workingDir: File = new File("."), + extraEnvironment: Map[String, String] = Map.empty): String = { + val builder = new ProcessBuilder(command: _*) + .directory(workingDir) + val environment = builder.environment() + for ((key, value) <- extraEnvironment) { + environment.put(key, value) + } + val process = builder.start() + new Thread("read stderr for " + command(0)) { + override def run() { + for (line <- Source.fromInputStream(process.getErrorStream).getLines) { + System.err.println(line) + } + } + }.start() + val output = new StringBuffer + val stdoutThread = new Thread("read stdout for " + command(0)) { + override def run() { + for (line <- Source.fromInputStream(process.getInputStream).getLines) { + output.append(line) + } + } + } + stdoutThread.start() + val exitCode = process.waitFor() + stdoutThread.join() // Wait for it to finish reading output + if (exitCode != 0) { + throw new SparkException("Process " + command + " exited with code " + exitCode) + } + output.toString + } + + /** + * A regular expression to match classes of the "core" Spark API that we want to skip when + * finding the call site of a method. + */ + private val SPARK_CLASS_REGEX = """^spark(\.api\.java)?(\.rdd)?\.[A-Z]""".r + + private[spark] class CallSiteInfo(val lastSparkMethod: String, val firstUserFile: String, + val firstUserLine: Int, val firstUserClass: String) /** * When called inside a class in the spark package, returns the name of the user code class * (outside the spark package) that called into Spark, as well as which Spark method they called. * This is used, for example, to tell users where in their code each RDD got created. */ - def getSparkCallSite: String = { + def getCallSiteInfo: CallSiteInfo = { val trace = Thread.currentThread.getStackTrace().filter( el => (!el.getMethodName.contains("getStackTrace"))) @@ -546,10 +622,11 @@ private object Utils extends Logging { var firstUserFile = "<unknown>" var firstUserLine = 0 var finished = false + var firstUserClass = "<unknown>" for (el <- trace) { if (!finished) { - if (el.getClassName.startsWith("spark.") && !el.getClassName.startsWith("spark.examples.")) { + if (SPARK_CLASS_REGEX.findFirstIn(el.getClassName) != None) { lastSparkMethod = if (el.getMethodName == "<init>") { // Spark method is a constructor; get its class name el.getClassName.substring(el.getClassName.lastIndexOf('.') + 1) @@ -560,25 +637,33 @@ private object Utils extends Logging { else { firstUserLine = el.getLineNumber firstUserFile = el.getFileName + firstUserClass = el.getClassName finished = true } } } - "%s at %s:%s".format(lastSparkMethod, firstUserFile, firstUserLine) + new CallSiteInfo(lastSparkMethod, firstUserFile, firstUserLine, firstUserClass) } - /** - * Try to find a free port to bind to on the local host. This should ideally never be needed, - * except that, unfortunately, some of the networking libraries we currently rely on (e.g. Spray) - * don't let users bind to port 0 and then figure out which free port they actually bound to. - * We work around this by binding a ServerSocket and immediately unbinding it. This is *not* - * necessarily guaranteed to work, but it's the best we can do. - */ - def findFreePort(): Int = { - val socket = new ServerSocket(0) - val portBound = socket.getLocalPort - socket.close() - portBound + def formatSparkCallSite = { + val callSiteInfo = getCallSiteInfo + "%s at %s:%s".format(callSiteInfo.lastSparkMethod, callSiteInfo.firstUserFile, + callSiteInfo.firstUserLine) + } + + /** Return a string containing part of a file from byte 'start' to 'end'. */ + def offsetBytes(path: String, start: Long, end: Long): String = { + val file = new File(path) + val length = file.length() + val effectiveEnd = math.min(length, end) + val effectiveStart = math.max(0, start) + val buff = new Array[Byte]((effectiveEnd-effectiveStart).toInt) + val stream = new FileInputStream(file) + + stream.skip(effectiveStart) + stream.read(buff) + stream.close() + Source.fromBytes(buff).mkString } /** @@ -608,4 +693,67 @@ private object Utils extends Logging { } return false } + + def isSpace(c: Char): Boolean = { + " \t\r\n".indexOf(c) != -1 + } + + /** + * Split a string of potentially quoted arguments from the command line the way that a shell + * would do it to determine arguments to a command. For example, if the string is 'a "b c" d', + * then it would be parsed as three arguments: 'a', 'b c' and 'd'. + */ + def splitCommandString(s: String): Seq[String] = { + val buf = new ArrayBuffer[String] + var inWord = false + var inSingleQuote = false + var inDoubleQuote = false + var curWord = new StringBuilder + def endWord() { + buf += curWord.toString + curWord.clear() + } + var i = 0 + while (i < s.length) { + var nextChar = s.charAt(i) + if (inDoubleQuote) { + if (nextChar == '"') { + inDoubleQuote = false + } else if (nextChar == '\\') { + if (i < s.length - 1) { + // Append the next character directly, because only " and \ may be escaped in + // double quotes after the shell's own expansion + curWord.append(s.charAt(i + 1)) + i += 1 + } + } else { + curWord.append(nextChar) + } + } else if (inSingleQuote) { + if (nextChar == '\'') { + inSingleQuote = false + } else { + curWord.append(nextChar) + } + // Backslashes are not treated specially in single quotes + } else if (nextChar == '"') { + inWord = true + inDoubleQuote = true + } else if (nextChar == '\'') { + inWord = true + inSingleQuote = true + } else if (!isSpace(nextChar)) { + curWord.append(nextChar) + inWord = true + } else if (inWord && isSpace(nextChar)) { + endWord() + inWord = false + } + i += 1 + } + if (inWord || inDoubleQuote || inSingleQuote) { + endWord() + } + return buf + } } diff --git a/core/src/main/scala/spark/api/java/JavaDoubleRDD.scala b/core/src/main/scala/spark/api/java/JavaDoubleRDD.scala index 16692c0440..8ce7df6213 100644 --- a/core/src/main/scala/spark/api/java/JavaDoubleRDD.scala +++ b/core/src/main/scala/spark/api/java/JavaDoubleRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java import spark.RDD @@ -98,33 +115,48 @@ class JavaDoubleRDD(val srdd: RDD[scala.Double]) extends JavaRDDLike[Double, Jav // Double RDD functions - /** Return the sum of the elements in this RDD. */ + /** Add up the elements in this RDD. */ def sum(): Double = srdd.sum() - /** Return a [[spark.StatCounter]] describing the elements in this RDD. */ + /** + * Return a [[spark.util.StatCounter]] object that captures the mean, variance and count + * of the RDD's elements in one operation. + */ def stats(): StatCounter = srdd.stats() - /** Return the mean of the elements in this RDD. */ + /** Compute the mean of this RDD's elements. */ def mean(): Double = srdd.mean() - /** Return the variance of the elements in this RDD. */ + /** Compute the variance of this RDD's elements. */ def variance(): Double = srdd.variance() - /** Return the standard deviation of the elements in this RDD. */ + /** Compute the standard deviation of this RDD's elements. */ def stdev(): Double = srdd.stdev() + /** + * Compute the sample standard deviation of this RDD's elements (which corrects for bias in + * estimating the standard deviation by dividing by N-1 instead of N). + */ + def sampleStdev(): Double = srdd.sampleStdev() + + /** + * Compute the sample variance of this RDD's elements (which corrects for bias in + * estimating the standard variance by dividing by N-1 instead of N). + */ + def sampleVariance(): Double = srdd.sampleVariance() + /** Return the approximate mean of the elements in this RDD. */ def meanApprox(timeout: Long, confidence: Double): PartialResult[BoundedDouble] = srdd.meanApprox(timeout, confidence) - /** Return the approximate mean of the elements in this RDD. */ + /** (Experimental) Approximate operation to return the mean within a timeout. */ def meanApprox(timeout: Long): PartialResult[BoundedDouble] = srdd.meanApprox(timeout) - /** Return the approximate sum of the elements in this RDD. */ + /** (Experimental) Approximate operation to return the sum within a timeout. */ def sumApprox(timeout: Long, confidence: Double): PartialResult[BoundedDouble] = srdd.sumApprox(timeout, confidence) - - /** Return the approximate sum of the elements in this RDD. */ + + /** (Experimental) Approximate operation to return the sum within a timeout. */ def sumApprox(timeout: Long): PartialResult[BoundedDouble] = srdd.sumApprox(timeout) } diff --git a/core/src/main/scala/spark/api/java/JavaPairRDD.scala b/core/src/main/scala/spark/api/java/JavaPairRDD.scala index 30084df4e2..ccc511dc5f 100644 --- a/core/src/main/scala/spark/api/java/JavaPairRDD.scala +++ b/core/src/main/scala/spark/api/java/JavaPairRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java import java.util.{List => JList} @@ -6,6 +23,7 @@ import java.util.Comparator import scala.Tuple2 import scala.collection.JavaConversions._ +import org.apache.hadoop.io.compress.CompressionCodec import org.apache.hadoop.mapred.JobConf import org.apache.hadoop.mapred.OutputFormat import org.apache.hadoop.mapreduce.{OutputFormat => NewOutputFormat} @@ -459,6 +477,16 @@ class JavaPairRDD[K, V](val rdd: RDD[(K, V)])(implicit val kManifest: ClassManif rdd.saveAsHadoopFile(path, keyClass, valueClass, outputFormatClass) } + /** Output the RDD to any Hadoop-supported file system, compressing with the supplied codec. */ + def saveAsHadoopFile[F <: OutputFormat[_, _]]( + path: String, + keyClass: Class[_], + valueClass: Class[_], + outputFormatClass: Class[F], + codec: Class[_ <: CompressionCodec]) { + rdd.saveAsHadoopFile(path, keyClass, valueClass, outputFormatClass, codec) + } + /** Output the RDD to any Hadoop-supported file system. */ def saveAsNewAPIHadoopFile[F <: NewOutputFormat[_, _]]( path: String, diff --git a/core/src/main/scala/spark/api/java/JavaRDD.scala b/core/src/main/scala/spark/api/java/JavaRDD.scala index e29f1e5899..c0bf2cf568 100644 --- a/core/src/main/scala/spark/api/java/JavaRDD.scala +++ b/core/src/main/scala/spark/api/java/JavaRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java import spark._ @@ -14,12 +31,18 @@ JavaRDDLike[T, JavaRDD[T]] { /** Persist this RDD with the default storage level (`MEMORY_ONLY`). */ def cache(): JavaRDD[T] = wrapRDD(rdd.cache()) - /** + /** * Set this RDD's storage level to persist its values across operations after the first time - * it is computed. Can only be called once on each RDD. + * it is computed. This can only be used to assign a new storage level if the RDD does not + * have a storage level set yet.. */ def persist(newLevel: StorageLevel): JavaRDD[T] = wrapRDD(rdd.persist(newLevel)) + /** + * Mark the RDD as non-persistent, and remove all blocks for it from memory and disk. + */ + def unpersist(): JavaRDD[T] = wrapRDD(rdd.unpersist()) + // Transformations (return a new RDD) /** @@ -31,7 +54,7 @@ JavaRDDLike[T, JavaRDD[T]] { * Return a new RDD containing the distinct elements in this RDD. */ def distinct(numPartitions: Int): JavaRDD[T] = wrapRDD(rdd.distinct(numPartitions)) - + /** * Return a new RDD containing only the elements that satisfy a predicate. */ @@ -54,7 +77,7 @@ JavaRDDLike[T, JavaRDD[T]] { */ def sample(withReplacement: Boolean, fraction: Double, seed: Int): JavaRDD[T] = wrapRDD(rdd.sample(withReplacement, fraction, seed)) - + /** * Return the union of this RDD and another one. Any identical elements will appear multiple * times (use `.distinct()` to eliminate them). @@ -63,7 +86,7 @@ JavaRDDLike[T, JavaRDD[T]] { /** * Return an RDD with the elements from `this` that are not in `other`. - * + * * Uses `this` partitioner/partition size, because even if `other` is huge, the resulting * RDD will be <= us. */ @@ -80,7 +103,6 @@ JavaRDDLike[T, JavaRDD[T]] { */ def subtract(other: JavaRDD[T], p: Partitioner): JavaRDD[T] = wrapRDD(rdd.subtract(other, p)) - } object JavaRDD { diff --git a/core/src/main/scala/spark/api/java/JavaRDDLike.scala b/core/src/main/scala/spark/api/java/JavaRDDLike.scala index d884529d7a..21b5abf053 100644 --- a/core/src/main/scala/spark/api/java/JavaRDDLike.scala +++ b/core/src/main/scala/spark/api/java/JavaRDDLike.scala @@ -1,9 +1,27 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java -import java.util.{List => JList} +import java.util.{List => JList, Comparator} import scala.Tuple2 import scala.collection.JavaConversions._ +import org.apache.hadoop.io.compress.CompressionCodec import spark.{SparkContext, Partition, RDD, TaskContext} import spark.api.java.JavaPairRDD._ import spark.api.java.function.{Function2 => JFunction2, Function => JFunction, _} @@ -182,6 +200,21 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable { JavaPairRDD.fromRDD(rdd.zip(other.rdd)(other.classManifest))(classManifest, other.classManifest) } + /** + * Zip this RDD's partitions with one (or more) RDD(s) and return a new RDD by + * applying a function to the zipped partitions. Assumes that all the RDDs have the + * *same number of partitions*, but does *not* require them to have the same number + * of elements in each partition. + */ + def zipPartitions[U, V]( + f: FlatMapFunction2[java.util.Iterator[T], java.util.Iterator[U], V], + other: JavaRDDLike[U, _]): JavaRDD[V] = { + def fn = (x: Iterator[T], y: Iterator[U]) => asScalaIterator( + f.apply(asJavaIterator(x), asJavaIterator(y)).iterator()) + JavaRDD.fromRDD( + rdd.zipPartitions(fn, other.rdd)(other.classManifest, f.elementType()))(f.elementType()) + } + // Actions (launch a job to return a value to the user program) /** @@ -295,6 +328,13 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable { */ def saveAsTextFile(path: String) = rdd.saveAsTextFile(path) + + /** + * Save this RDD as a compressed text file, using string representations of elements. + */ + def saveAsTextFile(path: String, codec: Class[_ <: CompressionCodec]) = + rdd.saveAsTextFile(path, codec) + /** * Save this RDD as a SequenceFile of serialized objects. */ @@ -336,4 +376,54 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable { def toDebugString(): String = { rdd.toDebugString } + + /** + * Returns the top K elements from this RDD as defined by + * the specified Comparator[T]. + * @param num the number of top elements to return + * @param comp the comparator that defines the order + * @return an array of top elements + */ + def top(num: Int, comp: Comparator[T]): JList[T] = { + import scala.collection.JavaConversions._ + val topElems = rdd.top(num)(Ordering.comparatorToOrdering(comp)) + val arr: java.util.Collection[T] = topElems.toSeq + new java.util.ArrayList(arr) + } + + /** + * Returns the top K elements from this RDD using the + * natural ordering for T. + * @param num the number of top elements to return + * @return an array of top elements + */ + def top(num: Int): JList[T] = { + val comp = com.google.common.collect.Ordering.natural().asInstanceOf[Comparator[T]] + top(num, comp) + } + + /** + * Returns the first K elements from this RDD as defined by + * the specified Comparator[T] and maintains the order. + * @param num the number of top elements to return + * @param comp the comparator that defines the order + * @return an array of top elements + */ + def takeOrdered(num: Int, comp: Comparator[T]): JList[T] = { + import scala.collection.JavaConversions._ + val topElems = rdd.takeOrdered(num)(Ordering.comparatorToOrdering(comp)) + val arr: java.util.Collection[T] = topElems.toSeq + new java.util.ArrayList(arr) + } + + /** + * Returns the first K elements from this RDD using the + * natural ordering for T while maintain the order. + * @param num the number of top elements to return + * @return an array of top elements + */ + def takeOrdered(num: Int): JList[T] = { + val comp = com.google.common.collect.Ordering.natural().asInstanceOf[Comparator[T]] + takeOrdered(num, comp) + } } diff --git a/core/src/main/scala/spark/api/java/JavaSparkContext.scala b/core/src/main/scala/spark/api/java/JavaSparkContext.scala index 5f18b1e15b..fe182e7ab6 100644 --- a/core/src/main/scala/spark/api/java/JavaSparkContext.scala +++ b/core/src/main/scala/spark/api/java/JavaSparkContext.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java import java.util.{Map => JMap} diff --git a/core/src/main/scala/spark/api/java/JavaSparkContextVarargsWorkaround.java b/core/src/main/scala/spark/api/java/JavaSparkContextVarargsWorkaround.java index 97344e73da..42b1de01b1 100644 --- a/core/src/main/scala/spark/api/java/JavaSparkContextVarargsWorkaround.java +++ b/core/src/main/scala/spark/api/java/JavaSparkContextVarargsWorkaround.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java; import java.util.Arrays; diff --git a/core/src/main/scala/spark/api/java/StorageLevels.java b/core/src/main/scala/spark/api/java/StorageLevels.java index 5e5845ac3a..f385636e83 100644 --- a/core/src/main/scala/spark/api/java/StorageLevels.java +++ b/core/src/main/scala/spark/api/java/StorageLevels.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java; import spark.storage.StorageLevel; diff --git a/core/src/main/scala/spark/api/java/function/DoubleFlatMapFunction.java b/core/src/main/scala/spark/api/java/function/DoubleFlatMapFunction.java index 3a8192be3a..8bc88d757f 100644 --- a/core/src/main/scala/spark/api/java/function/DoubleFlatMapFunction.java +++ b/core/src/main/scala/spark/api/java/function/DoubleFlatMapFunction.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function; diff --git a/core/src/main/scala/spark/api/java/function/DoubleFunction.java b/core/src/main/scala/spark/api/java/function/DoubleFunction.java index c6ef76d088..1aa1e5dae0 100644 --- a/core/src/main/scala/spark/api/java/function/DoubleFunction.java +++ b/core/src/main/scala/spark/api/java/function/DoubleFunction.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function; diff --git a/core/src/main/scala/spark/api/java/function/FlatMapFunction.scala b/core/src/main/scala/spark/api/java/function/FlatMapFunction.scala index e027cdacd3..9eb0cfe3f9 100644 --- a/core/src/main/scala/spark/api/java/function/FlatMapFunction.scala +++ b/core/src/main/scala/spark/api/java/function/FlatMapFunction.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function /** diff --git a/core/src/main/scala/spark/api/java/function/FlatMapFunction2.scala b/core/src/main/scala/spark/api/java/function/FlatMapFunction2.scala new file mode 100644 index 0000000000..dda98710c2 --- /dev/null +++ b/core/src/main/scala/spark/api/java/function/FlatMapFunction2.scala @@ -0,0 +1,28 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.api.java.function + +/** + * A function that takes two inputs and returns zero or more output records. + */ +abstract class FlatMapFunction2[A, B, C] extends Function2[A, B, java.lang.Iterable[C]] { + @throws(classOf[Exception]) + def call(a: A, b:B) : java.lang.Iterable[C] + + def elementType() : ClassManifest[C] = ClassManifest.Any.asInstanceOf[ClassManifest[C]] +} diff --git a/core/src/main/scala/spark/api/java/function/Function.java b/core/src/main/scala/spark/api/java/function/Function.java index dae8295f21..2a2ea0aacf 100644 --- a/core/src/main/scala/spark/api/java/function/Function.java +++ b/core/src/main/scala/spark/api/java/function/Function.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function; import scala.reflect.ClassManifest; diff --git a/core/src/main/scala/spark/api/java/function/Function2.java b/core/src/main/scala/spark/api/java/function/Function2.java index 69bf12c8c9..952d31ece4 100644 --- a/core/src/main/scala/spark/api/java/function/Function2.java +++ b/core/src/main/scala/spark/api/java/function/Function2.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function; import scala.reflect.ClassManifest; diff --git a/core/src/main/scala/spark/api/java/function/PairFlatMapFunction.java b/core/src/main/scala/spark/api/java/function/PairFlatMapFunction.java index b3cc4df6aa..4aad602da3 100644 --- a/core/src/main/scala/spark/api/java/function/PairFlatMapFunction.java +++ b/core/src/main/scala/spark/api/java/function/PairFlatMapFunction.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function; import scala.Tuple2; diff --git a/core/src/main/scala/spark/api/java/function/PairFunction.java b/core/src/main/scala/spark/api/java/function/PairFunction.java index 9fc6df4b88..ccfe64ecf1 100644 --- a/core/src/main/scala/spark/api/java/function/PairFunction.java +++ b/core/src/main/scala/spark/api/java/function/PairFunction.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function; import scala.Tuple2; diff --git a/core/src/main/scala/spark/api/java/function/VoidFunction.scala b/core/src/main/scala/spark/api/java/function/VoidFunction.scala index b0096cf2bf..f6fc0b0f7d 100644 --- a/core/src/main/scala/spark/api/java/function/VoidFunction.scala +++ b/core/src/main/scala/spark/api/java/function/VoidFunction.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function /** @@ -13,4 +30,4 @@ abstract class VoidFunction[T] extends Serializable { // return Unit), so it is implicitly converted to a Function1[T, Unit]: object VoidFunction { implicit def toFunction[T](f: VoidFunction[T]) : Function1[T, Unit] = ((x : T) => f.call(x)) -}
\ No newline at end of file +} diff --git a/core/src/main/scala/spark/api/java/function/WrappedFunction1.scala b/core/src/main/scala/spark/api/java/function/WrappedFunction1.scala index 923f5cdf4f..1758a38c4e 100644 --- a/core/src/main/scala/spark/api/java/function/WrappedFunction1.scala +++ b/core/src/main/scala/spark/api/java/function/WrappedFunction1.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function import scala.runtime.AbstractFunction1 diff --git a/core/src/main/scala/spark/api/java/function/WrappedFunction2.scala b/core/src/main/scala/spark/api/java/function/WrappedFunction2.scala index 2c6e9b1571..b093567d2c 100644 --- a/core/src/main/scala/spark/api/java/function/WrappedFunction2.scala +++ b/core/src/main/scala/spark/api/java/function/WrappedFunction2.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.java.function import scala.runtime.AbstractFunction2 diff --git a/core/src/main/scala/spark/api/python/PythonPartitioner.scala b/core/src/main/scala/spark/api/python/PythonPartitioner.scala index d618c098c2..31a719fbff 100644 --- a/core/src/main/scala/spark/api/python/PythonPartitioner.scala +++ b/core/src/main/scala/spark/api/python/PythonPartitioner.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.python import spark.Partitioner diff --git a/core/src/main/scala/spark/api/python/PythonRDD.scala b/core/src/main/scala/spark/api/python/PythonRDD.scala index 807119ca8c..af10822dbd 100644 --- a/core/src/main/scala/spark/api/python/PythonRDD.scala +++ b/core/src/main/scala/spark/api/python/PythonRDD.scala @@ -1,11 +1,27 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.api.python import java.io._ import java.net._ -import java.util.{List => JList, ArrayList => JArrayList, Collections} +import java.util.{List => JList, ArrayList => JArrayList, Map => JMap, Collections} import scala.collection.JavaConversions._ -import scala.io.Source import spark.api.java.{JavaSparkContext, JavaPairRDD, JavaRDD} import spark.broadcast.Broadcast @@ -16,16 +32,18 @@ import spark.rdd.PipedRDD private[spark] class PythonRDD[T: ClassManifest]( parent: RDD[T], command: Seq[String], - envVars: java.util.Map[String, String], + envVars: JMap[String, String], preservePartitoning: Boolean, pythonExec: String, broadcastVars: JList[Broadcast[Array[Byte]]], accumulator: Accumulator[JList[Array[Byte]]]) extends RDD[Array[Byte]](parent) { + val bufferSize = System.getProperty("spark.buffer.size", "65536").toInt + // Similar to Runtime.exec(), if we are given a single string, split it into words // using a standard StringTokenizer (i.e. by spaces) - def this(parent: RDD[T], command: String, envVars: java.util.Map[String, String], + def this(parent: RDD[T], command: String, envVars: JMap[String, String], preservePartitoning: Boolean, pythonExec: String, broadcastVars: JList[Broadcast[Array[Byte]]], accumulator: Accumulator[JList[Array[Byte]]]) = @@ -36,68 +54,57 @@ private[spark] class PythonRDD[T: ClassManifest]( override val partitioner = if (preservePartitoning) parent.partitioner else None - override def compute(split: Partition, context: TaskContext): Iterator[Array[Byte]] = { - val SPARK_HOME = new ProcessBuilder().environment().get("SPARK_HOME") - - val pb = new ProcessBuilder(Seq(pythonExec, SPARK_HOME + "/python/pyspark/worker.py")) - // Add the environmental variables to the process. - val currentEnvVars = pb.environment() - - for ((variable, value) <- envVars) { - currentEnvVars.put(variable, value) - } - val proc = pb.start() + override def compute(split: Partition, context: TaskContext): Iterator[Array[Byte]] = { + val startTime = System.currentTimeMillis val env = SparkEnv.get - - // Start a thread to print the process's stderr to ours - new Thread("stderr reader for " + pythonExec) { - override def run() { - for (line <- Source.fromInputStream(proc.getErrorStream).getLines) { - System.err.println(line) - } - } - }.start() + val worker = env.createPythonWorker(pythonExec, envVars.toMap) // Start a thread to feed the process input from our parent's iterator new Thread("stdin writer for " + pythonExec) { override def run() { SparkEnv.set(env) - val out = new PrintWriter(proc.getOutputStream) - val dOut = new DataOutputStream(proc.getOutputStream) + val stream = new BufferedOutputStream(worker.getOutputStream, bufferSize) + val dataOut = new DataOutputStream(stream) + val printOut = new PrintWriter(stream) // Partition index - dOut.writeInt(split.index) + dataOut.writeInt(split.index) // sparkFilesDir - PythonRDD.writeAsPickle(SparkFiles.getRootDirectory, dOut) + PythonRDD.writeAsPickle(SparkFiles.getRootDirectory, dataOut) // Broadcast variables - dOut.writeInt(broadcastVars.length) + dataOut.writeInt(broadcastVars.length) for (broadcast <- broadcastVars) { - dOut.writeLong(broadcast.id) - dOut.writeInt(broadcast.value.length) - dOut.write(broadcast.value) - dOut.flush() + dataOut.writeLong(broadcast.id) + dataOut.writeInt(broadcast.value.length) + dataOut.write(broadcast.value) } + dataOut.flush() // Serialized user code for (elem <- command) { - out.println(elem) + printOut.println(elem) } - out.flush() + printOut.flush() // Data values for (elem <- parent.iterator(split, context)) { - PythonRDD.writeAsPickle(elem, dOut) + PythonRDD.writeAsPickle(elem, dataOut) } - dOut.flush() - out.flush() - proc.getOutputStream.close() + dataOut.flush() + printOut.flush() + worker.shutdownOutput() } }.start() // Return an iterator that read lines from the process's stdout - val stream = new DataInputStream(proc.getInputStream) + val stream = new DataInputStream(new BufferedInputStream(worker.getInputStream, bufferSize)) return new Iterator[Array[Byte]] { def next(): Array[Byte] = { val obj = _nextObj - _nextObj = read() + if (hasNext) { + // FIXME: can deadlock if worker is waiting for us to + // respond to current message (currently irrelevant because + // output is shutdown before we read any input) + _nextObj = read() + } obj } @@ -108,6 +115,17 @@ private[spark] class PythonRDD[T: ClassManifest]( val obj = new Array[Byte](length) stream.readFully(obj) obj + case -3 => + // Timing data from worker + val bootTime = stream.readLong() + val initTime = stream.readLong() + val finishTime = stream.readLong() + val boot = bootTime - startTime + val init = initTime - bootTime + val finish = finishTime - initTime + val total = finishTime - startTime + logInfo("Times: total = %s, boot = %s, init = %s, finish = %s".format(total, boot, init, finish)) + read case -2 => // Signals that an exception has been thrown in python val exLength = stream.readInt() @@ -115,23 +133,21 @@ private[spark] class PythonRDD[T: ClassManifest]( stream.readFully(obj) throw new PythonException(new String(obj)) case -1 => - // We've finished the data section of the output, but we can still read some - // accumulator updates; let's do that, breaking when we get EOFException - while (true) { - val len2 = stream.readInt() + // We've finished the data section of the output, but we can still + // read some accumulator updates; let's do that, breaking when we + // get a negative length record. + var len2 = stream.readInt() + while (len2 >= 0) { val update = new Array[Byte](len2) stream.readFully(update) accumulator += Collections.singletonList(update) + len2 = stream.readInt() } new Array[Byte](0) } } catch { case eof: EOFException => { - val exitStatus = proc.waitFor() - if (exitStatus != 0) { - throw new Exception("Subprocess exited with status " + exitStatus) - } - new Array[Byte](0) + throw new SparkException("Python worker exited unexpectedly (crashed)", eof) } case e => throw e } @@ -159,7 +175,7 @@ private class PairwiseRDD(prev: RDD[Array[Byte]]) extends override def compute(split: Partition, context: TaskContext) = prev.iterator(split, context).grouped(2).map { case Seq(a, b) => (a, b) - case x => throw new Exception("PairwiseRDD: unexpected value: " + x) + case x => throw new SparkException("PairwiseRDD: unexpected value: " + x) } val asJavaPairRDD : JavaPairRDD[Array[Byte], Array[Byte]] = JavaPairRDD.fromRDD(this) } @@ -215,7 +231,7 @@ private[spark] object PythonRDD { dOut.write(s) dOut.writeByte(Pickle.STOP) } else { - throw new Exception("Unexpected RDD type") + throw new SparkException("Unexpected RDD type") } } @@ -279,6 +295,8 @@ class PythonAccumulatorParam(@transient serverHost: String, serverPort: Int) extends AccumulatorParam[JList[Array[Byte]]] { Utils.checkHost(serverHost, "Expected hostname") + + val bufferSize = System.getProperty("spark.buffer.size", "65536").toInt override def zero(value: JList[Array[Byte]]): JList[Array[Byte]] = new JArrayList @@ -292,7 +310,7 @@ class PythonAccumulatorParam(@transient serverHost: String, serverPort: Int) // This happens on the master, where we pass the updates to Python through a socket val socket = new Socket(serverHost, serverPort) val in = socket.getInputStream - val out = new DataOutputStream(socket.getOutputStream) + val out = new DataOutputStream(new BufferedOutputStream(socket.getOutputStream, bufferSize)) out.writeInt(val2.size) for (array <- val2) { out.writeInt(array.length) diff --git a/core/src/main/scala/spark/api/python/PythonWorkerFactory.scala b/core/src/main/scala/spark/api/python/PythonWorkerFactory.scala new file mode 100644 index 0000000000..078ad45ce8 --- /dev/null +++ b/core/src/main/scala/spark/api/python/PythonWorkerFactory.scala @@ -0,0 +1,130 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.api.python + +import java.io.{DataInputStream, IOException} +import java.net.{Socket, SocketException, InetAddress} + +import scala.collection.JavaConversions._ + +import spark._ + +private[spark] class PythonWorkerFactory(pythonExec: String, envVars: Map[String, String]) + extends Logging { + var daemon: Process = null + val daemonHost = InetAddress.getByAddress(Array(127, 0, 0, 1)) + var daemonPort: Int = 0 + + def create(): Socket = { + synchronized { + // Start the daemon if it hasn't been started + startDaemon() + + // Attempt to connect, restart and retry once if it fails + try { + new Socket(daemonHost, daemonPort) + } catch { + case exc: SocketException => { + logWarning("Python daemon unexpectedly quit, attempting to restart") + stopDaemon() + startDaemon() + new Socket(daemonHost, daemonPort) + } + case e => throw e + } + } + } + + def stop() { + stopDaemon() + } + + private def startDaemon() { + synchronized { + // Is it already running? + if (daemon != null) { + return + } + + try { + // Create and start the daemon + val sparkHome = new ProcessBuilder().environment().get("SPARK_HOME") + val pb = new ProcessBuilder(Seq(pythonExec, sparkHome + "/python/pyspark/daemon.py")) + val workerEnv = pb.environment() + workerEnv.putAll(envVars) + daemon = pb.start() + + // Redirect the stderr to ours + new Thread("stderr reader for " + pythonExec) { + override def run() { + scala.util.control.Exception.ignoring(classOf[IOException]) { + // FIXME HACK: We copy the stream on the level of bytes to + // attempt to dodge encoding problems. + val in = daemon.getErrorStream + var buf = new Array[Byte](1024) + var len = in.read(buf) + while (len != -1) { + System.err.write(buf, 0, len) + len = in.read(buf) + } + } + } + }.start() + + val in = new DataInputStream(daemon.getInputStream) + daemonPort = in.readInt() + + // Redirect further stdout output to our stderr + new Thread("stdout reader for " + pythonExec) { + override def run() { + scala.util.control.Exception.ignoring(classOf[IOException]) { + // FIXME HACK: We copy the stream on the level of bytes to + // attempt to dodge encoding problems. + var buf = new Array[Byte](1024) + var len = in.read(buf) + while (len != -1) { + System.err.write(buf, 0, len) + len = in.read(buf) + } + } + } + }.start() + } catch { + case e => { + stopDaemon() + throw e + } + } + + // Important: don't close daemon's stdin (daemon.getOutputStream) so it can correctly + // detect our disappearance. + } + } + + private def stopDaemon() { + synchronized { + // Request shutdown of existing daemon by sending SIGTERM + if (daemon != null) { + daemon.destroy() + } + + daemon = null + daemonPort = 0 + } + } +} diff --git a/core/src/main/scala/spark/broadcast/BitTorrentBroadcast.scala b/core/src/main/scala/spark/broadcast/BitTorrentBroadcast.scala index adcb2d2415..6f7d385379 100644 --- a/core/src/main/scala/spark/broadcast/BitTorrentBroadcast.scala +++ b/core/src/main/scala/spark/broadcast/BitTorrentBroadcast.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.broadcast import java.io._ diff --git a/core/src/main/scala/spark/broadcast/Broadcast.scala b/core/src/main/scala/spark/broadcast/Broadcast.scala index 415bde5d67..aba56a60ca 100644 --- a/core/src/main/scala/spark/broadcast/Broadcast.scala +++ b/core/src/main/scala/spark/broadcast/Broadcast.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.broadcast import java.io._ diff --git a/core/src/main/scala/spark/broadcast/BroadcastFactory.scala b/core/src/main/scala/spark/broadcast/BroadcastFactory.scala index 5c6184c3c7..d33d95c7d9 100644 --- a/core/src/main/scala/spark/broadcast/BroadcastFactory.scala +++ b/core/src/main/scala/spark/broadcast/BroadcastFactory.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.broadcast /** diff --git a/core/src/main/scala/spark/broadcast/HttpBroadcast.scala b/core/src/main/scala/spark/broadcast/HttpBroadcast.scala index 7e30b8f7d2..c565876950 100644 --- a/core/src/main/scala/spark/broadcast/HttpBroadcast.scala +++ b/core/src/main/scala/spark/broadcast/HttpBroadcast.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.broadcast import com.ning.compress.lzf.{LZFInputStream, LZFOutputStream} diff --git a/core/src/main/scala/spark/broadcast/MultiTracker.scala b/core/src/main/scala/spark/broadcast/MultiTracker.scala index 3fd77af73f..7855d44e9b 100644 --- a/core/src/main/scala/spark/broadcast/MultiTracker.scala +++ b/core/src/main/scala/spark/broadcast/MultiTracker.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.broadcast import java.io._ diff --git a/core/src/main/scala/spark/broadcast/SourceInfo.scala b/core/src/main/scala/spark/broadcast/SourceInfo.scala index c79bb93c38..b17ae63b5c 100644 --- a/core/src/main/scala/spark/broadcast/SourceInfo.scala +++ b/core/src/main/scala/spark/broadcast/SourceInfo.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.broadcast import java.util.BitSet diff --git a/core/src/main/scala/spark/broadcast/TreeBroadcast.scala b/core/src/main/scala/spark/broadcast/TreeBroadcast.scala index c55c476117..ea1e9a12c1 100644 --- a/core/src/main/scala/spark/broadcast/TreeBroadcast.scala +++ b/core/src/main/scala/spark/broadcast/TreeBroadcast.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.broadcast import java.io._ diff --git a/core/src/main/scala/spark/deploy/ApplicationDescription.scala b/core/src/main/scala/spark/deploy/ApplicationDescription.scala index 6659e53b25..a8b22fbef8 100644 --- a/core/src/main/scala/spark/deploy/ApplicationDescription.scala +++ b/core/src/main/scala/spark/deploy/ApplicationDescription.scala @@ -1,11 +1,29 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy private[spark] class ApplicationDescription( val name: String, - val cores: Int, + val maxCores: Int, /* Integer.MAX_VALUE denotes an unlimited number of cores */ val memoryPerSlave: Int, val command: Command, - val sparkHome: String) + val sparkHome: String, + val appUiUrl: String) extends Serializable { val user = System.getProperty("user.name", "<unknown>") diff --git a/core/src/main/scala/spark/deploy/Command.scala b/core/src/main/scala/spark/deploy/Command.scala index 577101e3c3..bad629e965 100644 --- a/core/src/main/scala/spark/deploy/Command.scala +++ b/core/src/main/scala/spark/deploy/Command.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy import scala.collection.Map diff --git a/core/src/main/scala/spark/deploy/DeployMessage.scala b/core/src/main/scala/spark/deploy/DeployMessage.scala index 51274acb1e..e1f8aff6f5 100644 --- a/core/src/main/scala/spark/deploy/DeployMessage.scala +++ b/core/src/main/scala/spark/deploy/DeployMessage.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy import spark.deploy.ExecutorState.ExecutorState diff --git a/core/src/main/scala/spark/deploy/ExecutorState.scala b/core/src/main/scala/spark/deploy/ExecutorState.scala index 5dc0c54552..08c9a3b725 100644 --- a/core/src/main/scala/spark/deploy/ExecutorState.scala +++ b/core/src/main/scala/spark/deploy/ExecutorState.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy private[spark] object ExecutorState diff --git a/core/src/main/scala/spark/deploy/JsonProtocol.scala b/core/src/main/scala/spark/deploy/JsonProtocol.scala index 71a641a9ef..64f89623e1 100644 --- a/core/src/main/scala/spark/deploy/JsonProtocol.scala +++ b/core/src/main/scala/spark/deploy/JsonProtocol.scala @@ -1,79 +1,82 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy import master.{ApplicationInfo, WorkerInfo} +import net.liftweb.json.JsonDSL._ import worker.ExecutorRunner -import cc.spray.json._ -/** - * spray-json helper class containing implicit conversion to json for marshalling responses - */ -private[spark] object JsonProtocol extends DefaultJsonProtocol { - implicit object WorkerInfoJsonFormat extends RootJsonWriter[WorkerInfo] { - def write(obj: WorkerInfo) = JsObject( - "id" -> JsString(obj.id), - "host" -> JsString(obj.host), - "port" -> JsNumber(obj.port), - "webuiaddress" -> JsString(obj.webUiAddress), - "cores" -> JsNumber(obj.cores), - "coresused" -> JsNumber(obj.coresUsed), - "memory" -> JsNumber(obj.memory), - "memoryused" -> JsNumber(obj.memoryUsed) - ) - } +private[spark] object JsonProtocol { + def writeWorkerInfo(obj: WorkerInfo) = { + ("id" -> obj.id) ~ + ("host" -> obj.host) ~ + ("port" -> obj.port) ~ + ("webuiaddress" -> obj.webUiAddress) ~ + ("cores" -> obj.cores) ~ + ("coresused" -> obj.coresUsed) ~ + ("memory" -> obj.memory) ~ + ("memoryused" -> obj.memoryUsed) + } - implicit object AppInfoJsonFormat extends RootJsonWriter[ApplicationInfo] { - def write(obj: ApplicationInfo) = JsObject( - "starttime" -> JsNumber(obj.startTime), - "id" -> JsString(obj.id), - "name" -> JsString(obj.desc.name), - "cores" -> JsNumber(obj.desc.cores), - "user" -> JsString(obj.desc.user), - "memoryperslave" -> JsNumber(obj.desc.memoryPerSlave), - "submitdate" -> JsString(obj.submitDate.toString)) + def writeApplicationInfo(obj: ApplicationInfo) = { + ("starttime" -> obj.startTime) ~ + ("id" -> obj.id) ~ + ("name" -> obj.desc.name) ~ + ("cores" -> obj.desc.maxCores) ~ + ("user" -> obj.desc.user) ~ + ("memoryperslave" -> obj.desc.memoryPerSlave) ~ + ("submitdate" -> obj.submitDate.toString) } - implicit object AppDescriptionJsonFormat extends RootJsonWriter[ApplicationDescription] { - def write(obj: ApplicationDescription) = JsObject( - "name" -> JsString(obj.name), - "cores" -> JsNumber(obj.cores), - "memoryperslave" -> JsNumber(obj.memoryPerSlave), - "user" -> JsString(obj.user) - ) + def writeApplicationDescription(obj: ApplicationDescription) = { + ("name" -> obj.name) ~ + ("cores" -> obj.maxCores) ~ + ("memoryperslave" -> obj.memoryPerSlave) ~ + ("user" -> obj.user) } - implicit object ExecutorRunnerJsonFormat extends RootJsonWriter[ExecutorRunner] { - def write(obj: ExecutorRunner) = JsObject( - "id" -> JsNumber(obj.execId), - "memory" -> JsNumber(obj.memory), - "appid" -> JsString(obj.appId), - "appdesc" -> obj.appDesc.toJson.asJsObject - ) + def writeExecutorRunner(obj: ExecutorRunner) = { + ("id" -> obj.execId) ~ + ("memory" -> obj.memory) ~ + ("appid" -> obj.appId) ~ + ("appdesc" -> writeApplicationDescription(obj.appDesc)) } - implicit object MasterStateJsonFormat extends RootJsonWriter[MasterState] { - def write(obj: MasterState) = JsObject( - "url" -> JsString("spark://" + obj.uri), - "workers" -> JsArray(obj.workers.toList.map(_.toJson)), - "cores" -> JsNumber(obj.workers.map(_.cores).sum), - "coresused" -> JsNumber(obj.workers.map(_.coresUsed).sum), - "memory" -> JsNumber(obj.workers.map(_.memory).sum), - "memoryused" -> JsNumber(obj.workers.map(_.memoryUsed).sum), - "activeapps" -> JsArray(obj.activeApps.toList.map(_.toJson)), - "completedapps" -> JsArray(obj.completedApps.toList.map(_.toJson)) - ) + def writeMasterState(obj: MasterState) = { + ("url" -> ("spark://" + obj.uri)) ~ + ("workers" -> obj.workers.toList.map(writeWorkerInfo)) ~ + ("cores" -> obj.workers.map(_.cores).sum) ~ + ("coresused" -> obj.workers.map(_.coresUsed).sum) ~ + ("memory" -> obj.workers.map(_.memory).sum) ~ + ("memoryused" -> obj.workers.map(_.memoryUsed).sum) ~ + ("activeapps" -> obj.activeApps.toList.map(writeApplicationInfo)) ~ + ("completedapps" -> obj.completedApps.toList.map(writeApplicationInfo)) } - implicit object WorkerStateJsonFormat extends RootJsonWriter[WorkerState] { - def write(obj: WorkerState) = JsObject( - "id" -> JsString(obj.workerId), - "masterurl" -> JsString(obj.masterUrl), - "masterwebuiurl" -> JsString(obj.masterWebUiUrl), - "cores" -> JsNumber(obj.cores), - "coresused" -> JsNumber(obj.coresUsed), - "memory" -> JsNumber(obj.memory), - "memoryused" -> JsNumber(obj.memoryUsed), - "executors" -> JsArray(obj.executors.toList.map(_.toJson)), - "finishedexecutors" -> JsArray(obj.finishedExecutors.toList.map(_.toJson)) - ) + def writeWorkerState(obj: WorkerState) = { + ("id" -> obj.workerId) ~ + ("masterurl" -> obj.masterUrl) ~ + ("masterwebuiurl" -> obj.masterWebUiUrl) ~ + ("cores" -> obj.cores) ~ + ("coresused" -> obj.coresUsed) ~ + ("memory" -> obj.memory) ~ + ("memoryused" -> obj.memoryUsed) ~ + ("executors" -> obj.executors.toList.map(writeExecutorRunner)) ~ + ("finishedexecutors" -> obj.finishedExecutors.toList.map(writeExecutorRunner)) } } diff --git a/core/src/main/scala/spark/deploy/LocalSparkCluster.scala b/core/src/main/scala/spark/deploy/LocalSparkCluster.scala index 55bb61b0cc..6b8e9f27af 100644 --- a/core/src/main/scala/spark/deploy/LocalSparkCluster.scala +++ b/core/src/main/scala/spark/deploy/LocalSparkCluster.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy import akka.actor.{ActorRef, Props, Actor, ActorSystem, Terminated} @@ -45,6 +62,7 @@ class LocalSparkCluster(numWorkers: Int, coresPerWorker: Int, memoryPerWorker: I // Stop the workers before the master so they don't get upset that it disconnected workerActorSystems.foreach(_.shutdown()) workerActorSystems.foreach(_.awaitTermination()) + masterActorSystems.foreach(_.shutdown()) masterActorSystems.foreach(_.awaitTermination()) } diff --git a/core/src/main/scala/spark/deploy/WebUI.scala b/core/src/main/scala/spark/deploy/WebUI.scala index ad1a1092b2..8ea7792ef4 100644 --- a/core/src/main/scala/spark/deploy/WebUI.scala +++ b/core/src/main/scala/spark/deploy/WebUI.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy import java.text.SimpleDateFormat @@ -6,7 +23,7 @@ import java.util.Date /** * Utilities used throughout the web UI. */ -private[spark] object WebUI { +private[spark] object DeployWebUI { val DATE_FORMAT = new SimpleDateFormat("yyyy/MM/dd HH:mm:ss") def formatDate(date: Date): String = DATE_FORMAT.format(date) diff --git a/core/src/main/scala/spark/deploy/client/Client.scala b/core/src/main/scala/spark/deploy/client/Client.scala index 4af44f9c16..edefa0292d 100644 --- a/core/src/main/scala/spark/deploy/client/Client.scala +++ b/core/src/main/scala/spark/deploy/client/Client.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.client import spark.deploy._ diff --git a/core/src/main/scala/spark/deploy/client/ClientListener.scala b/core/src/main/scala/spark/deploy/client/ClientListener.scala index e8c4083f9d..064024455e 100644 --- a/core/src/main/scala/spark/deploy/client/ClientListener.scala +++ b/core/src/main/scala/spark/deploy/client/ClientListener.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.client /** diff --git a/core/src/main/scala/spark/deploy/client/TestClient.scala b/core/src/main/scala/spark/deploy/client/TestClient.scala index ad92532b58..4f4daa141a 100644 --- a/core/src/main/scala/spark/deploy/client/TestClient.scala +++ b/core/src/main/scala/spark/deploy/client/TestClient.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.client import spark.util.AkkaUtils @@ -25,7 +42,7 @@ private[spark] object TestClient { val url = args(0) val (actorSystem, port) = AkkaUtils.createActorSystem("spark", Utils.localIpAddress, 0) val desc = new ApplicationDescription( - "TestClient", 1, 512, Command("spark.deploy.client.TestExecutor", Seq(), Map()), "dummy-spark-home") + "TestClient", 1, 512, Command("spark.deploy.client.TestExecutor", Seq(), Map()), "dummy-spark-home", "ignored") val listener = new TestListener val client = new Client(actorSystem, url, desc, listener) client.start() diff --git a/core/src/main/scala/spark/deploy/client/TestExecutor.scala b/core/src/main/scala/spark/deploy/client/TestExecutor.scala index 0e46db2272..8a22b6b89f 100644 --- a/core/src/main/scala/spark/deploy/client/TestExecutor.scala +++ b/core/src/main/scala/spark/deploy/client/TestExecutor.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.client private[spark] object TestExecutor { diff --git a/core/src/main/scala/spark/deploy/master/ApplicationInfo.scala b/core/src/main/scala/spark/deploy/master/ApplicationInfo.scala index 3591a94072..15ff919738 100644 --- a/core/src/main/scala/spark/deploy/master/ApplicationInfo.scala +++ b/core/src/main/scala/spark/deploy/master/ApplicationInfo.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.master import spark.deploy.ApplicationDescription @@ -10,7 +27,8 @@ private[spark] class ApplicationInfo( val id: String, val desc: ApplicationDescription, val submitDate: Date, - val driver: ActorRef) + val driver: ActorRef, + val appUiUrl: String) { var state = ApplicationState.WAITING var executors = new mutable.HashMap[Int, ExecutorInfo] @@ -37,7 +55,7 @@ private[spark] class ApplicationInfo( coresGranted -= exec.cores } - def coresLeft: Int = desc.cores - coresGranted + def coresLeft: Int = desc.maxCores - coresGranted private var _retryCount = 0 @@ -60,4 +78,5 @@ private[spark] class ApplicationInfo( System.currentTimeMillis() - startTime } } + } diff --git a/core/src/main/scala/spark/deploy/master/ApplicationState.scala b/core/src/main/scala/spark/deploy/master/ApplicationState.scala index 15016b388d..94f0ad8bae 100644 --- a/core/src/main/scala/spark/deploy/master/ApplicationState.scala +++ b/core/src/main/scala/spark/deploy/master/ApplicationState.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.master private[spark] object ApplicationState diff --git a/core/src/main/scala/spark/deploy/master/ExecutorInfo.scala b/core/src/main/scala/spark/deploy/master/ExecutorInfo.scala index 48e6055fb5..99b60f7d09 100644 --- a/core/src/main/scala/spark/deploy/master/ExecutorInfo.scala +++ b/core/src/main/scala/spark/deploy/master/ExecutorInfo.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.master import spark.deploy.ExecutorState diff --git a/core/src/main/scala/spark/deploy/master/Master.scala b/core/src/main/scala/spark/deploy/master/Master.scala index 160afe5239..9692af5295 100644 --- a/core/src/main/scala/spark/deploy/master/Master.scala +++ b/core/src/main/scala/spark/deploy/master/Master.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.master import akka.actor._ @@ -12,7 +29,9 @@ import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} import spark.deploy._ import spark.{Logging, SparkException, Utils} +import spark.metrics.MetricsSystem import spark.util.AkkaUtils +import ui.MasterWebUI private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Actor with Logging { @@ -35,8 +54,13 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act var firstApp: Option[ApplicationInfo] = None + val webUi = new MasterWebUI(self, webUiPort) + Utils.checkHost(host, "Expected hostname") + val metricsSystem = MetricsSystem.createMetricsSystem("master") + val masterSource = new MasterSource(this) + val masterPublicAddress = { val envVar = System.getenv("SPARK_PUBLIC_DNS") if (envVar != null) envVar else host @@ -51,19 +75,16 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act logInfo("Starting Spark master at spark://" + host + ":" + port) // Listen for remote client disconnection events, since they don't go through Akka's watch() context.system.eventStream.subscribe(self, classOf[RemoteClientLifeCycleEvent]) - startWebUi() + webUi.start() context.system.scheduler.schedule(0 millis, WORKER_TIMEOUT millis)(timeOutDeadWorkers()) + + metricsSystem.registerSource(masterSource) + metricsSystem.start() } - def startWebUi() { - val webUi = new MasterWebUI(context.system, self) - try { - AkkaUtils.startSprayServer(context.system, "0.0.0.0", webUiPort, webUi.handler) - } catch { - case e: Exception => - logError("Failed to create web UI", e) - System.exit(1) - } + override def postStop() { + webUi.stop() + metricsSystem.stop() } override def receive = { @@ -75,7 +96,7 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act } else { addWorker(id, host, workerPort, cores, memory, worker_webUiPort, publicAddress) context.watch(sender) // This doesn't work with remote actors but helps for testing - sender ! RegisteredWorker("http://" + masterPublicAddress + ":" + webUiPort) + sender ! RegisteredWorker("http://" + masterPublicAddress + ":" + webUi.boundPort.get) schedule() } } @@ -244,7 +265,7 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act def addApplication(desc: ApplicationDescription, driver: ActorRef): ApplicationInfo = { val now = System.currentTimeMillis() val date = new Date(now) - val app = new ApplicationInfo(now, newApplicationId(date), desc, date, driver) + val app = new ApplicationInfo(now, newApplicationId(date), desc, date, driver, desc.appUiUrl) apps += app idToApp(app.id) = app actorToApp(driver) = app @@ -275,9 +296,12 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act for (exec <- app.executors.values) { exec.worker.removeExecutor(exec) exec.worker.actor ! KillExecutor(exec.application.id, exec.id) + exec.state = ExecutorState.KILLED } app.markFinished(state) - app.driver ! ApplicationRemoved(state.toString) + if (state != ApplicationState.FINISHED) { + app.driver ! ApplicationRemoved(state.toString) + } schedule() } } diff --git a/core/src/main/scala/spark/deploy/master/MasterArguments.scala b/core/src/main/scala/spark/deploy/master/MasterArguments.scala index 3d28ecabb4..0ae0160767 100644 --- a/core/src/main/scala/spark/deploy/master/MasterArguments.scala +++ b/core/src/main/scala/spark/deploy/master/MasterArguments.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.master import spark.util.IntParam @@ -21,7 +38,10 @@ private[spark] class MasterArguments(args: Array[String]) { if (System.getenv("SPARK_MASTER_WEBUI_PORT") != null) { webUiPort = System.getenv("SPARK_MASTER_WEBUI_PORT").toInt } - + if (System.getProperty("master.ui.port") != null) { + webUiPort = System.getProperty("master.ui.port").toInt + } + parse(args.toList) def parse(args: List[String]): Unit = args match { diff --git a/core/src/main/scala/spark/deploy/master/MasterSource.scala b/core/src/main/scala/spark/deploy/master/MasterSource.scala new file mode 100644 index 0000000000..b8cfa6a773 --- /dev/null +++ b/core/src/main/scala/spark/deploy/master/MasterSource.scala @@ -0,0 +1,25 @@ +package spark.deploy.master + +import com.codahale.metrics.{Gauge, MetricRegistry} + +import spark.metrics.source.Source + +private[spark] class MasterSource(val master: Master) extends Source { + val metricRegistry = new MetricRegistry() + val sourceName = "master" + + // Gauge for worker numbers in cluster + metricRegistry.register(MetricRegistry.name("workers","number"), new Gauge[Int] { + override def getValue: Int = master.workers.size + }) + + // Gauge for application numbers in cluster + metricRegistry.register(MetricRegistry.name("apps", "number"), new Gauge[Int] { + override def getValue: Int = master.apps.size + }) + + // Gauge for waiting application numbers in cluster + metricRegistry.register(MetricRegistry.name("waitingApps", "number"), new Gauge[Int] { + override def getValue: Int = master.waitingApps.size + }) +} diff --git a/core/src/main/scala/spark/deploy/master/MasterWebUI.scala b/core/src/main/scala/spark/deploy/master/MasterWebUI.scala deleted file mode 100644 index a4e21c8130..0000000000 --- a/core/src/main/scala/spark/deploy/master/MasterWebUI.scala +++ /dev/null @@ -1,74 +0,0 @@ -package spark.deploy.master - -import akka.actor.{ActorRef, ActorSystem} -import akka.dispatch.Await -import akka.pattern.ask -import akka.util.{Duration, Timeout} -import akka.util.duration._ -import cc.spray.Directives -import cc.spray.directives._ -import cc.spray.typeconversion.TwirlSupport._ -import cc.spray.http.MediaTypes -import cc.spray.typeconversion.SprayJsonSupport._ - -import spark.deploy._ -import spark.deploy.JsonProtocol._ - -/** - * Web UI server for the standalone master. - */ -private[spark] -class MasterWebUI(val actorSystem: ActorSystem, master: ActorRef) extends Directives { - val RESOURCE_DIR = "spark/deploy/master/webui" - val STATIC_RESOURCE_DIR = "spark/deploy/static" - - implicit val timeout = Timeout(Duration.create(System.getProperty("spark.akka.askTimeout", "10").toLong, "seconds")) - - val handler = { - get { - (path("") & parameters('format ?)) { - case Some(js) if js.equalsIgnoreCase("json") => - val future = master ? RequestMasterState - respondWithMediaType(MediaTypes.`application/json`) { ctx => - ctx.complete(future.mapTo[MasterState]) - } - case _ => - completeWith { - val future = master ? RequestMasterState - future.map { - masterState => spark.deploy.master.html.index.render(masterState.asInstanceOf[MasterState]) - } - } - } ~ - path("app") { - parameters("appId", 'format ?) { - case (appId, Some(js)) if (js.equalsIgnoreCase("json")) => - val future = master ? RequestMasterState - val appInfo = for (masterState <- future.mapTo[MasterState]) yield { - masterState.activeApps.find(_.id == appId).getOrElse({ - masterState.completedApps.find(_.id == appId).getOrElse(null) - }) - } - respondWithMediaType(MediaTypes.`application/json`) { ctx => - ctx.complete(appInfo.mapTo[ApplicationInfo]) - } - case (appId, _) => - completeWith { - val future = master ? RequestMasterState - future.map { state => - val masterState = state.asInstanceOf[MasterState] - val app = masterState.activeApps.find(_.id == appId).getOrElse({ - masterState.completedApps.find(_.id == appId).getOrElse(null) - }) - spark.deploy.master.html.app_details.render(app) - } - } - } - } ~ - pathPrefix("static") { - getFromResourceDirectory(STATIC_RESOURCE_DIR) - } ~ - getFromResourceDirectory(RESOURCE_DIR) - } - } -} diff --git a/core/src/main/scala/spark/deploy/master/WorkerInfo.scala b/core/src/main/scala/spark/deploy/master/WorkerInfo.scala index 0c08c5f417..4135cfeb28 100644 --- a/core/src/main/scala/spark/deploy/master/WorkerInfo.scala +++ b/core/src/main/scala/spark/deploy/master/WorkerInfo.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.master import akka.actor.ActorRef diff --git a/core/src/main/scala/spark/deploy/master/WorkerState.scala b/core/src/main/scala/spark/deploy/master/WorkerState.scala index 0bf35014c8..3e50b7748d 100644 --- a/core/src/main/scala/spark/deploy/master/WorkerState.scala +++ b/core/src/main/scala/spark/deploy/master/WorkerState.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.master private[spark] object WorkerState extends Enumeration("ALIVE", "DEAD", "DECOMMISSIONED") { diff --git a/core/src/main/scala/spark/deploy/master/ui/ApplicationPage.scala b/core/src/main/scala/spark/deploy/master/ui/ApplicationPage.scala new file mode 100644 index 0000000000..32264af393 --- /dev/null +++ b/core/src/main/scala/spark/deploy/master/ui/ApplicationPage.scala @@ -0,0 +1,117 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.deploy.master.ui + +import akka.dispatch.Await +import akka.pattern.ask +import akka.util.duration._ + +import javax.servlet.http.HttpServletRequest + +import net.liftweb.json.JsonAST.JValue + +import scala.xml.Node + +import spark.deploy.{RequestMasterState, JsonProtocol, MasterState} +import spark.deploy.master.ExecutorInfo +import spark.ui.UIUtils + +private[spark] class ApplicationPage(parent: MasterWebUI) { + val master = parent.master + implicit val timeout = parent.timeout + + /** Executor details for a particular application */ + def renderJson(request: HttpServletRequest): JValue = { + val appId = request.getParameter("appId") + val stateFuture = (master ? RequestMasterState)(timeout).mapTo[MasterState] + val state = Await.result(stateFuture, 30 seconds) + val app = state.activeApps.find(_.id == appId).getOrElse({ + state.completedApps.find(_.id == appId).getOrElse(null) + }) + JsonProtocol.writeApplicationInfo(app) + } + + /** Executor details for a particular application */ + def render(request: HttpServletRequest): Seq[Node] = { + val appId = request.getParameter("appId") + val stateFuture = (master ? RequestMasterState)(timeout).mapTo[MasterState] + val state = Await.result(stateFuture, 30 seconds) + val app = state.activeApps.find(_.id == appId).getOrElse({ + state.completedApps.find(_.id == appId).getOrElse(null) + }) + + val executorHeaders = Seq("ExecutorID", "Worker", "Cores", "Memory", "State", "Logs") + val executors = app.executors.values.toSeq + val executorTable = UIUtils.listingTable(executorHeaders, executorRow, executors) + + val content = + <hr /> + <div class="row"> + <div class="span12"> + <ul class="unstyled"> + <li><strong>ID:</strong> {app.id}</li> + <li><strong>Description:</strong> {app.desc.name}</li> + <li><strong>User:</strong> {app.desc.user}</li> + <li><strong>Cores:</strong> + { + if (app.desc.maxCores == Integer.MAX_VALUE) { + "Unlimited %s granted".format(app.coresGranted) + } else { + "%s (%s granted, %s left)".format( + app.desc.maxCores, app.coresGranted, app.coresLeft) + } + } + </li> + <li><strong>Memory per Slave:</strong> {app.desc.memoryPerSlave}</li> + <li><strong>Submit Date:</strong> {app.submitDate}</li> + <li><strong>State:</strong> {app.state}</li> + <li><strong><a href={app.appUiUrl}>Application Detail UI</a></strong></li> + </ul> + </div> + </div> + + <hr/> + + <div class="row"> <!-- Executors --> + <div class="span12"> + <h3> Executor Summary </h3> + <br/> + {executorTable} + </div> + </div>; + UIUtils.basicSparkPage(content, "Application Info: " + app.desc.name) + } + + def executorRow(executor: ExecutorInfo): Seq[Node] = { + <tr> + <td>{executor.id}</td> + <td> + <a href={executor.worker.webUiAddress}>{executor.worker.id}</a> + </td> + <td>{executor.cores}</td> + <td>{executor.memory}</td> + <td>{executor.state}</td> + <td> + <a href={"%s/logPage?appId=%s&executorId=%s&logType=stdout" + .format(executor.worker.webUiAddress, executor.application.id, executor.id)}>stdout</a> + <a href={"%s/logPage?appId=%s&executorId=%s&logType=stderr" + .format(executor.worker.webUiAddress, executor.application.id, executor.id)}>stderr</a> + </td> + </tr> + } +} diff --git a/core/src/main/scala/spark/deploy/master/ui/IndexPage.scala b/core/src/main/scala/spark/deploy/master/ui/IndexPage.scala new file mode 100644 index 0000000000..b05197c1b9 --- /dev/null +++ b/core/src/main/scala/spark/deploy/master/ui/IndexPage.scala @@ -0,0 +1,138 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.deploy.master.ui + +import akka.dispatch.Await +import akka.pattern.ask +import akka.util.duration._ + +import javax.servlet.http.HttpServletRequest + +import scala.xml.Node + +import spark.deploy.{RequestMasterState, DeployWebUI, MasterState} +import spark.Utils +import spark.ui.UIUtils +import spark.deploy.master.{ApplicationInfo, WorkerInfo} + +private[spark] class IndexPage(parent: MasterWebUI) { + val master = parent.master + implicit val timeout = parent.timeout + + /** Index view listing applications and executors */ + def render(request: HttpServletRequest): Seq[Node] = { + val stateFuture = (master ? RequestMasterState)(timeout).mapTo[MasterState] + val state = Await.result(stateFuture, 30 seconds) + + val workerHeaders = Seq("Id", "Address", "State", "Cores", "Memory") + val workers = state.workers.sortBy(_.id) + val workerTable = UIUtils.listingTable(workerHeaders, workerRow, workers) + + val appHeaders = Seq("ID", "Description", "Cores", "Memory per Node", "Submit Time", "User", + "State", "Duration") + val activeApps = state.activeApps.sortBy(_.startTime).reverse + val activeAppsTable = UIUtils.listingTable(appHeaders, appRow, activeApps) + val completedApps = state.completedApps.sortBy(_.endTime).reverse + val completedAppsTable = UIUtils.listingTable(appHeaders, appRow, completedApps) + + val content = + <hr /> + <div class="row"> + <div class="span12"> + <ul class="unstyled"> + <li><strong>URL:</strong> {state.uri}</li> + <li><strong>Workers:</strong> {state.workers.size}</li> + <li><strong>Cores:</strong> {state.workers.map(_.cores).sum} Total, + {state.workers.map(_.coresUsed).sum} Used</li> + <li><strong>Memory:</strong> + {Utils.memoryMegabytesToString(state.workers.map(_.memory).sum)} Total, + {Utils.memoryMegabytesToString(state.workers.map(_.memoryUsed).sum)} Used</li> + <li><strong>Applications:</strong> + {state.activeApps.size} Running, + {state.completedApps.size} Completed </li> + </ul> + </div> + </div> + + <div class="row"> + <div class="span12"> + <h3> Workers </h3> + <br/> + {workerTable} + </div> + </div> + + <hr/> + + <div class="row"> + <div class="span12"> + <h3> Running Applications </h3> + <br/> + {activeAppsTable} + </div> + </div> + + <hr/> + + <div class="row"> + <div class="span12"> + <h3> Completed Applications </h3> + <br/> + {completedAppsTable} + </div> + </div>; + UIUtils.basicSparkPage(content, "Spark Master: " + state.uri) + } + + def workerRow(worker: WorkerInfo): Seq[Node] = { + <tr> + <td> + <a href={worker.webUiAddress}>{worker.id}</a> + </td> + <td>{worker.host}:{worker.port}</td> + <td>{worker.state}</td> + <td>{worker.cores} ({worker.coresUsed} Used)</td> + <td sorttable_customkey={"%s.%s".format(worker.memory, worker.memoryUsed)}> + {Utils.memoryMegabytesToString(worker.memory)} + ({Utils.memoryMegabytesToString(worker.memoryUsed)} Used) + </td> + </tr> + } + + + def appRow(app: ApplicationInfo): Seq[Node] = { + <tr> + <td> + <a href={"app?appId=" + app.id}>{app.id}</a> + </td> + <td> + <a href={app.appUiUrl}>{app.desc.name}</a> + </td> + <td> + {app.coresGranted} + </td> + <td sorttable_customkey={app.desc.memoryPerSlave.toString}> + {Utils.memoryMegabytesToString(app.desc.memoryPerSlave)} + </td> + <td>{DeployWebUI.formatDate(app.submitDate)}</td> + <td>{app.desc.user}</td> + <td>{app.state.toString}</td> + <td>{DeployWebUI.formatDuration(app.duration)}</td> + </tr> + } +} diff --git a/core/src/main/scala/spark/deploy/master/ui/MasterWebUI.scala b/core/src/main/scala/spark/deploy/master/ui/MasterWebUI.scala new file mode 100644 index 0000000000..dabc2d8dc7 --- /dev/null +++ b/core/src/main/scala/spark/deploy/master/ui/MasterWebUI.scala @@ -0,0 +1,74 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.deploy.master.ui + +import akka.actor.ActorRef +import akka.util.Duration + +import javax.servlet.http.HttpServletRequest + +import org.eclipse.jetty.server.{Handler, Server} + +import spark.{Logging, Utils} +import spark.ui.JettyUtils +import spark.ui.JettyUtils._ + +/** + * Web UI server for the standalone master. + */ +private[spark] +class MasterWebUI(val master: ActorRef, requestedPort: Int) extends Logging { + implicit val timeout = Duration.create( + System.getProperty("spark.akka.askTimeout", "10").toLong, "seconds") + val host = Utils.localHostName() + val port = requestedPort + + var server: Option[Server] = None + var boundPort: Option[Int] = None + + val applicationPage = new ApplicationPage(this) + val indexPage = new IndexPage(this) + + def start() { + try { + val (srv, bPort) = JettyUtils.startJettyServer("0.0.0.0", port, handlers) + server = Some(srv) + boundPort = Some(bPort) + logInfo("Started Master web UI at http://%s:%d".format(host, boundPort.get)) + } catch { + case e: Exception => + logError("Failed to create Master JettyUtils", e) + System.exit(1) + } + } + + val handlers = Array[(String, Handler)]( + ("/static", createStaticHandler(MasterWebUI.STATIC_RESOURCE_DIR)), + ("/app/json", (request: HttpServletRequest) => applicationPage.renderJson(request)), + ("/app", (request: HttpServletRequest) => applicationPage.render(request)), + ("*", (request: HttpServletRequest) => indexPage.render(request)) + ) + + def stop() { + server.foreach(_.stop()) + } +} + +private[spark] object MasterWebUI { + val STATIC_RESOURCE_DIR = "spark/ui/static" +} diff --git a/core/src/main/scala/spark/deploy/worker/ExecutorRunner.scala b/core/src/main/scala/spark/deploy/worker/ExecutorRunner.scala index 04a774658e..47d3390928 100644 --- a/core/src/main/scala/spark/deploy/worker/ExecutorRunner.scala +++ b/core/src/main/scala/spark/deploy/worker/ExecutorRunner.scala @@ -1,6 +1,24 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.worker import java.io._ +import java.lang.System.getenv import spark.deploy.{ExecutorState, ExecutorStateChanged, ApplicationDescription} import akka.actor.ActorRef import spark.{Utils, Logging} @@ -33,6 +51,9 @@ private[spark] class ExecutorRunner( var process: Process = null var shutdownHook: Thread = null + private def getAppEnv(key: String): Option[String] = + appDesc.command.environment.get(key).orElse(Option(getenv(key))) + def start() { workerThread = new Thread("ExecutorRunner for " + fullId) { override def run() { fetchAndRunExecutor() } @@ -40,7 +61,7 @@ private[spark] class ExecutorRunner( workerThread.start() // Shutdown hook that kills actors on shutdown. - shutdownHook = new Thread() { + shutdownHook = new Thread() { override def run() { if (process != null) { logInfo("Shutdown hook killing child process.") @@ -77,9 +98,31 @@ private[spark] class ExecutorRunner( def buildCommandSeq(): Seq[String] = { val command = appDesc.command - val script = if (System.getProperty("os.name").startsWith("Windows")) "run.cmd" else "run" - val runScript = new File(sparkHome, script).getCanonicalPath - Seq(runScript, command.mainClass) ++ (command.arguments ++ Seq(appId)).map(substituteVariables) + val runner = getAppEnv("JAVA_HOME").map(_ + "/bin/java").getOrElse("java") + // SPARK-698: do not call the run.cmd script, as process.destroy() + // fails to kill a process tree on Windows + Seq(runner) ++ buildJavaOpts() ++ Seq(command.mainClass) ++ + command.arguments.map(substituteVariables) + } + + /** + * Attention: this must always be aligned with the environment variables in the run scripts and + * the way the JAVA_OPTS are assembled there. + */ + def buildJavaOpts(): Seq[String] = { + val libraryOpts = getAppEnv("SPARK_LIBRARY_PATH") + .map(p => List("-Djava.library.path=" + p)) + .getOrElse(Nil) + val userOpts = getAppEnv("SPARK_JAVA_OPTS").map(Utils.splitCommandString).getOrElse(Nil) + val memoryOpts = Seq("-Xms" + memory + "M", "-Xmx" + memory + "M") + + // Figure out our classpath with the external compute-classpath script + val ext = if (System.getProperty("os.name").startsWith("Windows")) ".cmd" else ".sh" + val classPath = Utils.executeAndGetOutput( + Seq(sparkHome + "/bin/compute-classpath" + ext), + extraEnvironment=appDesc.command.environment) + + Seq("-cp", classPath) ++ libraryOpts ++ userOpts ++ memoryOpts } /** Spawn a thread that will redirect a given stream to a file */ @@ -115,7 +158,6 @@ private[spark] class ExecutorRunner( for ((key, value) <- appDesc.command.environment) { env.put(key, value) } - env.put("SPARK_MEM", memory.toString + "m") // In case we are running this from within the Spark Shell, avoid creating a "scala" // parent process for the executor command env.put("SPARK_LAUNCH_WITH_SCALA", "0") diff --git a/core/src/main/scala/spark/deploy/worker/Worker.scala b/core/src/main/scala/spark/deploy/worker/Worker.scala index 1a7da0f7bf..8fa0d12b82 100644 --- a/core/src/main/scala/spark/deploy/worker/Worker.scala +++ b/core/src/main/scala/spark/deploy/worker/Worker.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.worker import scala.collection.mutable.{ArrayBuffer, HashMap} @@ -6,6 +23,7 @@ import akka.util.duration._ import spark.{Logging, Utils} import spark.util.AkkaUtils import spark.deploy._ +import spark.metrics.MetricsSystem import akka.remote.{RemoteClientLifeCycleEvent, RemoteClientShutdown, RemoteClientDisconnected} import java.text.SimpleDateFormat import java.util.Date @@ -14,6 +32,7 @@ import spark.deploy.LaunchExecutor import spark.deploy.RegisterWorkerFailed import spark.deploy.master.Master import java.io.File +import ui.WorkerWebUI private[spark] class Worker( host: String, @@ -44,17 +63,24 @@ private[spark] class Worker( val envVar = System.getenv("SPARK_PUBLIC_DNS") if (envVar != null) envVar else host } + var webUi: WorkerWebUI = null var coresUsed = 0 var memoryUsed = 0 + val metricsSystem = MetricsSystem.createMetricsSystem("worker") + val workerSource = new WorkerSource(this) + def coresFree: Int = cores - coresUsed def memoryFree: Int = memory - memoryUsed def createWorkDir() { workDir = Option(workDirPath).map(new File(_)).getOrElse(new File(sparkHome, "work")) try { - if ( (workDir.exists() && !workDir.isDirectory) || (!workDir.exists() && !workDir.mkdirs()) ) { + // This sporadically fails - not sure why ... !workDir.exists() && !workDir.mkdirs() + // So attempting to create and then check if directory was created or not. + workDir.mkdirs() + if ( !workDir.exists() || !workDir.isDirectory) { logError("Failed to create work directory " + workDir) System.exit(1) } @@ -72,29 +98,22 @@ private[spark] class Worker( sparkHome = new File(Option(System.getenv("SPARK_HOME")).getOrElse(".")) logInfo("Spark home: " + sparkHome) createWorkDir() + webUi = new WorkerWebUI(this, workDir, Some(webUiPort)) + webUi.start() connectToMaster() - startWebUi() + + metricsSystem.registerSource(workerSource) + metricsSystem.start() } def connectToMaster() { logInfo("Connecting to master " + masterUrl) master = context.actorFor(Master.toAkkaUrl(masterUrl)) - master ! RegisterWorker(workerId, host, port, cores, memory, webUiPort, publicAddress) + master ! RegisterWorker(workerId, host, port, cores, memory, webUi.boundPort.get, publicAddress) context.system.eventStream.subscribe(self, classOf[RemoteClientLifeCycleEvent]) context.watch(master) // Doesn't work with remote actors, but useful for testing } - def startWebUi() { - val webUi = new WorkerWebUI(context.system, self, workDir) - try { - AkkaUtils.startSprayServer(context.system, "0.0.0.0", webUiPort, webUi.handler) - } catch { - case e: Exception => - logError("Failed to create web UI", e) - System.exit(1) - } - } - override def receive = { case RegisteredWorker(url) => masterWebUiUrl = url @@ -143,10 +162,10 @@ private[spark] class Worker( case Terminated(_) | RemoteClientDisconnected(_, _) | RemoteClientShutdown(_, _) => masterDisconnected() - + case RequestWorkerState => { sender ! WorkerState(host, port, workerId, executors.values.toList, - finishedExecutors.values.toList, masterUrl, cores, memory, + finishedExecutors.values.toList, masterUrl, cores, memory, coresUsed, memoryUsed, masterWebUiUrl) } } @@ -165,6 +184,8 @@ private[spark] class Worker( override def postStop() { executors.values.foreach(_.kill()) + webUi.stop() + metricsSystem.stop() } } diff --git a/core/src/main/scala/spark/deploy/worker/WorkerArguments.scala b/core/src/main/scala/spark/deploy/worker/WorkerArguments.scala index 2b96611ee3..9fcd3260ca 100644 --- a/core/src/main/scala/spark/deploy/worker/WorkerArguments.scala +++ b/core/src/main/scala/spark/deploy/worker/WorkerArguments.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.deploy.worker import spark.util.IntParam diff --git a/core/src/main/scala/spark/deploy/worker/WorkerSource.scala b/core/src/main/scala/spark/deploy/worker/WorkerSource.scala new file mode 100644 index 0000000000..39cb8e5690 --- /dev/null +++ b/core/src/main/scala/spark/deploy/worker/WorkerSource.scala @@ -0,0 +1,34 @@ +package spark.deploy.worker + +import com.codahale.metrics.{Gauge, MetricRegistry} + +import spark.metrics.source.Source + +private[spark] class WorkerSource(val worker: Worker) extends Source { + val sourceName = "worker" + val metricRegistry = new MetricRegistry() + + metricRegistry.register(MetricRegistry.name("executors", "number"), new Gauge[Int] { + override def getValue: Int = worker.executors.size + }) + + // Gauge for cores used of this worker + metricRegistry.register(MetricRegistry.name("coresUsed", "number"), new Gauge[Int] { + override def getValue: Int = worker.coresUsed + }) + + // Gauge for memory used of this worker + metricRegistry.register(MetricRegistry.name("memUsed", "MBytes"), new Gauge[Int] { + override def getValue: Int = worker.memoryUsed + }) + + // Gauge for cores free of this worker + metricRegistry.register(MetricRegistry.name("coresFree", "number"), new Gauge[Int] { + override def getValue: Int = worker.coresFree + }) + + // Gauge for memory free of this worker + metricRegistry.register(MetricRegistry.name("memFree", "MBytes"), new Gauge[Int] { + override def getValue: Int = worker.memoryFree + }) +} diff --git a/core/src/main/scala/spark/deploy/worker/WorkerWebUI.scala b/core/src/main/scala/spark/deploy/worker/WorkerWebUI.scala deleted file mode 100644 index 3235c50d1b..0000000000 --- a/core/src/main/scala/spark/deploy/worker/WorkerWebUI.scala +++ /dev/null @@ -1,57 +0,0 @@ -package spark.deploy.worker - -import akka.actor.{ActorRef, ActorSystem} -import akka.dispatch.Await -import akka.pattern.ask -import akka.util.{Duration, Timeout} -import akka.util.duration._ -import cc.spray.Directives -import cc.spray.typeconversion.TwirlSupport._ -import cc.spray.http.MediaTypes -import cc.spray.typeconversion.SprayJsonSupport._ - -import spark.deploy.{WorkerState, RequestWorkerState} -import spark.deploy.JsonProtocol._ -import java.io.File - -/** - * Web UI server for the standalone worker. - */ -private[spark] -class WorkerWebUI(val actorSystem: ActorSystem, worker: ActorRef, workDir: File) extends Directives { - val RESOURCE_DIR = "spark/deploy/worker/webui" - val STATIC_RESOURCE_DIR = "spark/deploy/static" - - implicit val timeout = Timeout(Duration.create(System.getProperty("spark.akka.askTimeout", "10").toLong, "seconds")) - - val handler = { - get { - (path("") & parameters('format ?)) { - case Some(js) if js.equalsIgnoreCase("json") => { - val future = worker ? RequestWorkerState - respondWithMediaType(MediaTypes.`application/json`) { ctx => - ctx.complete(future.mapTo[WorkerState]) - } - } - case _ => - completeWith{ - val future = worker ? RequestWorkerState - future.map { workerState => - spark.deploy.worker.html.index(workerState.asInstanceOf[WorkerState]) - } - } - } ~ - path("log") { - parameters("appId", "executorId", "logType") { (appId, executorId, logType) => - respondWithMediaType(cc.spray.http.MediaTypes.`text/plain`) { - getFromFileName(workDir.getPath() + "/" + appId + "/" + executorId + "/" + logType) - } - } - } ~ - pathPrefix("static") { - getFromResourceDirectory(STATIC_RESOURCE_DIR) - } ~ - getFromResourceDirectory(RESOURCE_DIR) - } - } -} diff --git a/core/src/main/scala/spark/deploy/worker/ui/IndexPage.scala b/core/src/main/scala/spark/deploy/worker/ui/IndexPage.scala new file mode 100644 index 0000000000..7548a26c2e --- /dev/null +++ b/core/src/main/scala/spark/deploy/worker/ui/IndexPage.scala @@ -0,0 +1,117 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.deploy.worker.ui + +import akka.dispatch.Await +import akka.pattern.ask +import akka.util.duration._ + +import javax.servlet.http.HttpServletRequest + +import net.liftweb.json.JsonAST.JValue + +import scala.xml.Node + +import spark.deploy.{RequestWorkerState, JsonProtocol, WorkerState} +import spark.deploy.worker.ExecutorRunner +import spark.Utils +import spark.ui.UIUtils + +private[spark] class IndexPage(parent: WorkerWebUI) { + val workerActor = parent.worker.self + val worker = parent.worker + val timeout = parent.timeout + + def renderJson(request: HttpServletRequest): JValue = { + val stateFuture = (workerActor ? RequestWorkerState)(timeout).mapTo[WorkerState] + val workerState = Await.result(stateFuture, 30 seconds) + JsonProtocol.writeWorkerState(workerState) + } + + def render(request: HttpServletRequest): Seq[Node] = { + val stateFuture = (workerActor ? RequestWorkerState)(timeout).mapTo[WorkerState] + val workerState = Await.result(stateFuture, 30 seconds) + + val executorHeaders = Seq("ExecutorID", "Cores", "Memory", "Job Details", "Logs") + val runningExecutorTable = + UIUtils.listingTable(executorHeaders, executorRow, workerState.executors) + val finishedExecutorTable = + UIUtils.listingTable(executorHeaders, executorRow, workerState.finishedExecutors) + + val content = + <hr /> + <div class="row"> <!-- Worker Details --> + <div class="span12"> + <ul class="unstyled"> + <li><strong>ID:</strong> {workerState.workerId}</li> + <li><strong> + Master URL:</strong> {workerState.masterUrl} + </li> + <li><strong>Cores:</strong> {workerState.cores} ({workerState.coresUsed} Used)</li> + <li><strong>Memory:</strong> {Utils.memoryMegabytesToString(workerState.memory)} + ({Utils.memoryMegabytesToString(workerState.memoryUsed)} Used)</li> + </ul> + <p><a href={workerState.masterWebUiUrl}>Back to Master</a></p> + </div> + </div> + <hr/> + + <div class="row"> <!-- Running Executors --> + <div class="span12"> + <h3> Running Executors {workerState.executors.size} </h3> + <br/> + {runningExecutorTable} + </div> + </div> + <hr/> + + <div class="row"> <!-- Finished Executors --> + <div class="span12"> + <h3> Finished Executors </h3> + <br/> + {finishedExecutorTable} + </div> + </div>; + + UIUtils.basicSparkPage(content, "Spark Worker on %s:%s".format(workerState.host, workerState.port)) + } + + def executorRow(executor: ExecutorRunner): Seq[Node] = { + <tr> + <td>{executor.execId}</td> + <td>{executor.cores}</td> + <td sorttable_customkey={executor.memory.toString}> + {Utils.memoryMegabytesToString(executor.memory)} + </td> + <td> + <ul class="unstyled"> + <li><strong>ID:</strong> {executor.appId}</li> + <li><strong>Name:</strong> {executor.appDesc.name}</li> + <li><strong>User:</strong> {executor.appDesc.user}</li> + </ul> + </td> + <td> + <a href={"logPage?appId=%s&executorId=%s&logType=stdout" + .format(executor.appId, executor.execId)}>stdout</a> + <a href={"logPage?appId=%s&executorId=%s&logType=stderr" + .format(executor.appId, executor.execId)}>stderr</a> + </td> + </tr> + } + +} diff --git a/core/src/main/scala/spark/deploy/worker/ui/WorkerWebUI.scala b/core/src/main/scala/spark/deploy/worker/ui/WorkerWebUI.scala new file mode 100644 index 0000000000..61d4cd6d99 --- /dev/null +++ b/core/src/main/scala/spark/deploy/worker/ui/WorkerWebUI.scala @@ -0,0 +1,182 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.deploy.worker.ui + +import akka.actor.ActorRef +import akka.util.{Duration, Timeout} + +import java.io.{FileInputStream, File} + +import javax.servlet.http.HttpServletRequest + +import org.eclipse.jetty.server.{Handler, Server} + +import spark.deploy.worker.Worker +import spark.{Utils, Logging} +import spark.ui.JettyUtils +import spark.ui.JettyUtils._ +import spark.ui.UIUtils + +/** + * Web UI server for the standalone worker. + */ +private[spark] +class WorkerWebUI(val worker: Worker, val workDir: File, requestedPort: Option[Int] = None) + extends Logging { + implicit val timeout = Timeout( + Duration.create(System.getProperty("spark.akka.askTimeout", "10").toLong, "seconds")) + val host = Utils.localHostName() + val port = requestedPort.getOrElse( + System.getProperty("worker.ui.port", WorkerWebUI.DEFAULT_PORT).toInt) + + var server: Option[Server] = None + var boundPort: Option[Int] = None + + val indexPage = new IndexPage(this) + + val handlers = Array[(String, Handler)]( + ("/static", createStaticHandler(WorkerWebUI.STATIC_RESOURCE_DIR)), + ("/log", (request: HttpServletRequest) => log(request)), + ("/logPage", (request: HttpServletRequest) => logPage(request)), + ("/json", (request: HttpServletRequest) => indexPage.renderJson(request)), + ("*", (request: HttpServletRequest) => indexPage.render(request)) + ) + + def start() { + try { + val (srv, bPort) = JettyUtils.startJettyServer("0.0.0.0", port, handlers) + server = Some(srv) + boundPort = Some(bPort) + logInfo("Started Worker web UI at http://%s:%d".format(host, bPort)) + } catch { + case e: Exception => + logError("Failed to create Worker JettyUtils", e) + System.exit(1) + } + } + + def log(request: HttpServletRequest): String = { + val defaultBytes = 100 * 1024 + val appId = request.getParameter("appId") + val executorId = request.getParameter("executorId") + val logType = request.getParameter("logType") + val offset = Option(request.getParameter("offset")).map(_.toLong) + val byteLength = Option(request.getParameter("byteLength")).map(_.toInt).getOrElse(defaultBytes) + val path = "%s/%s/%s/%s".format(workDir.getPath, appId, executorId, logType) + + val (startByte, endByte) = getByteRange(path, offset, byteLength) + val file = new File(path) + val logLength = file.length + + val pre = "==== Bytes %s-%s of %s of %s/%s/%s ====\n" + .format(startByte, endByte, logLength, appId, executorId, logType) + pre + Utils.offsetBytes(path, startByte, endByte) + } + + def logPage(request: HttpServletRequest): Seq[scala.xml.Node] = { + val defaultBytes = 100 * 1024 + val appId = request.getParameter("appId") + val executorId = request.getParameter("executorId") + val logType = request.getParameter("logType") + val offset = Option(request.getParameter("offset")).map(_.toLong) + val byteLength = Option(request.getParameter("byteLength")).map(_.toInt).getOrElse(defaultBytes) + val path = "%s/%s/%s/%s".format(workDir.getPath, appId, executorId, logType) + + val (startByte, endByte) = getByteRange(path, offset, byteLength) + val file = new File(path) + val logLength = file.length + + val logText = <node>{Utils.offsetBytes(path, startByte, endByte)}</node> + + val linkToMaster = <p><a href={worker.masterWebUiUrl}>Back to Master</a></p> + + val range = <span>Bytes {startByte.toString} - {endByte.toString} of {logLength}</span> + + val backButton = + if (startByte > 0) { + <a href={"?appId=%s&executorId=%s&logType=%s&offset=%s&byteLength=%s" + .format(appId, executorId, logType, math.max(startByte-byteLength, 0), + byteLength)}> + <button>Previous {Utils.memoryBytesToString(math.min(byteLength, startByte))}</button> + </a> + } + else { + <button disabled="disabled">Previous 0 B</button> + } + + val nextButton = + if (endByte < logLength) { + <a href={"?appId=%s&executorId=%s&logType=%s&offset=%s&byteLength=%s". + format(appId, executorId, logType, endByte, byteLength)}> + <button>Next {Utils.memoryBytesToString(math.min(byteLength, logLength-endByte))}</button> + </a> + } + else { + <button disabled="disabled">Next 0 B</button> + } + + val content = + <html> + <body> + {linkToMaster} + <hr /> + <div> + <div style="float:left;width:40%">{backButton}</div> + <div style="float:left;">{range}</div> + <div style="float:right;">{nextButton}</div> + </div> + <br /> + <div style="height:500px;overflow:auto;padding:5px;"> + <pre>{logText}</pre> + </div> + </body> + </html> + UIUtils.basicSparkPage(content, logType + " log page for " + appId) + } + + /** Determine the byte range for a log or log page. */ + def getByteRange(path: String, offset: Option[Long], byteLength: Int) + : (Long, Long) = { + val defaultBytes = 100 * 1024 + val maxBytes = 1024 * 1024 + + val file = new File(path) + val logLength = file.length() + val getOffset = offset.getOrElse(logLength-defaultBytes) + + val startByte = + if (getOffset < 0) 0L + else if (getOffset > logLength) logLength + else getOffset + + val logPageLength = math.min(byteLength, maxBytes) + + val endByte = math.min(startByte+logPageLength, logLength) + + (startByte, endByte) + } + + def stop() { + server.foreach(_.stop()) + } +} + +private[spark] object WorkerWebUI { + val STATIC_RESOURCE_DIR = "spark/ui/static" + val DEFAULT_PORT="8081" +} diff --git a/core/src/main/scala/spark/executor/Executor.scala b/core/src/main/scala/spark/executor/Executor.scala index 344face5e6..8a74a8d853 100644 --- a/core/src/main/scala/spark/executor/Executor.scala +++ b/core/src/main/scala/spark/executor/Executor.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.executor import java.io.{File, FileOutputStream} @@ -42,7 +59,8 @@ private[spark] class Executor(executorId: String, slaveHostname: String, propert // Create our ClassLoader and set it on this thread private val urlClassLoader = createClassLoader() - Thread.currentThread.setContextClassLoader(urlClassLoader) + private val replClassLoader = addReplClassLoaderIfNeeded(urlClassLoader) + Thread.currentThread.setContextClassLoader(replClassLoader) // Make any thread terminations due to uncaught exceptions kill the entire // executor process to avoid surprising stalls. @@ -51,7 +69,7 @@ private[spark] class Executor(executorId: String, slaveHostname: String, propert override def uncaughtException(thread: Thread, exception: Throwable) { try { logError("Uncaught exception in thread " + thread, exception) - + // We may have been called from a shutdown hook. If so, we must not call System.exit(). // (If we do, we will deadlock.) if (!Utils.inShutdown()) { @@ -69,9 +87,14 @@ private[spark] class Executor(executorId: String, slaveHostname: String, propert } ) + val executorSource = new ExecutorSource(this) + // Initialize Spark environment (using system properties read above) val env = SparkEnv.createFromSystemProperties(executorId, slaveHostname, 0, false, false) SparkEnv.set(env) + env.metricsSystem.registerSource(executorSource) + + private val akkaFrameSize = env.actorSystem.settings.config.getBytes("akka.remote.netty.message-frame-size") // Start worker thread pool val threadPool = new ThreadPoolExecutor( @@ -87,22 +110,26 @@ private[spark] class Executor(executorId: String, slaveHostname: String, propert override def run() { val startTime = System.currentTimeMillis() SparkEnv.set(env) - Thread.currentThread.setContextClassLoader(urlClassLoader) + Thread.currentThread.setContextClassLoader(replClassLoader) val ser = SparkEnv.get.closureSerializer.newInstance() logInfo("Running task ID " + taskId) context.statusUpdate(taskId, TaskState.RUNNING, EMPTY_BYTE_BUFFER) + var attemptedTask: Option[Task[Any]] = None + var taskStart: Long = 0 try { SparkEnv.set(env) Accumulators.clear() val (taskFiles, taskJars, taskBytes) = Task.deserializeWithDependencies(serializedTask) updateDependencies(taskFiles, taskJars) val task = ser.deserialize[Task[Any]](taskBytes, Thread.currentThread.getContextClassLoader) + attemptedTask = Some(task) logInfo("Its generation is " + task.generation) env.mapOutputTracker.updateGeneration(task.generation) - val taskStart = System.currentTimeMillis() + taskStart = System.currentTimeMillis() val value = task.run(taskId.toInt) val taskFinish = System.currentTimeMillis() task.metrics.foreach{ m => + m.hostname = Utils.localHostName m.executorDeserializeTime = (taskStart - startTime).toInt m.executorRunTime = (taskFinish - taskStart).toInt } @@ -113,6 +140,10 @@ private[spark] class Executor(executorId: String, slaveHostname: String, propert val result = new TaskResult(value, accumUpdates, task.metrics.getOrElse(null)) val serializedResult = ser.serialize(result) logInfo("Serialized size of result for " + taskId + " is " + serializedResult.limit) + if (serializedResult.limit >= (akkaFrameSize - 1024)) { + context.statusUpdate(taskId, TaskState.FAILED, ser.serialize(TaskResultTooBigFailure())) + return + } context.statusUpdate(taskId, TaskState.FINISHED, serializedResult) logInfo("Finished task ID " + taskId) } catch { @@ -122,7 +153,10 @@ private[spark] class Executor(executorId: String, slaveHostname: String, propert } case t: Throwable => { - val reason = ExceptionFailure(t) + val serviceTime = (System.currentTimeMillis() - taskStart).toInt + val metrics = attemptedTask.flatMap(t => t.metrics) + metrics.foreach{m => m.executorRunTime = serviceTime} + val reason = ExceptionFailure(t.getClass.getName, t.toString, t.getStackTrace, metrics) context.statusUpdate(taskId, TaskState.FAILED, ser.serialize(reason)) // TODO: Should we exit the whole executor here? On the one hand, the failed task may @@ -147,26 +181,31 @@ private[spark] class Executor(executorId: String, slaveHostname: String, propert val urls = currentJars.keySet.map { uri => new File(uri.split("/").last).toURI.toURL }.toArray - loader = new URLClassLoader(urls, loader) + new ExecutorURLClassLoader(urls, loader) + } - // If the REPL is in use, add another ClassLoader that will read - // new classes defined by the REPL as the user types code + /** + * If the REPL is in use, add another ClassLoader that will read + * new classes defined by the REPL as the user types code + */ + private def addReplClassLoaderIfNeeded(parent: ClassLoader): ClassLoader = { val classUri = System.getProperty("spark.repl.class.uri") if (classUri != null) { logInfo("Using REPL class URI: " + classUri) - loader = { - try { - val klass = Class.forName("spark.repl.ExecutorClassLoader") - .asInstanceOf[Class[_ <: ClassLoader]] - val constructor = klass.getConstructor(classOf[String], classOf[ClassLoader]) - constructor.newInstance(classUri, loader) - } catch { - case _: ClassNotFoundException => loader - } + try { + val klass = Class.forName("spark.repl.ExecutorClassLoader") + .asInstanceOf[Class[_ <: ClassLoader]] + val constructor = klass.getConstructor(classOf[String], classOf[ClassLoader]) + return constructor.newInstance(classUri, parent) + } catch { + case _: ClassNotFoundException => + logError("Could not find spark.repl.ExecutorClassLoader on classpath!") + System.exit(1) + null } + } else { + return parent } - - return new ExecutorURLClassLoader(Array(), loader) } /** diff --git a/core/src/main/scala/spark/executor/ExecutorBackend.scala b/core/src/main/scala/spark/executor/ExecutorBackend.scala index e97e509700..33a6f8a824 100644 --- a/core/src/main/scala/spark/executor/ExecutorBackend.scala +++ b/core/src/main/scala/spark/executor/ExecutorBackend.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.executor import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/executor/ExecutorExitCode.scala b/core/src/main/scala/spark/executor/ExecutorExitCode.scala index fd76029cb3..64b9fb88f8 100644 --- a/core/src/main/scala/spark/executor/ExecutorExitCode.scala +++ b/core/src/main/scala/spark/executor/ExecutorExitCode.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.executor /** diff --git a/core/src/main/scala/spark/executor/ExecutorSource.scala b/core/src/main/scala/spark/executor/ExecutorSource.scala new file mode 100644 index 0000000000..94116edfcf --- /dev/null +++ b/core/src/main/scala/spark/executor/ExecutorSource.scala @@ -0,0 +1,30 @@ +package spark.executor + +import com.codahale.metrics.{Gauge, MetricRegistry} + +import spark.metrics.source.Source + +class ExecutorSource(val executor: Executor) extends Source { + val metricRegistry = new MetricRegistry() + val sourceName = "executor" + + // Gauge for executor thread pool's actively executing task counts + metricRegistry.register(MetricRegistry.name("threadpool", "activeTask", "count"), new Gauge[Int] { + override def getValue: Int = executor.threadPool.getActiveCount() + }) + + // Gauge for executor thread pool's approximate total number of tasks that have been completed + metricRegistry.register(MetricRegistry.name("threadpool", "completeTask", "count"), new Gauge[Long] { + override def getValue: Long = executor.threadPool.getCompletedTaskCount() + }) + + // Gauge for executor thread pool's current number of threads + metricRegistry.register(MetricRegistry.name("threadpool", "currentPool", "size"), new Gauge[Int] { + override def getValue: Int = executor.threadPool.getPoolSize() + }) + + // Gauge got executor thread pool's largest number of threads that have ever simultaneously been in th pool + metricRegistry.register(MetricRegistry.name("threadpool", "maxPool", "size"), new Gauge[Int] { + override def getValue: Int = executor.threadPool.getMaximumPoolSize() + }) +} diff --git a/core/src/main/scala/spark/executor/ExecutorURLClassLoader.scala b/core/src/main/scala/spark/executor/ExecutorURLClassLoader.scala index 5beb4d049e..09d12fb65b 100644 --- a/core/src/main/scala/spark/executor/ExecutorURLClassLoader.scala +++ b/core/src/main/scala/spark/executor/ExecutorURLClassLoader.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.executor import java.net.{URLClassLoader, URL} diff --git a/core/src/main/scala/spark/executor/MesosExecutorBackend.scala b/core/src/main/scala/spark/executor/MesosExecutorBackend.scala index 10f3531df0..4961c42fad 100644 --- a/core/src/main/scala/spark/executor/MesosExecutorBackend.scala +++ b/core/src/main/scala/spark/executor/MesosExecutorBackend.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.executor import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/executor/StandaloneExecutorBackend.scala b/core/src/main/scala/spark/executor/StandaloneExecutorBackend.scala index ebe2ac68d8..f4003da732 100644 --- a/core/src/main/scala/spark/executor/StandaloneExecutorBackend.scala +++ b/core/src/main/scala/spark/executor/StandaloneExecutorBackend.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.executor import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/executor/TaskMetrics.scala b/core/src/main/scala/spark/executor/TaskMetrics.scala index a7c56c2371..3151627839 100644 --- a/core/src/main/scala/spark/executor/TaskMetrics.scala +++ b/core/src/main/scala/spark/executor/TaskMetrics.scala @@ -1,7 +1,29 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.executor class TaskMetrics extends Serializable { /** + * Host's name the task runs on + */ + var hostname: String = _ + + /** * Time taken on the executor to deserialize this task */ var executorDeserializeTime: Int = _ @@ -34,9 +56,14 @@ object TaskMetrics { class ShuffleReadMetrics extends Serializable { /** + * Time when shuffle finishs + */ + var shuffleFinishTime: Long = _ + + /** * Total number of blocks fetched in a shuffle (remote or local) */ - var totalBlocksFetched : Int = _ + var totalBlocksFetched: Int = _ /** * Number of remote blocks fetched in a shuffle diff --git a/core/src/main/scala/spark/metrics/MetricsConfig.scala b/core/src/main/scala/spark/metrics/MetricsConfig.scala new file mode 100644 index 0000000000..ed505b0aa7 --- /dev/null +++ b/core/src/main/scala/spark/metrics/MetricsConfig.scala @@ -0,0 +1,79 @@ +package spark.metrics + +import java.util.Properties +import java.io.{File, FileInputStream, InputStream, IOException} + +import scala.collection.mutable +import scala.util.matching.Regex + +import spark.Logging + +private[spark] class MetricsConfig(val configFile: Option[String]) extends Logging { + initLogging() + + val DEFAULT_PREFIX = "*" + val INSTANCE_REGEX = "^(\\*|[a-zA-Z]+)\\.(.+)".r + val METRICS_CONF = "metrics.properties" + + val properties = new Properties() + var propertyCategories: mutable.HashMap[String, Properties] = null + + private def setDefaultProperties(prop: Properties) { + // empty function, any default property can be set here + } + + def initialize() { + //Add default properties in case there's no properties file + setDefaultProperties(properties) + + // If spark.metrics.conf is not set, try to get file in class path + var is: InputStream = null + try { + is = configFile match { + case Some(f) => new FileInputStream(f) + case None => getClass.getClassLoader.getResourceAsStream(METRICS_CONF) + } + + if (is != null) { + properties.load(is) + } + } catch { + case e: Exception => logError("Error loading configure file", e) + } finally { + if (is != null) is.close() + } + + propertyCategories = subProperties(properties, INSTANCE_REGEX) + if (propertyCategories.contains(DEFAULT_PREFIX)) { + import scala.collection.JavaConversions._ + + val defaultProperty = propertyCategories(DEFAULT_PREFIX) + for { (inst, prop) <- propertyCategories + if (inst != DEFAULT_PREFIX) + (k, v) <- defaultProperty + if (prop.getProperty(k) == null) } { + prop.setProperty(k, v) + } + } + } + + def subProperties(prop: Properties, regex: Regex): mutable.HashMap[String, Properties] = { + val subProperties = new mutable.HashMap[String, Properties] + import scala.collection.JavaConversions._ + prop.foreach { kv => + if (regex.findPrefixOf(kv._1) != None) { + val regex(prefix, suffix) = kv._1 + subProperties.getOrElseUpdate(prefix, new Properties).setProperty(suffix, kv._2) + } + } + subProperties + } + + def getInstance(inst: String): Properties = { + propertyCategories.get(inst) match { + case Some(s) => s + case None => propertyCategories.getOrElse(DEFAULT_PREFIX, new Properties) + } + } +} + diff --git a/core/src/main/scala/spark/metrics/MetricsSystem.scala b/core/src/main/scala/spark/metrics/MetricsSystem.scala new file mode 100644 index 0000000000..2f87577ff3 --- /dev/null +++ b/core/src/main/scala/spark/metrics/MetricsSystem.scala @@ -0,0 +1,129 @@ +package spark.metrics + +import com.codahale.metrics.{JmxReporter, MetricSet, MetricRegistry} + +import java.util.Properties +import java.util.concurrent.TimeUnit + +import scala.collection.mutable + +import spark.Logging +import spark.metrics.sink.Sink +import spark.metrics.source.Source + +/** + * Spark Metrics System, created by specific "instance", combined by source, + * sink, periodically poll source metrics data to sink destinations. + * + * "instance" specify "who" (the role) use metrics system. In spark there are several roles + * like master, worker, executor, client driver, these roles will create metrics system + * for monitoring. So instance represents these roles. Currently in Spark, several instances + * have already implemented: master, worker, executor, driver. + * + * "source" specify "where" (source) to collect metrics data. In metrics system, there exists + * two kinds of source: + * 1. Spark internal source, like MasterSource, WorkerSource, etc, which will collect + * Spark component's internal state, these sources are related to instance and will be + * added after specific metrics system is created. + * 2. Common source, like JvmSource, which will collect low level state, is configured by + * configuration and loaded through reflection. + * + * "sink" specify "where" (destination) to output metrics data to. Several sinks can be + * coexisted and flush metrics to all these sinks. + * + * Metrics configuration format is like below: + * [instance].[sink|source].[name].[options] = xxxx + * + * [instance] can be "master", "worker", "executor", "driver", which means only the specified + * instance has this property. + * wild card "*" can be used to replace instance name, which means all the instances will have + * this property. + * + * [sink|source] means this property belongs to source or sink. This field can only be source or sink. + * + * [name] specify the name of sink or source, it is custom defined. + * + * [options] is the specific property of this source or sink. + */ +private[spark] class MetricsSystem private (val instance: String) extends Logging { + initLogging() + + val confFile = System.getProperty("spark.metrics.conf") + val metricsConfig = new MetricsConfig(Option(confFile)) + + val sinks = new mutable.ArrayBuffer[Sink] + val sources = new mutable.ArrayBuffer[Source] + val registry = new MetricRegistry() + + metricsConfig.initialize() + registerSources() + registerSinks() + + def start() { + sinks.foreach(_.start) + } + + def stop() { + sinks.foreach(_.stop) + } + + def registerSource(source: Source) { + sources += source + try { + registry.register(source.sourceName, source.metricRegistry) + } catch { + case e: IllegalArgumentException => logInfo("Metrics already registered", e) + } + } + + def registerSources() { + val instConfig = metricsConfig.getInstance(instance) + val sourceConfigs = metricsConfig.subProperties(instConfig, MetricsSystem.SOURCE_REGEX) + + // Register all the sources related to instance + sourceConfigs.foreach { kv => + val classPath = kv._2.getProperty("class") + try { + val source = Class.forName(classPath).newInstance() + registerSource(source.asInstanceOf[Source]) + } catch { + case e: Exception => logError("Source class " + classPath + " cannot be instantialized", e) + } + } + } + + def registerSinks() { + val instConfig = metricsConfig.getInstance(instance) + val sinkConfigs = metricsConfig.subProperties(instConfig, MetricsSystem.SINK_REGEX) + + sinkConfigs.foreach { kv => + val classPath = kv._2.getProperty("class") + try { + val sink = Class.forName(classPath) + .getConstructor(classOf[Properties], classOf[MetricRegistry]) + .newInstance(kv._2, registry) + sinks += sink.asInstanceOf[Sink] + } catch { + case e: Exception => logError("Sink class " + classPath + " cannot be instantialized", e) + } + } + } +} + +private[spark] object MetricsSystem { + val SINK_REGEX = "^sink\\.(.+)\\.(.+)".r + val SOURCE_REGEX = "^source\\.(.+)\\.(.+)".r + + val MINIMAL_POLL_UNIT = TimeUnit.SECONDS + val MINIMAL_POLL_PERIOD = 1 + + def checkMinimalPollingPeriod(pollUnit: TimeUnit, pollPeriod: Int) { + val period = MINIMAL_POLL_UNIT.convert(pollPeriod, pollUnit) + if (period < MINIMAL_POLL_PERIOD) { + throw new IllegalArgumentException("Polling period " + pollPeriod + " " + pollUnit + + " below than minimal polling period ") + } + } + + def createMetricsSystem(instance: String): MetricsSystem = new MetricsSystem(instance) +} diff --git a/core/src/main/scala/spark/metrics/sink/ConsoleSink.scala b/core/src/main/scala/spark/metrics/sink/ConsoleSink.scala new file mode 100644 index 0000000000..eaaac5d153 --- /dev/null +++ b/core/src/main/scala/spark/metrics/sink/ConsoleSink.scala @@ -0,0 +1,42 @@ +package spark.metrics.sink + +import com.codahale.metrics.{ConsoleReporter, MetricRegistry} + +import java.util.Properties +import java.util.concurrent.TimeUnit + +import spark.metrics.MetricsSystem + +class ConsoleSink(val property: Properties, val registry: MetricRegistry) extends Sink { + val CONSOLE_DEFAULT_PERIOD = 10 + val CONSOLE_DEFAULT_UNIT = "SECONDS" + + val CONSOLE_KEY_PERIOD = "period" + val CONSOLE_KEY_UNIT = "unit" + + val pollPeriod = Option(property.getProperty(CONSOLE_KEY_PERIOD)) match { + case Some(s) => s.toInt + case None => CONSOLE_DEFAULT_PERIOD + } + + val pollUnit = Option(property.getProperty(CONSOLE_KEY_UNIT)) match { + case Some(s) => TimeUnit.valueOf(s.toUpperCase()) + case None => TimeUnit.valueOf(CONSOLE_DEFAULT_UNIT) + } + + MetricsSystem.checkMinimalPollingPeriod(pollUnit, pollPeriod) + + val reporter: ConsoleReporter = ConsoleReporter.forRegistry(registry) + .convertDurationsTo(TimeUnit.MILLISECONDS) + .convertRatesTo(TimeUnit.SECONDS) + .build() + + override def start() { + reporter.start(pollPeriod, pollUnit) + } + + override def stop() { + reporter.stop() + } +} + diff --git a/core/src/main/scala/spark/metrics/sink/CsvSink.scala b/core/src/main/scala/spark/metrics/sink/CsvSink.scala new file mode 100644 index 0000000000..aa5bff0d34 --- /dev/null +++ b/core/src/main/scala/spark/metrics/sink/CsvSink.scala @@ -0,0 +1,51 @@ +package spark.metrics.sink + +import com.codahale.metrics.{CsvReporter, MetricRegistry} + +import java.io.File +import java.util.{Locale, Properties} +import java.util.concurrent.TimeUnit + +import spark.metrics.MetricsSystem + +class CsvSink(val property: Properties, val registry: MetricRegistry) extends Sink { + val CSV_KEY_PERIOD = "period" + val CSV_KEY_UNIT = "unit" + val CSV_KEY_DIR = "directory" + + val CSV_DEFAULT_PERIOD = 10 + val CSV_DEFAULT_UNIT = "SECONDS" + val CSV_DEFAULT_DIR = "/tmp/" + + val pollPeriod = Option(property.getProperty(CSV_KEY_PERIOD)) match { + case Some(s) => s.toInt + case None => CSV_DEFAULT_PERIOD + } + + val pollUnit = Option(property.getProperty(CSV_KEY_UNIT)) match { + case Some(s) => TimeUnit.valueOf(s.toUpperCase()) + case None => TimeUnit.valueOf(CSV_DEFAULT_UNIT) + } + + MetricsSystem.checkMinimalPollingPeriod(pollUnit, pollPeriod) + + val pollDir = Option(property.getProperty(CSV_KEY_DIR)) match { + case Some(s) => s + case None => CSV_DEFAULT_DIR + } + + val reporter: CsvReporter = CsvReporter.forRegistry(registry) + .formatFor(Locale.US) + .convertDurationsTo(TimeUnit.MILLISECONDS) + .convertRatesTo(TimeUnit.SECONDS) + .build(new File(pollDir)) + + override def start() { + reporter.start(pollPeriod, pollUnit) + } + + override def stop() { + reporter.stop() + } +} + diff --git a/core/src/main/scala/spark/metrics/sink/JmxSink.scala b/core/src/main/scala/spark/metrics/sink/JmxSink.scala new file mode 100644 index 0000000000..6a40885b78 --- /dev/null +++ b/core/src/main/scala/spark/metrics/sink/JmxSink.scala @@ -0,0 +1,18 @@ +package spark.metrics.sink + +import com.codahale.metrics.{JmxReporter, MetricRegistry} + +import java.util.Properties + +class JmxSink(val property: Properties, val registry: MetricRegistry) extends Sink { + val reporter: JmxReporter = JmxReporter.forRegistry(registry).build() + + override def start() { + reporter.start() + } + + override def stop() { + reporter.stop() + } + +} diff --git a/core/src/main/scala/spark/metrics/sink/Sink.scala b/core/src/main/scala/spark/metrics/sink/Sink.scala new file mode 100644 index 0000000000..3ffdcbdaba --- /dev/null +++ b/core/src/main/scala/spark/metrics/sink/Sink.scala @@ -0,0 +1,6 @@ +package spark.metrics.sink + +trait Sink { + def start: Unit + def stop: Unit +}
\ No newline at end of file diff --git a/core/src/main/scala/spark/metrics/source/JvmSource.scala b/core/src/main/scala/spark/metrics/source/JvmSource.scala new file mode 100644 index 0000000000..79f505079c --- /dev/null +++ b/core/src/main/scala/spark/metrics/source/JvmSource.scala @@ -0,0 +1,15 @@ +package spark.metrics.source + +import com.codahale.metrics.MetricRegistry +import com.codahale.metrics.jvm.{GarbageCollectorMetricSet, MemoryUsageGaugeSet} + +class JvmSource extends Source { + val sourceName = "jvm" + val metricRegistry = new MetricRegistry() + + val gcMetricSet = new GarbageCollectorMetricSet + val memGaugeSet = new MemoryUsageGaugeSet + + metricRegistry.registerAll(gcMetricSet) + metricRegistry.registerAll(memGaugeSet) +} diff --git a/core/src/main/scala/spark/metrics/source/Source.scala b/core/src/main/scala/spark/metrics/source/Source.scala new file mode 100644 index 0000000000..5607e2c40a --- /dev/null +++ b/core/src/main/scala/spark/metrics/source/Source.scala @@ -0,0 +1,8 @@ +package spark.metrics.source + +import com.codahale.metrics.MetricRegistry + +trait Source { + def sourceName: String + def metricRegistry: MetricRegistry +} diff --git a/core/src/main/scala/spark/network/BufferMessage.scala b/core/src/main/scala/spark/network/BufferMessage.scala new file mode 100644 index 0000000000..e566aeac13 --- /dev/null +++ b/core/src/main/scala/spark/network/BufferMessage.scala @@ -0,0 +1,111 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network + +import java.nio.ByteBuffer + +import scala.collection.mutable.ArrayBuffer + +import spark.storage.BlockManager + + +private[spark] +class BufferMessage(id_ : Int, val buffers: ArrayBuffer[ByteBuffer], var ackId: Int) + extends Message(Message.BUFFER_MESSAGE, id_) { + + val initialSize = currentSize() + var gotChunkForSendingOnce = false + + def size = initialSize + + def currentSize() = { + if (buffers == null || buffers.isEmpty) { + 0 + } else { + buffers.map(_.remaining).reduceLeft(_ + _) + } + } + + def getChunkForSending(maxChunkSize: Int): Option[MessageChunk] = { + if (maxChunkSize <= 0) { + throw new Exception("Max chunk size is " + maxChunkSize) + } + + if (size == 0 && gotChunkForSendingOnce == false) { + val newChunk = new MessageChunk( + new MessageChunkHeader(typ, id, 0, 0, ackId, senderAddress), null) + gotChunkForSendingOnce = true + return Some(newChunk) + } + + while(!buffers.isEmpty) { + val buffer = buffers(0) + if (buffer.remaining == 0) { + BlockManager.dispose(buffer) + buffers -= buffer + } else { + val newBuffer = if (buffer.remaining <= maxChunkSize) { + buffer.duplicate() + } else { + buffer.slice().limit(maxChunkSize).asInstanceOf[ByteBuffer] + } + buffer.position(buffer.position + newBuffer.remaining) + val newChunk = new MessageChunk(new MessageChunkHeader( + typ, id, size, newBuffer.remaining, ackId, senderAddress), newBuffer) + gotChunkForSendingOnce = true + return Some(newChunk) + } + } + None + } + + def getChunkForReceiving(chunkSize: Int): Option[MessageChunk] = { + // STRONG ASSUMPTION: BufferMessage created when receiving data has ONLY ONE data buffer + if (buffers.size > 1) { + throw new Exception("Attempting to get chunk from message with multiple data buffers") + } + val buffer = buffers(0) + if (buffer.remaining > 0) { + if (buffer.remaining < chunkSize) { + throw new Exception("Not enough space in data buffer for receiving chunk") + } + val newBuffer = buffer.slice().limit(chunkSize).asInstanceOf[ByteBuffer] + buffer.position(buffer.position + newBuffer.remaining) + val newChunk = new MessageChunk(new MessageChunkHeader( + typ, id, size, newBuffer.remaining, ackId, senderAddress), newBuffer) + return Some(newChunk) + } + None + } + + def flip() { + buffers.foreach(_.flip) + } + + def hasAckId() = (ackId != 0) + + def isCompletelyReceived() = !buffers(0).hasRemaining + + override def toString = { + if (hasAckId) { + "BufferAckMessage(aid = " + ackId + ", id = " + id + ", size = " + size + ")" + } else { + "BufferMessage(id = " + id + ", size = " + size + ")" + } + } +} diff --git a/core/src/main/scala/spark/network/Connection.scala b/core/src/main/scala/spark/network/Connection.scala index 00a0433a44..b66c00b58c 100644 --- a/core/src/main/scala/spark/network/Connection.scala +++ b/core/src/main/scala/spark/network/Connection.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.network import spark._ @@ -13,12 +30,13 @@ import java.net._ private[spark] abstract class Connection(val channel: SocketChannel, val selector: Selector, - val socketRemoteConnectionManagerId: ConnectionManagerId) extends Logging { + val socketRemoteConnectionManagerId: ConnectionManagerId) + extends Logging { + def this(channel_ : SocketChannel, selector_ : Selector) = { this(channel_, selector_, - ConnectionManagerId.fromSocketAddress( - channel_.socket.getRemoteSocketAddress().asInstanceOf[InetSocketAddress] - )) + ConnectionManagerId.fromSocketAddress( + channel_.socket.getRemoteSocketAddress().asInstanceOf[InetSocketAddress])) } channel.configureBlocking(false) @@ -32,17 +50,19 @@ abstract class Connection(val channel: SocketChannel, val selector: Selector, var onKeyInterestChangeCallback: (Connection, Int) => Unit = null val remoteAddress = getRemoteAddress() - + // Read channels typically do not register for write and write does not for read // Now, we do have write registering for read too (temporarily), but this is to detect // channel close NOT to actually read/consume data on it ! // How does this work if/when we move to SSL ? - + // What is the interest to register with selector for when we want this connection to be selected def registerInterest() - // What is the interest to register with selector for when we want this connection to be de-selected - // Traditionally, 0 - but in our case, for example, for close-detection on SendingConnection hack, it will be - // SelectionKey.OP_READ (until we fix it properly) + + // What is the interest to register with selector for when we want this connection to + // be de-selected + // Traditionally, 0 - but in our case, for example, for close-detection on SendingConnection hack, + // it will be SelectionKey.OP_READ (until we fix it properly) def unregisterInterest() // On receiving a read event, should we change the interest for this channel or not ? @@ -64,12 +84,14 @@ abstract class Connection(val channel: SocketChannel, val selector: Selector, // Returns whether we have to register for further reads or not. def read(): Boolean = { - throw new UnsupportedOperationException("Cannot read on connection of type " + this.getClass.toString) + throw new UnsupportedOperationException( + "Cannot read on connection of type " + this.getClass.toString) } // Returns whether we have to register for further writes or not. def write(): Boolean = { - throw new UnsupportedOperationException("Cannot write on connection of type " + this.getClass.toString) + throw new UnsupportedOperationException( + "Cannot write on connection of type " + this.getClass.toString) } def close() { @@ -81,11 +103,17 @@ abstract class Connection(val channel: SocketChannel, val selector: Selector, callOnCloseCallback() } - def onClose(callback: Connection => Unit) {onCloseCallback = callback} + def onClose(callback: Connection => Unit) { + onCloseCallback = callback + } - def onException(callback: (Connection, Exception) => Unit) {onExceptionCallback = callback} + def onException(callback: (Connection, Exception) => Unit) { + onExceptionCallback = callback + } - def onKeyInterestChange(callback: (Connection, Int) => Unit) {onKeyInterestChangeCallback = callback} + def onKeyInterestChange(callback: (Connection, Int) => Unit) { + onKeyInterestChangeCallback = callback + } def callOnExceptionCallback(e: Exception) { if (onExceptionCallback != null) { @@ -95,7 +123,7 @@ abstract class Connection(val channel: SocketChannel, val selector: Selector, " and OnExceptionCallback not registered", e) } } - + def callOnCloseCallback() { if (onCloseCallback != null) { onCloseCallback(this) @@ -132,24 +160,25 @@ abstract class Connection(val channel: SocketChannel, val selector: Selector, print(" (" + position + ", " + length + ")") buffer.position(curPosition) } - } -private[spark] class SendingConnection(val address: InetSocketAddress, selector_ : Selector, - remoteId_ : ConnectionManagerId) -extends Connection(SocketChannel.open, selector_, remoteId_) { +private[spark] +class SendingConnection(val address: InetSocketAddress, selector_ : Selector, + remoteId_ : ConnectionManagerId) + extends Connection(SocketChannel.open, selector_, remoteId_) { class Outbox(fair: Int = 0) { val messages = new Queue[Message]() - val defaultChunkSize = 65536 //32768 //16384 + val defaultChunkSize = 65536 //32768 //16384 var nextMessageToBeUsed = 0 def addMessage(message: Message) { - messages.synchronized{ + messages.synchronized{ /*messages += message*/ messages.enqueue(message) - logDebug("Added [" + message + "] to outbox for sending to [" + getRemoteConnectionManagerId() + "]") + logDebug("Added [" + message + "] to outbox for sending to " + + "[" + getRemoteConnectionManagerId() + "]") } } @@ -174,7 +203,7 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { message.started = true message.startTime = System.currentTimeMillis } - return chunk + return chunk } else { /*logInfo("Finished sending [" + message + "] to [" + getRemoteConnectionManagerId() + "]")*/ message.finishTime = System.currentTimeMillis @@ -185,7 +214,7 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { } None } - + private def getChunkRR(): Option[MessageChunk] = { messages.synchronized { while (!messages.isEmpty) { @@ -197,12 +226,14 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { messages.enqueue(message) nextMessageToBeUsed = nextMessageToBeUsed + 1 if (!message.started) { - logDebug("Starting to send [" + message + "] to [" + getRemoteConnectionManagerId() + "]") + logDebug( + "Starting to send [" + message + "] to [" + getRemoteConnectionManagerId() + "]") message.started = true message.startTime = System.currentTimeMillis } - logTrace("Sending chunk from [" + message+ "] to [" + getRemoteConnectionManagerId() + "]") - return chunk + logTrace( + "Sending chunk from [" + message+ "] to [" + getRemoteConnectionManagerId() + "]") + return chunk } else { message.finishTime = System.currentTimeMillis logDebug("Finished sending [" + message + "] to [" + getRemoteConnectionManagerId() + @@ -213,7 +244,7 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { None } } - + private val outbox = new Outbox(1) val currentBuffers = new ArrayBuffer[ByteBuffer]() @@ -228,11 +259,11 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { // it does - so let us keep it for now. changeConnectionKeyInterest(SelectionKey.OP_WRITE | DEFAULT_INTEREST) } - + override def unregisterInterest() { changeConnectionKeyInterest(DEFAULT_INTEREST) } - + def send(message: Message) { outbox.synchronized { outbox.addMessage(message) @@ -262,12 +293,14 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { // selection - though need not necessarily always complete successfully. val connected = channel.finishConnect if (!force && !connected) { - logInfo("finish connect failed [" + address + "], " + outbox.messages.size + " messages pending") + logInfo( + "finish connect failed [" + address + "], " + outbox.messages.size + " messages pending") return false } // Fallback to previous behavior - assume finishConnect completed - // This will happen only when finishConnect failed for some repeated number of times (10 or so) + // This will happen only when finishConnect failed for some repeated number of times + // (10 or so) // Is highly unlikely unless there was an unclean close of socket, etc registerInterest() logInfo("Connected to [" + address + "], " + outbox.messages.size + " messages pending") @@ -283,13 +316,13 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { } override def write(): Boolean = { - try{ - while(true) { + try { + while (true) { if (currentBuffers.size == 0) { outbox.synchronized { outbox.getChunk() match { case Some(chunk) => { - currentBuffers ++= chunk.buffers + currentBuffers ++= chunk.buffers } case None => { // changeConnectionKeyInterest(0) @@ -299,7 +332,7 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { } } } - + if (currentBuffers.size > 0) { val buffer = currentBuffers(0) val remainingBytes = buffer.remaining @@ -314,7 +347,7 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { } } } catch { - case e: Exception => { + case e: Exception => { logWarning("Error writing in connection to " + getRemoteConnectionManagerId(), e) callOnExceptionCallback(e) close() @@ -336,7 +369,8 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { if (length == -1) { // EOF close() } else if (length > 0) { - logWarning("Unexpected data read from SendingConnection to " + getRemoteConnectionManagerId()) + logWarning( + "Unexpected data read from SendingConnection to " + getRemoteConnectionManagerId()) } } catch { case e: Exception => @@ -355,30 +389,32 @@ extends Connection(SocketChannel.open, selector_, remoteId_) { // Must be created within selector loop - else deadlock -private[spark] class ReceivingConnection(channel_ : SocketChannel, selector_ : Selector) -extends Connection(channel_, selector_) { - +private[spark] class ReceivingConnection(channel_ : SocketChannel, selector_ : Selector) + extends Connection(channel_, selector_) { + class Inbox() { val messages = new HashMap[Int, BufferMessage]() - + def getChunk(header: MessageChunkHeader): Option[MessageChunk] = { - + def createNewMessage: BufferMessage = { val newMessage = Message.create(header).asInstanceOf[BufferMessage] newMessage.started = true newMessage.startTime = System.currentTimeMillis - logDebug("Starting to receive [" + newMessage + "] from [" + getRemoteConnectionManagerId() + "]") + logDebug( + "Starting to receive [" + newMessage + "] from [" + getRemoteConnectionManagerId() + "]") messages += ((newMessage.id, newMessage)) newMessage } - + val message = messages.getOrElseUpdate(header.id, createNewMessage) - logTrace("Receiving chunk of [" + message + "] from [" + getRemoteConnectionManagerId() + "]") + logTrace( + "Receiving chunk of [" + message + "] from [" + getRemoteConnectionManagerId() + "]") message.getChunkForReceiving(header.chunkSize) } - + def getMessageForChunk(chunk: MessageChunk): Option[BufferMessage] = { - messages.get(chunk.header.id) + messages.get(chunk.header.id) } def removeMessage(message: Message) { @@ -387,12 +423,14 @@ extends Connection(channel_, selector_) { } @volatile private var inferredRemoteManagerId: ConnectionManagerId = null + override def getRemoteConnectionManagerId(): ConnectionManagerId = { val currId = inferredRemoteManagerId if (currId != null) currId else super.getRemoteConnectionManagerId() } - // The reciever's remote address is the local socket on remote side : which is NOT the connection manager id of the receiver. + // The reciever's remote address is the local socket on remote side : which is NOT + // the connection manager id of the receiver. // We infer that from the messages we receive on the receiver socket. private def processConnectionManagerId(header: MessageChunkHeader) { val currId = inferredRemoteManagerId @@ -428,7 +466,8 @@ extends Connection(channel_, selector_) { } headerBuffer.flip if (headerBuffer.remaining != MessageChunkHeader.HEADER_SIZE) { - throw new Exception("Unexpected number of bytes (" + headerBuffer.remaining + ") in the header") + throw new Exception( + "Unexpected number of bytes (" + headerBuffer.remaining + ") in the header") } val header = MessageChunkHeader.create(headerBuffer) headerBuffer.clear() @@ -451,9 +490,9 @@ extends Connection(channel_, selector_) { case _ => throw new Exception("Message of unknown type received") } } - + if (currentChunk == null) throw new Exception("No message chunk to receive data") - + val bytesRead = channel.read(currentChunk.buffer) if (bytesRead == 0) { // re-register for read event ... @@ -464,14 +503,15 @@ extends Connection(channel_, selector_) { } /*logDebug("Read " + bytesRead + " bytes for the buffer")*/ - + if (currentChunk.buffer.remaining == 0) { /*println("Filled buffer at " + System.currentTimeMillis)*/ val bufferMessage = inbox.getMessageForChunk(currentChunk).get if (bufferMessage.isCompletelyReceived) { bufferMessage.flip bufferMessage.finishTime = System.currentTimeMillis - logDebug("Finished receiving [" + bufferMessage + "] from [" + getRemoteConnectionManagerId() + "] in " + bufferMessage.timeTaken) + logDebug("Finished receiving [" + bufferMessage + "] from " + + "[" + getRemoteConnectionManagerId() + "] in " + bufferMessage.timeTaken) if (onReceiveCallback != null) { onReceiveCallback(this, bufferMessage) } @@ -481,7 +521,7 @@ extends Connection(channel_, selector_) { } } } catch { - case e: Exception => { + case e: Exception => { logWarning("Error reading from connection to " + getRemoteConnectionManagerId(), e) callOnExceptionCallback(e) close() @@ -491,7 +531,7 @@ extends Connection(channel_, selector_) { // should not happen - to keep scala compiler happy return true } - + def onReceive(callback: (Connection, Message) => Unit) {onReceiveCallback = callback} override def changeInterestForRead(): Boolean = true @@ -505,7 +545,7 @@ extends Connection(channel_, selector_) { // it does - so let us keep it for now. changeConnectionKeyInterest(SelectionKey.OP_READ) } - + override def unregisterInterest() { changeConnectionKeyInterest(0) } diff --git a/core/src/main/scala/spark/network/ConnectionManager.scala b/core/src/main/scala/spark/network/ConnectionManager.scala index 0c6bdb1559..6c4e7dc03e 100644 --- a/core/src/main/scala/spark/network/ConnectionManager.scala +++ b/core/src/main/scala/spark/network/ConnectionManager.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.network import spark._ @@ -18,20 +35,7 @@ import akka.dispatch.{Await, Promise, ExecutionContext, Future} import akka.util.Duration import akka.util.duration._ -private[spark] case class ConnectionManagerId(host: String, port: Int) { - // DEBUG code - Utils.checkHost(host) - assert (port > 0) - - def toSocketAddress() = new InetSocketAddress(host, port) -} -private[spark] object ConnectionManagerId { - def fromSocketAddress(socketAddress: InetSocketAddress): ConnectionManagerId = { - new ConnectionManagerId(socketAddress.getHostName(), socketAddress.getPort()) - } -} - private[spark] class ConnectionManager(port: Int) extends Logging { class MessageStatus( @@ -45,7 +49,7 @@ private[spark] class ConnectionManager(port: Int) extends Logging { def markDone() { completionHandler(this) } } - + private val selector = SelectorProvider.provider.openSelector() private val handleMessageExecutor = new ThreadPoolExecutor( @@ -80,7 +84,7 @@ private[spark] class ConnectionManager(port: Int) extends Logging { serverChannel.configureBlocking(false) serverChannel.socket.setReuseAddress(true) - serverChannel.socket.setReceiveBufferSize(256 * 1024) + serverChannel.socket.setReceiveBufferSize(256 * 1024) serverChannel.socket.bind(new InetSocketAddress(port)) serverChannel.register(selector, SelectionKey.OP_ACCEPT) @@ -188,6 +192,38 @@ private[spark] class ConnectionManager(port: Int) extends Logging { } ) } + // MUST be called within selector loop - else deadlock. + private def triggerForceCloseByException(key: SelectionKey, e: Exception) { + try { + key.interestOps(0) + } catch { + // ignore exceptions + case e: Exception => logDebug("Ignoring exception", e) + } + + val conn = connectionsByKey.getOrElse(key, null) + if (conn == null) return + + // Pushing to connect threadpool + handleConnectExecutor.execute(new Runnable { + override def run() { + try { + conn.callOnExceptionCallback(e) + } catch { + // ignore exceptions + case e: Exception => logDebug("Ignoring exception", e) + } + try { + conn.close() + } catch { + // ignore exceptions + case e: Exception => logDebug("Ignoring exception", e) + } + } + }) + } + + def run() { try { while(!selectorThread.isInterrupted) { @@ -200,29 +236,76 @@ private[spark] class ConnectionManager(port: Int) extends Logging { while(!keyInterestChangeRequests.isEmpty) { val (key, ops) = keyInterestChangeRequests.dequeue - val connection = connectionsByKey.getOrElse(key, null) - if (connection != null) { - val lastOps = key.interestOps() - key.interestOps(ops) - - // hot loop - prevent materialization of string if trace not enabled. - if (isTraceEnabled()) { - def intToOpStr(op: Int): String = { - val opStrs = ArrayBuffer[String]() - if ((op & SelectionKey.OP_READ) != 0) opStrs += "READ" - if ((op & SelectionKey.OP_WRITE) != 0) opStrs += "WRITE" - if ((op & SelectionKey.OP_CONNECT) != 0) opStrs += "CONNECT" - if ((op & SelectionKey.OP_ACCEPT) != 0) opStrs += "ACCEPT" - if (opStrs.size > 0) opStrs.reduceLeft(_ + " | " + _) else " " - } - logTrace("Changed key for connection to [" + connection.getRemoteConnectionManagerId() + - "] changed from [" + intToOpStr(lastOps) + "] to [" + intToOpStr(ops) + "]") + try { + if (key.isValid) { + val connection = connectionsByKey.getOrElse(key, null) + if (connection != null) { + val lastOps = key.interestOps() + key.interestOps(ops) + + // hot loop - prevent materialization of string if trace not enabled. + if (isTraceEnabled()) { + def intToOpStr(op: Int): String = { + val opStrs = ArrayBuffer[String]() + if ((op & SelectionKey.OP_READ) != 0) opStrs += "READ" + if ((op & SelectionKey.OP_WRITE) != 0) opStrs += "WRITE" + if ((op & SelectionKey.OP_CONNECT) != 0) opStrs += "CONNECT" + if ((op & SelectionKey.OP_ACCEPT) != 0) opStrs += "ACCEPT" + if (opStrs.size > 0) opStrs.reduceLeft(_ + " | " + _) else " " + } + + logTrace("Changed key for connection to [" + connection.getRemoteConnectionManagerId() + + "] changed from [" + intToOpStr(lastOps) + "] to [" + intToOpStr(ops) + "]") + } + } + } else { + logInfo("Key not valid ? " + key) + throw new CancelledKeyException() + } + } catch { + case e: CancelledKeyException => { + logInfo("key already cancelled ? " + key, e) + triggerForceCloseByException(key, e) + } + case e: Exception => { + logError("Exception processing key " + key, e) + triggerForceCloseByException(key, e) } } } - val selectedKeysCount = selector.select() + val selectedKeysCount = + try { + selector.select() + } catch { + // Explicitly only dealing with CancelledKeyException here since other exceptions should be dealt with differently. + case e: CancelledKeyException => { + // Some keys within the selectors list are invalid/closed. clear them. + val allKeys = selector.keys().iterator() + + while (allKeys.hasNext()) { + val key = allKeys.next() + try { + if (! key.isValid) { + logInfo("Key not valid ? " + key) + throw new CancelledKeyException() + } + } catch { + case e: CancelledKeyException => { + logInfo("key already cancelled ? " + key, e) + triggerForceCloseByException(key, e) + } + case e: Exception => { + logError("Exception processing key " + key, e) + triggerForceCloseByException(key, e) + } + } + } + } + 0 + } + if (selectedKeysCount == 0) { logDebug("Selector selected " + selectedKeysCount + " of " + selector.keys.size + " keys") } @@ -230,23 +313,40 @@ private[spark] class ConnectionManager(port: Int) extends Logging { logInfo("Selector thread was interrupted!") return } - - val selectedKeys = selector.selectedKeys().iterator() - while (selectedKeys.hasNext()) { - val key = selectedKeys.next - selectedKeys.remove() - if (key.isValid) { - if (key.isAcceptable) { - acceptConnection(key) - } else - if (key.isConnectable) { - triggerConnect(key) - } else - if (key.isReadable) { - triggerRead(key) - } else - if (key.isWritable) { - triggerWrite(key) + + if (0 != selectedKeysCount) { + val selectedKeys = selector.selectedKeys().iterator() + while (selectedKeys.hasNext()) { + val key = selectedKeys.next + selectedKeys.remove() + try { + if (key.isValid) { + if (key.isAcceptable) { + acceptConnection(key) + } else + if (key.isConnectable) { + triggerConnect(key) + } else + if (key.isReadable) { + triggerRead(key) + } else + if (key.isWritable) { + triggerWrite(key) + } + } else { + logInfo("Key not valid ? " + key) + throw new CancelledKeyException() + } + } catch { + // weird, but we saw this happening - even though key.isValid was true, key.isAcceptable would throw CancelledKeyException. + case e: CancelledKeyException => { + logInfo("key already cancelled ? " + key, e) + triggerForceCloseByException(key, e) + } + case e: Exception => { + logError("Exception processing key " + key, e) + triggerForceCloseByException(key, e) + } } } } @@ -255,7 +355,7 @@ private[spark] class ConnectionManager(port: Int) extends Logging { case e: Exception => logError("Error in select loop", e) } } - + def acceptConnection(key: SelectionKey) { val serverChannel = key.channel.asInstanceOf[ServerSocketChannel] @@ -367,7 +467,7 @@ private[spark] class ConnectionManager(port: Int) extends Logging { def receiveMessage(connection: Connection, message: Message) { val connectionManagerId = ConnectionManagerId.fromSocketAddress(message.senderAddress) - logDebug("Received [" + message + "] from [" + connectionManagerId + "]") + logDebug("Received [" + message + "] from [" + connectionManagerId + "]") val runnable = new Runnable() { val creationTime = System.currentTimeMillis def run() { @@ -387,11 +487,11 @@ private[spark] class ConnectionManager(port: Int) extends Logging { if (bufferMessage.hasAckId) { val sentMessageStatus = messageStatuses.synchronized { messageStatuses.get(bufferMessage.ackId) match { - case Some(status) => { - messageStatuses -= bufferMessage.ackId + case Some(status) => { + messageStatuses -= bufferMessage.ackId status } - case None => { + case None => { throw new Exception("Could not find reference for received ack message " + message.id) null } @@ -411,7 +511,7 @@ private[spark] class ConnectionManager(port: Int) extends Logging { logDebug("Not calling back as callback is null") None } - + if (ackMessage.isDefined) { if (!ackMessage.get.isInstanceOf[BufferMessage]) { logDebug("Response to " + bufferMessage + " is not a buffer message, it is of type " + ackMessage.get.getClass()) @@ -421,7 +521,7 @@ private[spark] class ConnectionManager(port: Int) extends Logging { } } - sendMessage(connectionManagerId, ackMessage.getOrElse { + sendMessage(connectionManagerId, ackMessage.getOrElse { Message.createBufferMessage(bufferMessage.id) }) } @@ -492,17 +592,17 @@ private[spark] object ConnectionManager { def main(args: Array[String]) { val manager = new ConnectionManager(9999) - manager.onReceiveMessage((msg: Message, id: ConnectionManagerId) => { + manager.onReceiveMessage((msg: Message, id: ConnectionManagerId) => { println("Received [" + msg + "] from [" + id + "]") None }) - + /*testSequentialSending(manager)*/ /*System.gc()*/ /*testParallelSending(manager)*/ /*System.gc()*/ - + /*testParallelDecreasingSending(manager)*/ /*System.gc()*/ @@ -514,9 +614,9 @@ private[spark] object ConnectionManager { println("--------------------------") println("Sequential Sending") println("--------------------------") - val size = 10 * 1024 * 1024 + val size = 10 * 1024 * 1024 val count = 10 - + val buffer = ByteBuffer.allocate(size).put(Array.tabulate[Byte](size)(x => x.toByte)) buffer.flip @@ -532,7 +632,7 @@ private[spark] object ConnectionManager { println("--------------------------") println("Parallel Sending") println("--------------------------") - val size = 10 * 1024 * 1024 + val size = 10 * 1024 * 1024 val count = 10 val buffer = ByteBuffer.allocate(size).put(Array.tabulate[Byte](size)(x => x.toByte)) @@ -547,12 +647,12 @@ private[spark] object ConnectionManager { if (!g.isDefined) println("Failed") }) val finishTime = System.currentTimeMillis - + val mb = size * count / 1024.0 / 1024.0 val ms = finishTime - startTime val tput = mb * 1000.0 / ms println("--------------------------") - println("Started at " + startTime + ", finished at " + finishTime) + println("Started at " + startTime + ", finished at " + finishTime) println("Sent " + count + " messages of size " + size + " in " + ms + " ms (" + tput + " MB/s)") println("--------------------------") println() @@ -562,7 +662,7 @@ private[spark] object ConnectionManager { println("--------------------------") println("Parallel Decreasing Sending") println("--------------------------") - val size = 10 * 1024 * 1024 + val size = 10 * 1024 * 1024 val count = 10 val buffers = Array.tabulate(count)(i => ByteBuffer.allocate(size * (i + 1)).put(Array.tabulate[Byte](size * (i + 1))(x => x.toByte))) buffers.foreach(_.flip) @@ -577,7 +677,7 @@ private[spark] object ConnectionManager { if (!g.isDefined) println("Failed") }) val finishTime = System.currentTimeMillis - + val ms = finishTime - startTime val tput = mb * 1000.0 / ms println("--------------------------") @@ -591,7 +691,7 @@ private[spark] object ConnectionManager { println("--------------------------") println("Continuous Sending") println("--------------------------") - val size = 10 * 1024 * 1024 + val size = 10 * 1024 * 1024 val count = 10 val buffer = ByteBuffer.allocate(size).put(Array.tabulate[Byte](size)(x => x.toByte)) diff --git a/core/src/main/scala/spark/network/ConnectionManagerId.scala b/core/src/main/scala/spark/network/ConnectionManagerId.scala new file mode 100644 index 0000000000..9d5c518293 --- /dev/null +++ b/core/src/main/scala/spark/network/ConnectionManagerId.scala @@ -0,0 +1,38 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network + +import java.net.InetSocketAddress + +import spark.Utils + + +private[spark] case class ConnectionManagerId(host: String, port: Int) { + // DEBUG code + Utils.checkHost(host) + assert (port > 0) + + def toSocketAddress() = new InetSocketAddress(host, port) +} + + +private[spark] object ConnectionManagerId { + def fromSocketAddress(socketAddress: InetSocketAddress): ConnectionManagerId = { + new ConnectionManagerId(socketAddress.getHostName(), socketAddress.getPort()) + } +} diff --git a/core/src/main/scala/spark/network/ConnectionManagerTest.scala b/core/src/main/scala/spark/network/ConnectionManagerTest.scala index 533e4610f3..9e3827aaf5 100644 --- a/core/src/main/scala/spark/network/ConnectionManagerTest.scala +++ b/core/src/main/scala/spark/network/ConnectionManagerTest.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.network import spark._ diff --git a/core/src/main/scala/spark/network/Message.scala b/core/src/main/scala/spark/network/Message.scala index 34fac9e776..a25457ea35 100644 --- a/core/src/main/scala/spark/network/Message.scala +++ b/core/src/main/scala/spark/network/Message.scala @@ -1,56 +1,27 @@ -package spark.network - -import spark._ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ -import scala.collection.mutable.ArrayBuffer +package spark.network import java.nio.ByteBuffer -import java.net.InetAddress import java.net.InetSocketAddress -import storage.BlockManager - -private[spark] class MessageChunkHeader( - val typ: Long, - val id: Int, - val totalSize: Int, - val chunkSize: Int, - val other: Int, - val address: InetSocketAddress) { - lazy val buffer = { - // No need to change this, at 'use' time, we do a reverse lookup of the hostname. Refer to network.Connection - val ip = address.getAddress.getAddress() - val port = address.getPort() - ByteBuffer. - allocate(MessageChunkHeader.HEADER_SIZE). - putLong(typ). - putInt(id). - putInt(totalSize). - putInt(chunkSize). - putInt(other). - putInt(ip.size). - put(ip). - putInt(port). - position(MessageChunkHeader.HEADER_SIZE). - flip.asInstanceOf[ByteBuffer] - } - - override def toString = "" + this.getClass.getSimpleName + ":" + id + " of type " + typ + - " and sizes " + totalSize + " / " + chunkSize + " bytes" -} -private[spark] class MessageChunk(val header: MessageChunkHeader, val buffer: ByteBuffer) { - val size = if (buffer == null) 0 else buffer.remaining - lazy val buffers = { - val ab = new ArrayBuffer[ByteBuffer]() - ab += header.buffer - if (buffer != null) { - ab += buffer - } - ab - } +import scala.collection.mutable.ArrayBuffer - override def toString = "" + this.getClass.getSimpleName + " (id = " + header.id + ", size = " + size + ")" -} private[spark] abstract class Message(val typ: Long, val id: Int) { var senderAddress: InetSocketAddress = null @@ -59,120 +30,16 @@ private[spark] abstract class Message(val typ: Long, val id: Int) { var finishTime = -1L def size: Int - + def getChunkForSending(maxChunkSize: Int): Option[MessageChunk] - + def getChunkForReceiving(chunkSize: Int): Option[MessageChunk] - + def timeTaken(): String = (finishTime - startTime).toString + " ms" override def toString = this.getClass.getSimpleName + "(id = " + id + ", size = " + size + ")" } -private[spark] class BufferMessage(id_ : Int, val buffers: ArrayBuffer[ByteBuffer], var ackId: Int) -extends Message(Message.BUFFER_MESSAGE, id_) { - - val initialSize = currentSize() - var gotChunkForSendingOnce = false - - def size = initialSize - - def currentSize() = { - if (buffers == null || buffers.isEmpty) { - 0 - } else { - buffers.map(_.remaining).reduceLeft(_ + _) - } - } - - def getChunkForSending(maxChunkSize: Int): Option[MessageChunk] = { - if (maxChunkSize <= 0) { - throw new Exception("Max chunk size is " + maxChunkSize) - } - - if (size == 0 && gotChunkForSendingOnce == false) { - val newChunk = new MessageChunk(new MessageChunkHeader(typ, id, 0, 0, ackId, senderAddress), null) - gotChunkForSendingOnce = true - return Some(newChunk) - } - - while(!buffers.isEmpty) { - val buffer = buffers(0) - if (buffer.remaining == 0) { - BlockManager.dispose(buffer) - buffers -= buffer - } else { - val newBuffer = if (buffer.remaining <= maxChunkSize) { - buffer.duplicate() - } else { - buffer.slice().limit(maxChunkSize).asInstanceOf[ByteBuffer] - } - buffer.position(buffer.position + newBuffer.remaining) - val newChunk = new MessageChunk(new MessageChunkHeader( - typ, id, size, newBuffer.remaining, ackId, senderAddress), newBuffer) - gotChunkForSendingOnce = true - return Some(newChunk) - } - } - None - } - - def getChunkForReceiving(chunkSize: Int): Option[MessageChunk] = { - // STRONG ASSUMPTION: BufferMessage created when receiving data has ONLY ONE data buffer - if (buffers.size > 1) { - throw new Exception("Attempting to get chunk from message with multiple data buffers") - } - val buffer = buffers(0) - if (buffer.remaining > 0) { - if (buffer.remaining < chunkSize) { - throw new Exception("Not enough space in data buffer for receiving chunk") - } - val newBuffer = buffer.slice().limit(chunkSize).asInstanceOf[ByteBuffer] - buffer.position(buffer.position + newBuffer.remaining) - val newChunk = new MessageChunk(new MessageChunkHeader( - typ, id, size, newBuffer.remaining, ackId, senderAddress), newBuffer) - return Some(newChunk) - } - None - } - - def flip() { - buffers.foreach(_.flip) - } - - def hasAckId() = (ackId != 0) - - def isCompletelyReceived() = !buffers(0).hasRemaining - - override def toString = { - if (hasAckId) { - "BufferAckMessage(aid = " + ackId + ", id = " + id + ", size = " + size + ")" - } else { - "BufferMessage(id = " + id + ", size = " + size + ")" - } - } -} - -private[spark] object MessageChunkHeader { - val HEADER_SIZE = 40 - - def create(buffer: ByteBuffer): MessageChunkHeader = { - if (buffer.remaining != HEADER_SIZE) { - throw new IllegalArgumentException("Cannot convert buffer data to Message") - } - val typ = buffer.getLong() - val id = buffer.getInt() - val totalSize = buffer.getInt() - val chunkSize = buffer.getInt() - val other = buffer.getInt() - val ipSize = buffer.getInt() - val ipBytes = new Array[Byte](ipSize) - buffer.get(ipBytes) - val ip = InetAddress.getByAddress(ipBytes) - val port = buffer.getInt() - new MessageChunkHeader(typ, id, totalSize, chunkSize, other, new InetSocketAddress(ip, port)) - } -} private[spark] object Message { val BUFFER_MESSAGE = 1111111111L @@ -181,14 +48,16 @@ private[spark] object Message { def getNewId() = synchronized { lastId += 1 - if (lastId == 0) lastId += 1 + if (lastId == 0) { + lastId += 1 + } lastId } def createBufferMessage(dataBuffers: Seq[ByteBuffer], ackId: Int): BufferMessage = { if (dataBuffers == null) { return new BufferMessage(getNewId(), new ArrayBuffer[ByteBuffer], ackId) - } + } if (dataBuffers.exists(_ == null)) { throw new Exception("Attempting to create buffer message with null buffer") } @@ -197,7 +66,7 @@ private[spark] object Message { def createBufferMessage(dataBuffers: Seq[ByteBuffer]): BufferMessage = createBufferMessage(dataBuffers, 0) - + def createBufferMessage(dataBuffer: ByteBuffer, ackId: Int): BufferMessage = { if (dataBuffer == null) { return createBufferMessage(Array(ByteBuffer.allocate(0)), ackId) @@ -205,15 +74,18 @@ private[spark] object Message { return createBufferMessage(Array(dataBuffer), ackId) } } - - def createBufferMessage(dataBuffer: ByteBuffer): BufferMessage = + + def createBufferMessage(dataBuffer: ByteBuffer): BufferMessage = createBufferMessage(dataBuffer, 0) - - def createBufferMessage(ackId: Int): BufferMessage = createBufferMessage(new Array[ByteBuffer](0), ackId) + + def createBufferMessage(ackId: Int): BufferMessage = { + createBufferMessage(new Array[ByteBuffer](0), ackId) + } def create(header: MessageChunkHeader): Message = { val newMessage: Message = header.typ match { - case BUFFER_MESSAGE => new BufferMessage(header.id, ArrayBuffer(ByteBuffer.allocate(header.totalSize)), header.other) + case BUFFER_MESSAGE => new BufferMessage(header.id, + ArrayBuffer(ByteBuffer.allocate(header.totalSize)), header.other) } newMessage.senderAddress = header.address newMessage diff --git a/core/src/main/scala/spark/network/MessageChunk.scala b/core/src/main/scala/spark/network/MessageChunk.scala new file mode 100644 index 0000000000..784db5ab62 --- /dev/null +++ b/core/src/main/scala/spark/network/MessageChunk.scala @@ -0,0 +1,42 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network + +import java.nio.ByteBuffer + +import scala.collection.mutable.ArrayBuffer + + +private[network] +class MessageChunk(val header: MessageChunkHeader, val buffer: ByteBuffer) { + + val size = if (buffer == null) 0 else buffer.remaining + + lazy val buffers = { + val ab = new ArrayBuffer[ByteBuffer]() + ab += header.buffer + if (buffer != null) { + ab += buffer + } + ab + } + + override def toString = { + "" + this.getClass.getSimpleName + " (id = " + header.id + ", size = " + size + ")" + } +} diff --git a/core/src/main/scala/spark/network/MessageChunkHeader.scala b/core/src/main/scala/spark/network/MessageChunkHeader.scala new file mode 100644 index 0000000000..18d0cbcc14 --- /dev/null +++ b/core/src/main/scala/spark/network/MessageChunkHeader.scala @@ -0,0 +1,75 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network + +import java.net.InetAddress +import java.net.InetSocketAddress +import java.nio.ByteBuffer + + +private[spark] class MessageChunkHeader( + val typ: Long, + val id: Int, + val totalSize: Int, + val chunkSize: Int, + val other: Int, + val address: InetSocketAddress) { + lazy val buffer = { + // No need to change this, at 'use' time, we do a reverse lookup of the hostname. + // Refer to network.Connection + val ip = address.getAddress.getAddress() + val port = address.getPort() + ByteBuffer. + allocate(MessageChunkHeader.HEADER_SIZE). + putLong(typ). + putInt(id). + putInt(totalSize). + putInt(chunkSize). + putInt(other). + putInt(ip.size). + put(ip). + putInt(port). + position(MessageChunkHeader.HEADER_SIZE). + flip.asInstanceOf[ByteBuffer] + } + + override def toString = "" + this.getClass.getSimpleName + ":" + id + " of type " + typ + + " and sizes " + totalSize + " / " + chunkSize + " bytes" +} + + +private[spark] object MessageChunkHeader { + val HEADER_SIZE = 40 + + def create(buffer: ByteBuffer): MessageChunkHeader = { + if (buffer.remaining != HEADER_SIZE) { + throw new IllegalArgumentException("Cannot convert buffer data to Message") + } + val typ = buffer.getLong() + val id = buffer.getInt() + val totalSize = buffer.getInt() + val chunkSize = buffer.getInt() + val other = buffer.getInt() + val ipSize = buffer.getInt() + val ipBytes = new Array[Byte](ipSize) + buffer.get(ipBytes) + val ip = InetAddress.getByAddress(ipBytes) + val port = buffer.getInt() + new MessageChunkHeader(typ, id, totalSize, chunkSize, other, new InetSocketAddress(ip, port)) + } +} diff --git a/core/src/main/scala/spark/network/ReceiverTest.scala b/core/src/main/scala/spark/network/ReceiverTest.scala index a174d5f403..2bbc736f40 100644 --- a/core/src/main/scala/spark/network/ReceiverTest.scala +++ b/core/src/main/scala/spark/network/ReceiverTest.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.network import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/network/SenderTest.scala b/core/src/main/scala/spark/network/SenderTest.scala index a4ff69e4d2..542c54c36b 100644 --- a/core/src/main/scala/spark/network/SenderTest.scala +++ b/core/src/main/scala/spark/network/SenderTest.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.network import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/network/netty/FileHeader.scala b/core/src/main/scala/spark/network/netty/FileHeader.scala new file mode 100644 index 0000000000..bf46d32aa3 --- /dev/null +++ b/core/src/main/scala/spark/network/netty/FileHeader.scala @@ -0,0 +1,74 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty + +import io.netty.buffer._ + +import spark.Logging + +private[spark] class FileHeader ( + val fileLen: Int, + val blockId: String) extends Logging { + + lazy val buffer = { + val buf = Unpooled.buffer() + buf.capacity(FileHeader.HEADER_SIZE) + buf.writeInt(fileLen) + buf.writeInt(blockId.length) + blockId.foreach((x: Char) => buf.writeByte(x)) + //padding the rest of header + if (FileHeader.HEADER_SIZE - buf.readableBytes > 0 ) { + buf.writeZero(FileHeader.HEADER_SIZE - buf.readableBytes) + } else { + throw new Exception("too long header " + buf.readableBytes) + logInfo("too long header") + } + buf + } + +} + +private[spark] object FileHeader { + + val HEADER_SIZE = 40 + + def getFileLenOffset = 0 + def getFileLenSize = Integer.SIZE/8 + + def create(buf: ByteBuf): FileHeader = { + val length = buf.readInt + val idLength = buf.readInt + val idBuilder = new StringBuilder(idLength) + for (i <- 1 to idLength) { + idBuilder += buf.readByte().asInstanceOf[Char] + } + val blockId = idBuilder.toString() + new FileHeader(length, blockId) + } + + + def main (args:Array[String]){ + + val header = new FileHeader(25,"block_0"); + val buf = header.buffer; + val newheader = FileHeader.create(buf); + System.out.println("id="+newheader.blockId+",size="+newheader.fileLen) + + } +} + diff --git a/core/src/main/scala/spark/network/netty/ShuffleCopier.scala b/core/src/main/scala/spark/network/netty/ShuffleCopier.scala new file mode 100644 index 0000000000..b01f6369f6 --- /dev/null +++ b/core/src/main/scala/spark/network/netty/ShuffleCopier.scala @@ -0,0 +1,118 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty + +import java.util.concurrent.Executors + +import io.netty.buffer.ByteBuf +import io.netty.channel.ChannelHandlerContext +import io.netty.util.CharsetUtil + +import spark.Logging +import spark.network.ConnectionManagerId + +import scala.collection.JavaConverters._ + + +private[spark] class ShuffleCopier extends Logging { + + def getBlock(host: String, port: Int, blockId: String, + resultCollectCallback: (String, Long, ByteBuf) => Unit) { + + val handler = new ShuffleCopier.ShuffleClientHandler(resultCollectCallback) + val connectTimeout = System.getProperty("spark.shuffle.netty.connect.timeout", "60000").toInt + val fc = new FileClient(handler, connectTimeout) + + try { + fc.init() + fc.connect(host, port) + fc.sendRequest(blockId) + fc.waitForClose() + fc.close() + } catch { + // Handle any socket-related exceptions in FileClient + case e: Exception => { + logError("Shuffle copy of block " + blockId + " from " + host + ":" + port + " failed", e) + handler.handleError(blockId) + } + } + } + + def getBlock(cmId: ConnectionManagerId, blockId: String, + resultCollectCallback: (String, Long, ByteBuf) => Unit) { + getBlock(cmId.host, cmId.port, blockId, resultCollectCallback) + } + + def getBlocks(cmId: ConnectionManagerId, + blocks: Seq[(String, Long)], + resultCollectCallback: (String, Long, ByteBuf) => Unit) { + + for ((blockId, size) <- blocks) { + getBlock(cmId, blockId, resultCollectCallback) + } + } +} + + +private[spark] object ShuffleCopier extends Logging { + + private class ShuffleClientHandler(resultCollectCallBack: (String, Long, ByteBuf) => Unit) + extends FileClientHandler with Logging { + + override def handle(ctx: ChannelHandlerContext, in: ByteBuf, header: FileHeader) { + logDebug("Received Block: " + header.blockId + " (" + header.fileLen + "B)"); + resultCollectCallBack(header.blockId, header.fileLen.toLong, in.readBytes(header.fileLen)) + } + + override def handleError(blockId: String) { + if (!isComplete) { + resultCollectCallBack(blockId, -1, null) + } + } + } + + def echoResultCollectCallBack(blockId: String, size: Long, content: ByteBuf) { + if (size != -1) { + logInfo("File: " + blockId + " content is : \" " + content.toString(CharsetUtil.UTF_8) + "\"") + } + } + + def main(args: Array[String]) { + if (args.length < 3) { + System.err.println("Usage: ShuffleCopier <host> <port> <shuffle_block_id> <threads>") + System.exit(1) + } + val host = args(0) + val port = args(1).toInt + val file = args(2) + val threads = if (args.length > 3) args(3).toInt else 10 + + val copiers = Executors.newFixedThreadPool(80) + val tasks = (for (i <- Range(0, threads)) yield { + Executors.callable(new Runnable() { + def run() { + val copier = new ShuffleCopier() + copier.getBlock(host, port, file, echoResultCollectCallBack) + } + }) + }).asJava + copiers.invokeAll(tasks) + copiers.shutdown + System.exit(0) + } +} diff --git a/core/src/main/scala/spark/network/netty/ShuffleSender.scala b/core/src/main/scala/spark/network/netty/ShuffleSender.scala new file mode 100644 index 0000000000..cdf88b03a0 --- /dev/null +++ b/core/src/main/scala/spark/network/netty/ShuffleSender.scala @@ -0,0 +1,70 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.network.netty + +import java.io.File + +import spark.Logging + + +private[spark] class ShuffleSender(portIn: Int, val pResolver: PathResolver) extends Logging { + + val server = new FileServer(pResolver, portIn) + server.start() + + def stop() { + server.stop() + } + + def port: Int = server.getPort() +} + + +/** + * An application for testing the shuffle sender as a standalone program. + */ +private[spark] object ShuffleSender { + + def main(args: Array[String]) { + if (args.length < 3) { + System.err.println( + "Usage: ShuffleSender <port> <subDirsPerLocalDir> <list of shuffle_block_directories>") + System.exit(1) + } + + val port = args(0).toInt + val subDirsPerLocalDir = args(1).toInt + val localDirs = args.drop(2).map(new File(_)) + + val pResovler = new PathResolver { + override def getAbsolutePath(blockId: String): String = { + if (!blockId.startsWith("shuffle_")) { + throw new Exception("Block " + blockId + " is not a shuffle block") + } + // Figure out which local directory it hashes to, and which subdirectory in that + val hash = math.abs(blockId.hashCode) + val dirId = hash % localDirs.length + val subDirId = (hash / localDirs.length) % subDirsPerLocalDir + val subDir = new File(localDirs(dirId), "%02x".format(subDirId)) + val file = new File(subDir, blockId) + return file.getAbsolutePath + } + } + val sender = new ShuffleSender(port, pResovler) + } +} diff --git a/core/src/main/scala/spark/package.scala b/core/src/main/scala/spark/package.scala index 389ec4da3e..b244bfbf06 100644 --- a/core/src/main/scala/spark/package.scala +++ b/core/src/main/scala/spark/package.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + /** * Core Spark functionality. [[spark.SparkContext]] serves as the main entry point to Spark, while * [[spark.RDD]] is the data type representing a distributed collection, and provides most diff --git a/core/src/main/scala/spark/partial/ApproximateActionListener.scala b/core/src/main/scala/spark/partial/ApproximateActionListener.scala index de2dce161a..691d939150 100644 --- a/core/src/main/scala/spark/partial/ApproximateActionListener.scala +++ b/core/src/main/scala/spark/partial/ApproximateActionListener.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial import spark._ diff --git a/core/src/main/scala/spark/partial/ApproximateEvaluator.scala b/core/src/main/scala/spark/partial/ApproximateEvaluator.scala index 75713b2eaa..5eae144dfb 100644 --- a/core/src/main/scala/spark/partial/ApproximateEvaluator.scala +++ b/core/src/main/scala/spark/partial/ApproximateEvaluator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial /** diff --git a/core/src/main/scala/spark/partial/BoundedDouble.scala b/core/src/main/scala/spark/partial/BoundedDouble.scala index 463c33d6e2..8bdbe6c012 100644 --- a/core/src/main/scala/spark/partial/BoundedDouble.scala +++ b/core/src/main/scala/spark/partial/BoundedDouble.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial /** diff --git a/core/src/main/scala/spark/partial/CountEvaluator.scala b/core/src/main/scala/spark/partial/CountEvaluator.scala index daf2c5170c..6aa92094eb 100644 --- a/core/src/main/scala/spark/partial/CountEvaluator.scala +++ b/core/src/main/scala/spark/partial/CountEvaluator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial import cern.jet.stat.Probability diff --git a/core/src/main/scala/spark/partial/GroupedCountEvaluator.scala b/core/src/main/scala/spark/partial/GroupedCountEvaluator.scala index 01fbb8a11b..ebe2e5a1e3 100644 --- a/core/src/main/scala/spark/partial/GroupedCountEvaluator.scala +++ b/core/src/main/scala/spark/partial/GroupedCountEvaluator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial import java.util.{HashMap => JHashMap} diff --git a/core/src/main/scala/spark/partial/GroupedMeanEvaluator.scala b/core/src/main/scala/spark/partial/GroupedMeanEvaluator.scala index c622df5220..2dadbbd5fb 100644 --- a/core/src/main/scala/spark/partial/GroupedMeanEvaluator.scala +++ b/core/src/main/scala/spark/partial/GroupedMeanEvaluator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial import java.util.{HashMap => JHashMap} diff --git a/core/src/main/scala/spark/partial/GroupedSumEvaluator.scala b/core/src/main/scala/spark/partial/GroupedSumEvaluator.scala index 20fa55cff2..ae2b63f7cb 100644 --- a/core/src/main/scala/spark/partial/GroupedSumEvaluator.scala +++ b/core/src/main/scala/spark/partial/GroupedSumEvaluator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial import java.util.{HashMap => JHashMap} diff --git a/core/src/main/scala/spark/partial/MeanEvaluator.scala b/core/src/main/scala/spark/partial/MeanEvaluator.scala index 762c85400d..5ddcad7075 100644 --- a/core/src/main/scala/spark/partial/MeanEvaluator.scala +++ b/core/src/main/scala/spark/partial/MeanEvaluator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial import cern.jet.stat.Probability diff --git a/core/src/main/scala/spark/partial/PartialResult.scala b/core/src/main/scala/spark/partial/PartialResult.scala index 200ed4ea1e..922a9f9bc6 100644 --- a/core/src/main/scala/spark/partial/PartialResult.scala +++ b/core/src/main/scala/spark/partial/PartialResult.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial class PartialResult[R](initialVal: R, isFinal: Boolean) { diff --git a/core/src/main/scala/spark/partial/StudentTCacher.scala b/core/src/main/scala/spark/partial/StudentTCacher.scala index 443abba5cd..f3bb987d46 100644 --- a/core/src/main/scala/spark/partial/StudentTCacher.scala +++ b/core/src/main/scala/spark/partial/StudentTCacher.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial import cern.jet.stat.Probability diff --git a/core/src/main/scala/spark/partial/SumEvaluator.scala b/core/src/main/scala/spark/partial/SumEvaluator.scala index 58fb60f441..4083abef03 100644 --- a/core/src/main/scala/spark/partial/SumEvaluator.scala +++ b/core/src/main/scala/spark/partial/SumEvaluator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.partial import cern.jet.stat.Probability diff --git a/core/src/main/scala/spark/rdd/BlockRDD.scala b/core/src/main/scala/spark/rdd/BlockRDD.scala index 7348c4f15b..0ebb722d73 100644 --- a/core/src/main/scala/spark/rdd/BlockRDD.scala +++ b/core/src/main/scala/spark/rdd/BlockRDD.scala @@ -1,7 +1,24 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd -import scala.collection.mutable.HashMap import spark.{RDD, SparkContext, SparkEnv, Partition, TaskContext} +import spark.storage.BlockManager private[spark] class BlockRDDPartition(val blockId: String, idx: Int) extends Partition { val index = idx @@ -11,12 +28,7 @@ private[spark] class BlockRDD[T: ClassManifest](sc: SparkContext, @transient blockIds: Array[String]) extends RDD[T](sc, Nil) { - @transient lazy val locations_ = { - val blockManager = SparkEnv.get.blockManager - /*val locations = blockIds.map(id => blockManager.getLocations(id))*/ - val locations = blockManager.getLocations(blockIds) - HashMap(blockIds.zip(locations):_*) - } + @transient lazy val locations_ = BlockManager.blockIdsToExecutorLocations(blockIds, SparkEnv.get) override def getPartitions: Array[Partition] = (0 until blockIds.size).map(i => { new BlockRDDPartition(blockIds(i), i).asInstanceOf[Partition] diff --git a/core/src/main/scala/spark/rdd/CartesianRDD.scala b/core/src/main/scala/spark/rdd/CartesianRDD.scala index 38600b8be4..150e5bca29 100644 --- a/core/src/main/scala/spark/rdd/CartesianRDD.scala +++ b/core/src/main/scala/spark/rdd/CartesianRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import java.io.{ObjectOutputStream, IOException} diff --git a/core/src/main/scala/spark/rdd/CheckpointRDD.scala b/core/src/main/scala/spark/rdd/CheckpointRDD.scala index 43ee39c993..6794e0e201 100644 --- a/core/src/main/scala/spark/rdd/CheckpointRDD.scala +++ b/core/src/main/scala/spark/rdd/CheckpointRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark._ @@ -43,7 +60,7 @@ class CheckpointRDD[T: ClassManifest](sc: SparkContext, val checkpointPath: Stri checkpointData.get.cpFile = Some(checkpointPath) override def getPreferredLocations(split: Partition): Seq[String] = { - val status = fs.getFileStatus(new Path(checkpointPath)) + val status = fs.getFileStatus(new Path(checkpointPath, CheckpointRDD.splitIdToFile(split.index))) val locations = fs.getFileBlockLocations(status, 0, status.getLen) locations.headOption.toList.flatMap(_.getHosts).filter(_ != "localhost") } diff --git a/core/src/main/scala/spark/rdd/CoGroupedRDD.scala b/core/src/main/scala/spark/rdd/CoGroupedRDD.scala index a6235491ca..c540cd36eb 100644 --- a/core/src/main/scala/spark/rdd/CoGroupedRDD.scala +++ b/core/src/main/scala/spark/rdd/CoGroupedRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import java.io.{ObjectOutputStream, IOException} @@ -6,7 +23,7 @@ import java.util.{HashMap => JHashMap} import scala.collection.JavaConversions import scala.collection.mutable.ArrayBuffer -import spark.{Aggregator, Logging, Partition, Partitioner, RDD, SparkEnv, TaskContext} +import spark.{Aggregator, Partition, Partitioner, RDD, SparkEnv, TaskContext} import spark.{Dependency, OneToOneDependency, ShuffleDependency} @@ -49,12 +66,17 @@ private[spark] class CoGroupAggregator * * @param rdds parent RDDs. * @param part partitioner used to partition the shuffle output. - * @param mapSideCombine flag indicating whether to merge values before shuffle step. + * @param mapSideCombine flag indicating whether to merge values before shuffle step. If the flag + * is on, Spark does an extra pass over the data on the map side to merge + * all values belonging to the same key together. This can reduce the amount + * of data shuffled if and only if the number of distinct keys is very small, + * and the ratio of key size to value size is also very small. */ class CoGroupedRDD[K]( @transient var rdds: Seq[RDD[(K, _)]], part: Partitioner, - val mapSideCombine: Boolean = true) + val mapSideCombine: Boolean = false, + val serializerClass: String = null) extends RDD[(K, Seq[Seq[_]])](rdds.head.context, Nil) { private val aggr = new CoGroupAggregator @@ -68,9 +90,9 @@ class CoGroupedRDD[K]( logInfo("Adding shuffle dependency with " + rdd) if (mapSideCombine) { val mapSideCombinedRDD = rdd.mapPartitions(aggr.combineValuesByKey(_), true) - new ShuffleDependency[Any, ArrayBuffer[Any]](mapSideCombinedRDD, part) + new ShuffleDependency[Any, ArrayBuffer[Any]](mapSideCombinedRDD, part, serializerClass) } else { - new ShuffleDependency[Any, Any](rdd.asInstanceOf[RDD[(Any, Any)]], part) + new ShuffleDependency[Any, Any](rdd.asInstanceOf[RDD[(Any, Any)]], part, serializerClass) } } } @@ -112,6 +134,7 @@ class CoGroupedRDD[K]( } } + val ser = SparkEnv.get.serializerManager.get(serializerClass) for ((dep, depNum) <- split.deps.zipWithIndex) dep match { case NarrowCoGroupSplitDep(rdd, _, itsSplit) => { // Read them from the parent @@ -124,12 +147,12 @@ class CoGroupedRDD[K]( val fetcher = SparkEnv.get.shuffleFetcher if (mapSideCombine) { // With map side combine on, for each key, the shuffle fetcher returns a list of values. - fetcher.fetch[K, Seq[Any]](shuffleId, split.index, context.taskMetrics).foreach { + fetcher.fetch[K, Seq[Any]](shuffleId, split.index, context.taskMetrics, ser).foreach { case (key, values) => getSeq(key)(depNum) ++= values } } else { // With map side combine off, for each key the shuffle fetcher returns a single value. - fetcher.fetch[K, Any](shuffleId, split.index, context.taskMetrics).foreach { + fetcher.fetch[K, Any](shuffleId, split.index, context.taskMetrics, ser).foreach { case (key, value) => getSeq(key)(depNum) += value } } diff --git a/core/src/main/scala/spark/rdd/CoalescedRDD.scala b/core/src/main/scala/spark/rdd/CoalescedRDD.scala index 6d862c0c28..2b5bf18541 100644 --- a/core/src/main/scala/spark/rdd/CoalescedRDD.scala +++ b/core/src/main/scala/spark/rdd/CoalescedRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{Dependency, OneToOneDependency, NarrowDependency, RDD, Partition, TaskContext} diff --git a/core/src/main/scala/spark/rdd/EmptyRDD.scala b/core/src/main/scala/spark/rdd/EmptyRDD.scala new file mode 100644 index 0000000000..d7d4db5d30 --- /dev/null +++ b/core/src/main/scala/spark/rdd/EmptyRDD.scala @@ -0,0 +1,33 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.rdd + +import spark.{RDD, SparkContext, SparkEnv, Partition, TaskContext} + + +/** + * An RDD that is empty, i.e. has no element in it. + */ +class EmptyRDD[T: ClassManifest](sc: SparkContext) extends RDD[T](sc, Nil) { + + override def getPartitions: Array[Partition] = Array.empty + + override def compute(split: Partition, context: TaskContext): Iterator[T] = { + throw new UnsupportedOperationException("empty RDD") + } +} diff --git a/core/src/main/scala/spark/rdd/FilteredRDD.scala b/core/src/main/scala/spark/rdd/FilteredRDD.scala index c84ec39d21..783508cfd1 100644 --- a/core/src/main/scala/spark/rdd/FilteredRDD.scala +++ b/core/src/main/scala/spark/rdd/FilteredRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{OneToOneDependency, RDD, Partition, TaskContext} diff --git a/core/src/main/scala/spark/rdd/FlatMappedRDD.scala b/core/src/main/scala/spark/rdd/FlatMappedRDD.scala index 8ebc778925..ed75eac3ff 100644 --- a/core/src/main/scala/spark/rdd/FlatMappedRDD.scala +++ b/core/src/main/scala/spark/rdd/FlatMappedRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{RDD, Partition, TaskContext} diff --git a/core/src/main/scala/spark/rdd/GlommedRDD.scala b/core/src/main/scala/spark/rdd/GlommedRDD.scala index e16c7ba881..1573f8a289 100644 --- a/core/src/main/scala/spark/rdd/GlommedRDD.scala +++ b/core/src/main/scala/spark/rdd/GlommedRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{RDD, Partition, TaskContext} diff --git a/core/src/main/scala/spark/rdd/HadoopRDD.scala b/core/src/main/scala/spark/rdd/HadoopRDD.scala index cbf5512e24..d0fdeb741e 100644 --- a/core/src/main/scala/spark/rdd/HadoopRDD.scala +++ b/core/src/main/scala/spark/rdd/HadoopRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import java.io.EOFException @@ -15,6 +32,7 @@ import org.apache.hadoop.mapred.RecordReader import org.apache.hadoop.mapred.Reporter import org.apache.hadoop.util.ReflectionUtils +import spark.deploy.SparkHadoopUtil import spark.{Dependency, Logging, Partition, RDD, SerializableWritable, SparkContext, TaskContext} import spark.util.NextIterator import org.apache.hadoop.conf.Configurable @@ -50,6 +68,7 @@ class HadoopRDD[K, V]( private val confBroadcast = sc.broadcast(new SerializableWritable(conf)) override def getPartitions: Array[Partition] = { + SparkHadoopUtil.addCredentials(conf); val inputFormat = createInputFormat(conf) if (inputFormat.isInstanceOf[Configurable]) { inputFormat.asInstanceOf[Configurable].setConf(conf) diff --git a/core/src/main/scala/spark/rdd/JdbcRDD.scala b/core/src/main/scala/spark/rdd/JdbcRDD.scala new file mode 100644 index 0000000000..59132437d2 --- /dev/null +++ b/core/src/main/scala/spark/rdd/JdbcRDD.scala @@ -0,0 +1,120 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.rdd + +import java.sql.{Connection, ResultSet} + +import spark.{Logging, Partition, RDD, SparkContext, TaskContext} +import spark.util.NextIterator + +private[spark] class JdbcPartition(idx: Int, val lower: Long, val upper: Long) extends Partition { + override def index = idx +} + +/** + * An RDD that executes an SQL query on a JDBC connection and reads results. + * For usage example, see test case JdbcRDDSuite. + * + * @param getConnection a function that returns an open Connection. + * The RDD takes care of closing the connection. + * @param sql the text of the query. + * The query must contain two ? placeholders for parameters used to partition the results. + * E.g. "select title, author from books where ? <= id and id <= ?" + * @param lowerBound the minimum value of the first placeholder + * @param upperBound the maximum value of the second placeholder + * The lower and upper bounds are inclusive. + * @param numPartitions the number of partitions. + * Given a lowerBound of 1, an upperBound of 20, and a numPartitions of 2, + * the query would be executed twice, once with (1, 10) and once with (11, 20) + * @param mapRow a function from a ResultSet to a single row of the desired result type(s). + * This should only call getInt, getString, etc; the RDD takes care of calling next. + * The default maps a ResultSet to an array of Object. + */ +class JdbcRDD[T: ClassManifest]( + sc: SparkContext, + getConnection: () => Connection, + sql: String, + lowerBound: Long, + upperBound: Long, + numPartitions: Int, + mapRow: (ResultSet) => T = JdbcRDD.resultSetToObjectArray _) + extends RDD[T](sc, Nil) with Logging { + + override def getPartitions: Array[Partition] = { + // bounds are inclusive, hence the + 1 here and - 1 on end + val length = 1 + upperBound - lowerBound + (0 until numPartitions).map(i => { + val start = lowerBound + ((i * length) / numPartitions).toLong + val end = lowerBound + (((i + 1) * length) / numPartitions).toLong - 1 + new JdbcPartition(i, start, end) + }).toArray + } + + override def compute(thePart: Partition, context: TaskContext) = new NextIterator[T] { + context.addOnCompleteCallback{ () => closeIfNeeded() } + val part = thePart.asInstanceOf[JdbcPartition] + val conn = getConnection() + val stmt = conn.prepareStatement(sql, ResultSet.TYPE_FORWARD_ONLY, ResultSet.CONCUR_READ_ONLY) + + // setFetchSize(Integer.MIN_VALUE) is a mysql driver specific way to force streaming results, + // rather than pulling entire resultset into memory. + // see http://dev.mysql.com/doc/refman/5.0/en/connector-j-reference-implementation-notes.html + if (conn.getMetaData.getURL.matches("jdbc:mysql:.*")) { + stmt.setFetchSize(Integer.MIN_VALUE) + logInfo("statement fetch size set to: " + stmt.getFetchSize + " to force MySQL streaming ") + } + + stmt.setLong(1, part.lower) + stmt.setLong(2, part.upper) + val rs = stmt.executeQuery() + + override def getNext: T = { + if (rs.next()) { + mapRow(rs) + } else { + finished = true + null.asInstanceOf[T] + } + } + + override def close() { + try { + if (null != rs && ! rs.isClosed()) rs.close() + } catch { + case e: Exception => logWarning("Exception closing resultset", e) + } + try { + if (null != stmt && ! stmt.isClosed()) stmt.close() + } catch { + case e: Exception => logWarning("Exception closing statement", e) + } + try { + if (null != conn && ! stmt.isClosed()) conn.close() + logInfo("closed connection") + } catch { + case e: Exception => logWarning("Exception closing connection", e) + } + } + } +} + +object JdbcRDD { + def resultSetToObjectArray(rs: ResultSet) = { + Array.tabulate[Object](rs.getMetaData.getColumnCount)(i => rs.getObject(i + 1)) + } +} diff --git a/core/src/main/scala/spark/rdd/MapPartitionsRDD.scala b/core/src/main/scala/spark/rdd/MapPartitionsRDD.scala index d283c5b2bb..af8f0a112f 100644 --- a/core/src/main/scala/spark/rdd/MapPartitionsRDD.scala +++ b/core/src/main/scala/spark/rdd/MapPartitionsRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{RDD, Partition, TaskContext} diff --git a/core/src/main/scala/spark/rdd/MapPartitionsWithIndexRDD.scala b/core/src/main/scala/spark/rdd/MapPartitionsWithIndexRDD.scala index afb7504ba1..3b4e9518fd 100644 --- a/core/src/main/scala/spark/rdd/MapPartitionsWithIndexRDD.scala +++ b/core/src/main/scala/spark/rdd/MapPartitionsWithIndexRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{RDD, Partition, TaskContext} diff --git a/core/src/main/scala/spark/rdd/MappedRDD.scala b/core/src/main/scala/spark/rdd/MappedRDD.scala index af07311b6d..8b411dd85d 100644 --- a/core/src/main/scala/spark/rdd/MappedRDD.scala +++ b/core/src/main/scala/spark/rdd/MappedRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{RDD, Partition, TaskContext} diff --git a/core/src/main/scala/spark/rdd/NewHadoopRDD.scala b/core/src/main/scala/spark/rdd/NewHadoopRDD.scala index 901d01ef30..17fe805fd4 100644 --- a/core/src/main/scala/spark/rdd/NewHadoopRDD.scala +++ b/core/src/main/scala/spark/rdd/NewHadoopRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import java.text.SimpleDateFormat diff --git a/core/src/main/scala/spark/rdd/ParallelCollectionRDD.scala b/core/src/main/scala/spark/rdd/ParallelCollectionRDD.scala index 07585a88ce..16ba0c26f8 100644 --- a/core/src/main/scala/spark/rdd/ParallelCollectionRDD.scala +++ b/core/src/main/scala/spark/rdd/ParallelCollectionRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import scala.collection.immutable.NumericRange diff --git a/core/src/main/scala/spark/rdd/PartitionPruningRDD.scala b/core/src/main/scala/spark/rdd/PartitionPruningRDD.scala index 41ff62dd22..191cfde565 100644 --- a/core/src/main/scala/spark/rdd/PartitionPruningRDD.scala +++ b/core/src/main/scala/spark/rdd/PartitionPruningRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{NarrowDependency, RDD, SparkEnv, Partition, TaskContext} diff --git a/core/src/main/scala/spark/rdd/PipedRDD.scala b/core/src/main/scala/spark/rdd/PipedRDD.scala index 962a1b21ad..2cefdc78b0 100644 --- a/core/src/main/scala/spark/rdd/PipedRDD.scala +++ b/core/src/main/scala/spark/rdd/PipedRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import java.io.PrintWriter @@ -9,6 +26,7 @@ import scala.collection.mutable.ArrayBuffer import scala.io.Source import spark.{RDD, SparkEnv, Partition, TaskContext} +import spark.broadcast.Broadcast /** @@ -18,14 +36,21 @@ import spark.{RDD, SparkEnv, Partition, TaskContext} class PipedRDD[T: ClassManifest]( prev: RDD[T], command: Seq[String], - envVars: Map[String, String]) + envVars: Map[String, String], + printPipeContext: (String => Unit) => Unit, + printRDDElement: (T, String => Unit) => Unit) extends RDD[String](prev) { - def this(prev: RDD[T], command: Seq[String]) = this(prev, command, Map()) - // Similar to Runtime.exec(), if we are given a single string, split it into words // using a standard StringTokenizer (i.e. by spaces) - def this(prev: RDD[T], command: String) = this(prev, PipedRDD.tokenize(command)) + def this( + prev: RDD[T], + command: String, + envVars: Map[String, String] = Map(), + printPipeContext: (String => Unit) => Unit = null, + printRDDElement: (T, String => Unit) => Unit = null) = + this(prev, PipedRDD.tokenize(command), envVars, printPipeContext, printRDDElement) + override def getPartitions: Array[Partition] = firstParent[T].partitions @@ -52,8 +77,17 @@ class PipedRDD[T: ClassManifest]( override def run() { SparkEnv.set(env) val out = new PrintWriter(proc.getOutputStream) + + // input the pipe context firstly + if (printPipeContext != null) { + printPipeContext(out.println(_)) + } for (elem <- firstParent[T].iterator(split, context)) { - out.println(elem) + if (printRDDElement != null) { + printRDDElement(elem, out.println(_)) + } else { + out.println(elem) + } } out.close() } diff --git a/core/src/main/scala/spark/rdd/SampledRDD.scala b/core/src/main/scala/spark/rdd/SampledRDD.scala index 243673f151..574c9b141d 100644 --- a/core/src/main/scala/spark/rdd/SampledRDD.scala +++ b/core/src/main/scala/spark/rdd/SampledRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import java.util.Random diff --git a/core/src/main/scala/spark/rdd/ShuffledRDD.scala b/core/src/main/scala/spark/rdd/ShuffledRDD.scala index 4e33b7dd5c..0137f80953 100644 --- a/core/src/main/scala/spark/rdd/ShuffledRDD.scala +++ b/core/src/main/scala/spark/rdd/ShuffledRDD.scala @@ -1,8 +1,26 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import spark.{Partitioner, RDD, SparkEnv, ShuffleDependency, Partition, TaskContext} import spark.SparkContext._ + private[spark] class ShuffledRDDPartition(val idx: Int) extends Partition { override val index = idx override def hashCode(): Int = idx @@ -12,13 +30,15 @@ private[spark] class ShuffledRDDPartition(val idx: Int) extends Partition { * The resulting RDD from a shuffle (e.g. repartitioning of data). * @param prev the parent RDD. * @param part the partitioner used to partition the RDD + * @param serializerClass class name of the serializer to use. * @tparam K the key class. * @tparam V the value class. */ class ShuffledRDD[K, V]( @transient prev: RDD[(K, V)], - part: Partitioner) - extends RDD[(K, V)](prev.context, List(new ShuffleDependency(prev, part))) { + part: Partitioner, + serializerClass: String = null) + extends RDD[(K, V)](prev.context, List(new ShuffleDependency(prev, part, serializerClass))) { override val partitioner = Some(part) @@ -28,6 +48,7 @@ class ShuffledRDD[K, V]( override def compute(split: Partition, context: TaskContext): Iterator[(K, V)] = { val shuffledId = dependencies.head.asInstanceOf[ShuffleDependency[K, V]].shuffleId - SparkEnv.get.shuffleFetcher.fetch[K, V](shuffledId, split.index, context.taskMetrics) + SparkEnv.get.shuffleFetcher.fetch[K, V](shuffledId, split.index, context.taskMetrics, + SparkEnv.get.serializerManager.get(serializerClass)) } } diff --git a/core/src/main/scala/spark/rdd/SubtractedRDD.scala b/core/src/main/scala/spark/rdd/SubtractedRDD.scala index 481e03b349..0402b9f250 100644 --- a/core/src/main/scala/spark/rdd/SubtractedRDD.scala +++ b/core/src/main/scala/spark/rdd/SubtractedRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import java.util.{HashMap => JHashMap} @@ -12,6 +29,7 @@ import spark.SparkEnv import spark.ShuffleDependency import spark.OneToOneDependency + /** * An optimized version of cogroup for set difference/subtraction. * @@ -31,7 +49,9 @@ import spark.OneToOneDependency private[spark] class SubtractedRDD[K: ClassManifest, V: ClassManifest, W: ClassManifest]( @transient var rdd1: RDD[(K, V)], @transient var rdd2: RDD[(K, W)], - part: Partitioner) extends RDD[(K, V)](rdd1.context, Nil) { + part: Partitioner, + val serializerClass: String = null) + extends RDD[(K, V)](rdd1.context, Nil) { override def getDependencies: Seq[Dependency[_]] = { Seq(rdd1, rdd2).map { rdd => @@ -40,7 +60,7 @@ private[spark] class SubtractedRDD[K: ClassManifest, V: ClassManifest, W: ClassM new OneToOneDependency(rdd) } else { logInfo("Adding shuffle dependency with " + rdd) - new ShuffleDependency(rdd.asInstanceOf[RDD[(K, Any)]], part) + new ShuffleDependency(rdd.asInstanceOf[RDD[(K, Any)]], part, serializerClass) } } } @@ -65,6 +85,7 @@ private[spark] class SubtractedRDD[K: ClassManifest, V: ClassManifest, W: ClassM override def compute(p: Partition, context: TaskContext): Iterator[(K, V)] = { val partition = p.asInstanceOf[CoGroupPartition] + val serializer = SparkEnv.get.serializerManager.get(serializerClass) val map = new JHashMap[K, ArrayBuffer[V]] def getSeq(k: K): ArrayBuffer[V] = { val seq = map.get(k) @@ -77,12 +98,16 @@ private[spark] class SubtractedRDD[K: ClassManifest, V: ClassManifest, W: ClassM } } def integrate(dep: CoGroupSplitDep, op: ((K, V)) => Unit) = dep match { - case NarrowCoGroupSplitDep(rdd, _, itsSplit) => + case NarrowCoGroupSplitDep(rdd, _, itsSplit) => { for (t <- rdd.iterator(itsSplit, context)) op(t.asInstanceOf[(K, V)]) - case ShuffleCoGroupSplitDep(shuffleId) => - for (t <- SparkEnv.get.shuffleFetcher.fetch(shuffleId, partition.index, context.taskMetrics)) + } + case ShuffleCoGroupSplitDep(shuffleId) => { + val iter = SparkEnv.get.shuffleFetcher.fetch(shuffleId, partition.index, + context.taskMetrics, serializer) + for (t <- iter) op(t.asInstanceOf[(K, V)]) + } } // the first dep is rdd1; add all values to the map integrate(partition.deps(0), t => getSeq(t._1) += t._2) diff --git a/core/src/main/scala/spark/rdd/UnionRDD.scala b/core/src/main/scala/spark/rdd/UnionRDD.scala index 2c52a67e22..2776826f18 100644 --- a/core/src/main/scala/spark/rdd/UnionRDD.scala +++ b/core/src/main/scala/spark/rdd/UnionRDD.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import scala.collection.mutable.ArrayBuffer diff --git a/core/src/main/scala/spark/rdd/ZippedPartitionsRDD.scala b/core/src/main/scala/spark/rdd/ZippedPartitionsRDD.scala index fc3f29ffcd..6a4fa13ad6 100644 --- a/core/src/main/scala/spark/rdd/ZippedPartitionsRDD.scala +++ b/core/src/main/scala/spark/rdd/ZippedPartitionsRDD.scala @@ -1,6 +1,23 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd -import spark.{OneToOneDependency, RDD, SparkContext, Partition, TaskContext} +import spark.{Utils, OneToOneDependency, RDD, SparkContext, Partition, TaskContext} import java.io.{ObjectOutputStream, IOException} private[spark] class ZippedPartitionsPartition( @@ -38,9 +55,27 @@ abstract class ZippedPartitionsBaseRDD[V: ClassManifest]( } override def getPreferredLocations(s: Partition): Seq[String] = { + // Note that as number of rdd's increase and/or number of slaves in cluster increase, the computed preferredLocations below + // become diminishingly small : so we might need to look at alternate strategies to alleviate this. + // If there are no (or very small number of preferred locations), we will end up transferred the blocks to 'any' node in the + // cluster - paying with n/w and cache cost. + // Maybe pick a node which figures max amount of time ? + // Choose node which is hosting 'larger' of some subset of blocks ? + // Look at rack locality to ensure chosen host is atleast rack local to both hosting node ?, etc (would be good to defer this if possible) val splits = s.asInstanceOf[ZippedPartitionsPartition].partitions - val preferredLocations = rdds.zip(splits).map(x => x._1.preferredLocations(x._2)) - preferredLocations.reduce((x, y) => x.intersect(y)) + val rddSplitZip = rdds.zip(splits) + + // exact match. + val exactMatchPreferredLocations = rddSplitZip.map(x => x._1.preferredLocations(x._2)) + val exactMatchLocations = exactMatchPreferredLocations.reduce((x, y) => x.intersect(y)) + + // Remove exact match and then do host local match. + val exactMatchHosts = exactMatchLocations.map(Utils.parseHostPort(_)._1) + val matchPreferredHosts = exactMatchPreferredLocations.map(locs => locs.map(Utils.parseHostPort(_)._1)) + .reduce((x, y) => x.intersect(y)) + val otherNodeLocalLocations = matchPreferredHosts.filter { s => !exactMatchHosts.contains(s) } + + otherNodeLocalLocations ++ exactMatchLocations } override def clearDependencies() { diff --git a/core/src/main/scala/spark/rdd/ZippedRDD.scala b/core/src/main/scala/spark/rdd/ZippedRDD.scala index 35b0e06785..b1c43b3195 100644 --- a/core/src/main/scala/spark/rdd/ZippedRDD.scala +++ b/core/src/main/scala/spark/rdd/ZippedRDD.scala @@ -1,6 +1,23 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd -import spark.{OneToOneDependency, RDD, SparkContext, Partition, TaskContext} +import spark.{Utils, OneToOneDependency, RDD, SparkContext, Partition, TaskContext} import java.io.{ObjectOutputStream, IOException} @@ -48,8 +65,27 @@ class ZippedRDD[T: ClassManifest, U: ClassManifest]( } override def getPreferredLocations(s: Partition): Seq[String] = { + // Note that as number of slaves in cluster increase, the computed preferredLocations can become small : so we might need + // to look at alternate strategies to alleviate this. (If there are no (or very small number of preferred locations), we + // will end up transferred the blocks to 'any' node in the cluster - paying with n/w and cache cost. + // Maybe pick one or the other ? (so that atleast one block is local ?). + // Choose node which is hosting 'larger' of the blocks ? + // Look at rack locality to ensure chosen host is atleast rack local to both hosting node ?, etc (would be good to defer this if possible) val (partition1, partition2) = s.asInstanceOf[ZippedPartition[T, U]].partitions - rdd1.preferredLocations(partition1).intersect(rdd2.preferredLocations(partition2)) + val pref1 = rdd1.preferredLocations(partition1) + val pref2 = rdd2.preferredLocations(partition2) + + // exact match - instance local and host local. + val exactMatchLocations = pref1.intersect(pref2) + + // remove locations which are already handled via exactMatchLocations, and intersect where both partitions are node local. + val otherNodeLocalPref1 = pref1.filter(loc => ! exactMatchLocations.contains(loc)).map(loc => Utils.parseHostPort(loc)._1) + val otherNodeLocalPref2 = pref2.filter(loc => ! exactMatchLocations.contains(loc)).map(loc => Utils.parseHostPort(loc)._1) + val otherNodeLocalLocations = otherNodeLocalPref1.intersect(otherNodeLocalPref2) + + + // Can have mix of instance local (hostPort) and node local (host) locations as preference ! + exactMatchLocations ++ otherNodeLocalLocations } override def clearDependencies() { diff --git a/core/src/main/scala/spark/scheduler/ActiveJob.scala b/core/src/main/scala/spark/scheduler/ActiveJob.scala index 5a4e9a582d..71cc94edb6 100644 --- a/core/src/main/scala/spark/scheduler/ActiveJob.scala +++ b/core/src/main/scala/spark/scheduler/ActiveJob.scala @@ -1,7 +1,26 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import spark.TaskContext +import java.util.Properties + /** * Tracks information about an active job in the DAGScheduler. */ @@ -11,7 +30,8 @@ private[spark] class ActiveJob( val func: (TaskContext, Iterator[_]) => _, val partitions: Array[Int], val callSite: String, - val listener: JobListener) { + val listener: JobListener, + val properties: Properties) { val numPartitions = partitions.length val finished = Array.fill[Boolean](numPartitions)(false) diff --git a/core/src/main/scala/spark/scheduler/DAGScheduler.scala b/core/src/main/scala/spark/scheduler/DAGScheduler.scala index 1440b93e65..9b45fc2938 100644 --- a/core/src/main/scala/spark/scheduler/DAGScheduler.scala +++ b/core/src/main/scala/spark/scheduler/DAGScheduler.scala @@ -1,18 +1,34 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler -import cluster.TaskInfo +import java.io.NotSerializableException +import java.util.Properties +import java.util.concurrent.{LinkedBlockingQueue, TimeUnit} import java.util.concurrent.atomic.AtomicInteger -import java.util.concurrent.LinkedBlockingQueue -import java.util.concurrent.TimeUnit import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet, Map} import spark._ import spark.executor.TaskMetrics -import spark.partial.ApproximateActionListener -import spark.partial.ApproximateEvaluator -import spark.partial.PartialResult -import spark.storage.BlockManagerMaster +import spark.partial.{ApproximateActionListener, ApproximateEvaluator, PartialResult} +import spark.scheduler.cluster.TaskInfo +import spark.storage.{BlockManager, BlockManagerMaster} import spark.util.{MetadataCleaner, TimeStampedHashMap} /** @@ -34,6 +50,11 @@ class DAGScheduler( } taskSched.setListener(this) + // Called by TaskScheduler to report task's starting. + override def taskStarted(task: Task[_], taskInfo: TaskInfo) { + eventQueue.put(BeginEvent(task, taskInfo)) + } + // Called by TaskScheduler to report task completions or failures. override def taskEnded( task: Task[_], @@ -93,6 +114,8 @@ class DAGScheduler( // stray messages to detect. val failedGeneration = new HashMap[String, Long] + val idToActiveJob = new HashMap[Int, ActiveJob] + val waiting = new HashSet[Stage] // Stages we need to run whose parents aren't done val running = new HashSet[Stage] // Stages we are running right now val failed = new HashSet[Stage] // Stages that must be resubmitted due to fetch failures @@ -117,9 +140,8 @@ class DAGScheduler( private def getCacheLocs(rdd: RDD[_]): Array[List[String]] = { if (!cacheLocs.contains(rdd.id)) { val blockIds = rdd.partitions.indices.map(index=> "rdd_%d_%d".format(rdd.id, index)).toArray - cacheLocs(rdd.id) = blockManagerMaster.getLocations(blockIds).map { - locations => locations.map(_.hostPort).toList - }.toArray + val locs = BlockManager.blockIdsToExecutorLocations(blockIds, env, blockManagerMaster) + cacheLocs(rdd.id) = blockIds.map(locs.getOrElse(_, Nil)) } cacheLocs(rdd.id) } @@ -148,7 +170,13 @@ class DAGScheduler( * as a result stage for the final RDD used directly in an action. The stage will also be given * the provided priority. */ - private def newStage(rdd: RDD[_], shuffleDep: Option[ShuffleDependency[_,_]], priority: Int): Stage = { + private def newStage( + rdd: RDD[_], + shuffleDep: Option[ShuffleDependency[_,_]], + priority: Int, + callSite: Option[String] = None) + : Stage = + { if (shuffleDep != None) { // Kind of ugly: need to register RDDs with the cache and map output tracker here // since we can't do it in the RDD constructor because # of partitions is unknown @@ -156,7 +184,7 @@ class DAGScheduler( mapOutputTracker.registerShuffle(shuffleDep.get.shuffleId, rdd.partitions.size) } val id = nextStageId.getAndIncrement() - val stage = new Stage(id, rdd, shuffleDep, getParentStages(rdd, priority), priority) + val stage = new Stage(id, rdd, shuffleDep, getParentStages(rdd, priority), priority, callSite) idToStage(id) = stage stageToInfos(stage) = StageInfo(stage) stage @@ -226,13 +254,15 @@ class DAGScheduler( partitions: Seq[Int], callSite: String, allowLocal: Boolean, - resultHandler: (Int, U) => Unit) + resultHandler: (Int, U) => Unit, + properties: Properties = null) : (JobSubmitted, JobWaiter[U]) = { assert(partitions.size > 0) val waiter = new JobWaiter(partitions.size, resultHandler) val func2 = func.asInstanceOf[(TaskContext, Iterator[_]) => _] - val toSubmit = JobSubmitted(finalRdd, func2, partitions.toArray, allowLocal, callSite, waiter) + val toSubmit = JobSubmitted(finalRdd, func2, partitions.toArray, allowLocal, callSite, waiter, + properties) return (toSubmit, waiter) } @@ -242,17 +272,27 @@ class DAGScheduler( partitions: Seq[Int], callSite: String, allowLocal: Boolean, - resultHandler: (Int, U) => Unit) + resultHandler: (Int, U) => Unit, + properties: Properties = null) { if (partitions.size == 0) { return } - val (toSubmit, waiter) = prepareJob( - finalRdd, func, partitions, callSite, allowLocal, resultHandler) + + // Check to make sure we are not launching a task on a partition that does not exist. + val maxPartitions = finalRdd.partitions.length + partitions.find(p => p >= maxPartitions).foreach { p => + throw new IllegalArgumentException( + "Attempting to access a non-existent partition: " + p + ". " + + "Total number of partitions: " + maxPartitions) + } + + val (toSubmit: JobSubmitted, waiter: JobWaiter[_]) = prepareJob( + finalRdd, func, partitions, callSite, allowLocal, resultHandler, properties) eventQueue.put(toSubmit) waiter.awaitResult() match { case JobSucceeded => {} - case JobFailed(exception: Exception) => + case JobFailed(exception: Exception, _) => logInfo("Failed to run " + callSite) throw exception } @@ -263,13 +303,14 @@ class DAGScheduler( func: (TaskContext, Iterator[T]) => U, evaluator: ApproximateEvaluator[U, R], callSite: String, - timeout: Long) + timeout: Long, + properties: Properties = null) : PartialResult[R] = { val listener = new ApproximateActionListener(rdd, func, evaluator, timeout) val func2 = func.asInstanceOf[(TaskContext, Iterator[_]) => _] val partitions = (0 until rdd.partitions.size).toArray - eventQueue.put(JobSubmitted(rdd, func2, partitions, false, callSite, listener)) + eventQueue.put(JobSubmitted(rdd, func2, partitions, false, callSite, listener, properties)) return listener.awaitResult() // Will throw an exception if the job fails } @@ -279,20 +320,22 @@ class DAGScheduler( */ private[scheduler] def processEvent(event: DAGSchedulerEvent): Boolean = { event match { - case JobSubmitted(finalRDD, func, partitions, allowLocal, callSite, listener) => + case JobSubmitted(finalRDD, func, partitions, allowLocal, callSite, listener, properties) => val runId = nextRunId.getAndIncrement() - val finalStage = newStage(finalRDD, None, runId) - val job = new ActiveJob(runId, finalStage, func, partitions, callSite, listener) + val finalStage = newStage(finalRDD, None, runId, Some(callSite)) + val job = new ActiveJob(runId, finalStage, func, partitions, callSite, listener, properties) clearCacheLocs() logInfo("Got job " + job.runId + " (" + callSite + ") with " + partitions.length + " output partitions (allowLocal=" + allowLocal + ")") - logInfo("Final stage: " + finalStage + " (" + finalStage.origin + ")") + logInfo("Final stage: " + finalStage + " (" + finalStage.name + ")") logInfo("Parents of final stage: " + finalStage.parents) logInfo("Missing parents: " + getMissingParentStages(finalStage)) if (allowLocal && finalStage.parents.size == 0 && partitions.length == 1) { // Compute very short actions like first() or take() with no parent stages locally. runLocally(job) } else { + sparkListeners.foreach(_.onJobStart(SparkListenerJobStart(job, properties))) + idToActiveJob(runId) = job activeJobs += job resultStageToJob(finalStage) = job submitStage(finalStage) @@ -304,7 +347,12 @@ class DAGScheduler( case ExecutorLost(execId) => handleExecutorLost(execId) + case begin: BeginEvent => + sparkListeners.foreach(_.onTaskStart(SparkListenerTaskStart(begin.task, begin.taskInfo))) + case completion: CompletionEvent => + sparkListeners.foreach(_.onTaskEnd(SparkListenerTaskEnd(completion.task, + completion.reason, completion.taskInfo, completion.taskMetrics))) handleTaskCompletion(completion) case TaskSetFailed(taskSet, reason) => @@ -315,6 +363,7 @@ class DAGScheduler( for (job <- activeJobs) { val error = new SparkException("Job cancelled because SparkContext was shut down") job.listener.jobFailed(error) + sparkListeners.foreach(_.onJobEnd(SparkListenerJobEnd(job, JobFailed(error, None)))) } return true } @@ -334,7 +383,7 @@ class DAGScheduler( submitStage(stage) } } - + /** * Check for waiting or failed stages which are now eligible for resubmission. * Ordinarily run on every iteration of the event loop. @@ -462,11 +511,26 @@ class DAGScheduler( } } if (tasks.size > 0) { + // Preemptively serialize a task to make sure it can be serialized. We are catching this + // exception here because it would be fairly hard to catch the non-serializable exception + // down the road, where we have several different implementations for local scheduler and + // cluster schedulers. + try { + SparkEnv.get.closureSerializer.newInstance().serialize(tasks.head) + } catch { + case e: NotSerializableException => + abortStage(stage, e.toString) + running -= stage + return + } + + sparkListeners.foreach(_.onStageSubmitted(SparkListenerStageSubmitted(stage, tasks.size))) logInfo("Submitting " + tasks.size + " missing tasks from " + stage + " (" + stage.rdd + ")") myPending ++= tasks logDebug("New pending tasks: " + myPending) + val properties = idToActiveJob(stage.priority).properties taskSched.submitTasks( - new TaskSet(tasks.toArray, stage.id, stage.newAttemptId(), stage.priority)) + new TaskSet(tasks.toArray, stage.id, stage.newAttemptId(), stage.priority, properties)) if (!stage.submissionTime.isDefined) { stage.submissionTime = Some(System.currentTimeMillis()) } @@ -490,7 +554,8 @@ class DAGScheduler( case Some(t) => "%.03f".format((System.currentTimeMillis() - t) / 1000.0) case _ => "Unkown" } - logInfo("%s (%s) finished in %s s".format(stage, stage.origin, serviceTime)) + logInfo("%s (%s) finished in %s s".format(stage, stage.name, serviceTime)) + stage.completionTime = Some(System.currentTimeMillis) val stageComp = StageCompleted(stageToInfos(stage)) sparkListeners.foreach{_.onStageCompleted(stageComp)} running -= stage @@ -512,9 +577,11 @@ class DAGScheduler( job.numFinished += 1 // If the whole job has finished, remove it if (job.numFinished == job.numPartitions) { + idToActiveJob -= stage.priority activeJobs -= job resultStageToJob -= stage markStageAsFinished(stage) + sparkListeners.foreach(_.onJobEnd(SparkListenerJobEnd(job, JobSucceeded))) } job.listener.taskSucceeded(rt.outputId, event.result) } @@ -553,7 +620,7 @@ class DAGScheduler( if (stage.outputLocs.count(_ == Nil) != 0) { // Some tasks had failed; let's resubmit this stage // TODO: Lower-level scheduler should also deal with this - logInfo("Resubmitting " + stage + " (" + stage.origin + + logInfo("Resubmitting " + stage + " (" + stage.name + ") because some of its tasks had failed: " + stage.outputLocs.zipWithIndex.filter(_._1 == Nil).map(_._2).mkString(", ")) submitStage(stage) @@ -585,7 +652,7 @@ class DAGScheduler( running -= failedStage failed += failedStage // TODO: Cancel running tasks in the stage - logInfo("Marking " + failedStage + " (" + failedStage.origin + + logInfo("Marking " + failedStage + " (" + failedStage.name + ") for resubmision due to a fetch failure") // Mark the map whose fetch failed as broken in the map stage val mapStage = shuffleToMapStage(shuffleId) @@ -593,7 +660,7 @@ class DAGScheduler( mapStage.removeOutputLoc(mapId, bmAddress) mapOutputTracker.unregisterMapOutput(shuffleId, mapId, bmAddress) } - logInfo("The failed fetch was from " + mapStage + " (" + mapStage.origin + + logInfo("The failed fetch was from " + mapStage + " (" + mapStage.name + "); marking it for resubmission") failed += mapStage // Remember that a fetch failed now; this is used to resubmit the broken @@ -604,8 +671,11 @@ class DAGScheduler( handleExecutorLost(bmAddress.executorId, Some(task.generation)) } + case ExceptionFailure(className, description, stackTrace, metrics) => + // Do nothing here, left up to the TaskScheduler to decide how to handle user failures + case other => - // Non-fetch failure -- probably a bug in user code; abort all jobs depending on this stage + // Unrecognized failure - abort all jobs depending on this stage abortStage(idToStage(task.stageId), task + " failed: " + other) } } @@ -638,7 +708,7 @@ class DAGScheduler( "(generation " + currentGeneration + ")") } } - + private def handleExecutorGained(execId: String, hostPort: String) { // remove from failedGeneration(execId) ? if (failedGeneration.contains(execId)) { @@ -653,9 +723,13 @@ class DAGScheduler( */ private def abortStage(failedStage: Stage, reason: String) { val dependentStages = resultStageToJob.keys.filter(x => stageDependsOn(x, failedStage)).toSeq + failedStage.completionTime = Some(System.currentTimeMillis()) for (resultStage <- dependentStages) { val job = resultStageToJob(resultStage) - job.listener.jobFailed(new SparkException("Job failed: " + reason)) + val error = new SparkException("Job failed: " + reason) + job.listener.jobFailed(error) + sparkListeners.foreach(_.onJobEnd(SparkListenerJobEnd(job, JobFailed(error, Some(failedStage))))) + idToActiveJob -= resultStage.priority activeJobs -= job resultStageToJob -= resultStage } @@ -728,10 +802,14 @@ class DAGScheduler( sizeBefore = shuffleToMapStage.size shuffleToMapStage.clearOldValues(cleanupTime) logInfo("shuffleToMapStage " + sizeBefore + " --> " + shuffleToMapStage.size) - + sizeBefore = pendingTasks.size pendingTasks.clearOldValues(cleanupTime) logInfo("pendingTasks " + sizeBefore + " --> " + pendingTasks.size) + + sizeBefore = stageToInfos.size + stageToInfos.clearOldValues(cleanupTime) + logInfo("stageToInfos " + sizeBefore + " --> " + stageToInfos.size) } def stop() { diff --git a/core/src/main/scala/spark/scheduler/DAGSchedulerEvent.scala b/core/src/main/scala/spark/scheduler/DAGSchedulerEvent.scala index b46bb863f0..3b4ee6287a 100644 --- a/core/src/main/scala/spark/scheduler/DAGSchedulerEvent.scala +++ b/core/src/main/scala/spark/scheduler/DAGSchedulerEvent.scala @@ -1,5 +1,24 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler +import java.util.Properties + import spark.scheduler.cluster.TaskInfo import scala.collection.mutable.Map @@ -20,9 +39,12 @@ private[spark] case class JobSubmitted( partitions: Array[Int], allowLocal: Boolean, callSite: String, - listener: JobListener) + listener: JobListener, + properties: Properties = null) extends DAGSchedulerEvent +private[spark] case class BeginEvent(task: Task[_], taskInfo: TaskInfo) extends DAGSchedulerEvent + private[spark] case class CompletionEvent( task: Task[_], reason: TaskEndReason, diff --git a/core/src/main/scala/spark/scheduler/DAGSchedulerSource.scala b/core/src/main/scala/spark/scheduler/DAGSchedulerSource.scala new file mode 100644 index 0000000000..87d27cc70d --- /dev/null +++ b/core/src/main/scala/spark/scheduler/DAGSchedulerSource.scala @@ -0,0 +1,30 @@ +package spark.scheduler + +import com.codahale.metrics.{Gauge,MetricRegistry} + +import spark.metrics.source.Source + +private[spark] class DAGSchedulerSource(val dagScheduler: DAGScheduler) extends Source { + val metricRegistry = new MetricRegistry() + val sourceName = "DAGScheduler" + + metricRegistry.register(MetricRegistry.name("stage", "failedStages", "number"), new Gauge[Int] { + override def getValue: Int = dagScheduler.failed.size + }) + + metricRegistry.register(MetricRegistry.name("stage", "runningStages", "number"), new Gauge[Int] { + override def getValue: Int = dagScheduler.running.size + }) + + metricRegistry.register(MetricRegistry.name("stage", "waitingStages", "number"), new Gauge[Int] { + override def getValue: Int = dagScheduler.waiting.size + }) + + metricRegistry.register(MetricRegistry.name("job", "allJobs", "number"), new Gauge[Int] { + override def getValue: Int = dagScheduler.nextRunId.get() + }) + + metricRegistry.register(MetricRegistry.name("job", "activeJobs", "number"), new Gauge[Int] { + override def getValue: Int = dagScheduler.activeJobs.size + }) +} diff --git a/core/src/main/scala/spark/scheduler/InputFormatInfo.scala b/core/src/main/scala/spark/scheduler/InputFormatInfo.scala index 287f731787..65f8c3200e 100644 --- a/core/src/main/scala/spark/scheduler/InputFormatInfo.scala +++ b/core/src/main/scala/spark/scheduler/InputFormatInfo.scala @@ -1,13 +1,32 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import spark.Logging import scala.collection.immutable.Set import org.apache.hadoop.mapred.{FileInputFormat, JobConf} +import org.apache.hadoop.security.UserGroupInformation import org.apache.hadoop.util.ReflectionUtils import org.apache.hadoop.mapreduce.Job import org.apache.hadoop.conf.Configuration import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} import scala.collection.JavaConversions._ +import spark.deploy.SparkHadoopUtil /** @@ -70,6 +89,7 @@ class InputFormatInfo(val configuration: Configuration, val inputFormatClazz: Cl // This method does not expect failures, since validate has already passed ... private def prefLocsFromMapreduceInputFormat(): Set[SplitInfo] = { val conf = new JobConf(configuration) + SparkHadoopUtil.addCredentials(conf); FileInputFormat.setInputPaths(conf, path) val instance: org.apache.hadoop.mapreduce.InputFormat[_, _] = @@ -89,6 +109,7 @@ class InputFormatInfo(val configuration: Configuration, val inputFormatClazz: Cl // This method does not expect failures, since validate has already passed ... private def prefLocsFromMapredInputFormat(): Set[SplitInfo] = { val jobConf = new JobConf(configuration) + SparkHadoopUtil.addCredentials(jobConf); FileInputFormat.setInputPaths(jobConf, path) val instance: org.apache.hadoop.mapred.InputFormat[_, _] = diff --git a/core/src/main/scala/spark/scheduler/JobListener.scala b/core/src/main/scala/spark/scheduler/JobListener.scala index f46b9d551d..af108b8fec 100644 --- a/core/src/main/scala/spark/scheduler/JobListener.scala +++ b/core/src/main/scala/spark/scheduler/JobListener.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler /** diff --git a/core/src/main/scala/spark/scheduler/JobLogger.scala b/core/src/main/scala/spark/scheduler/JobLogger.scala new file mode 100644 index 0000000000..f7565b8c57 --- /dev/null +++ b/core/src/main/scala/spark/scheduler/JobLogger.scala @@ -0,0 +1,337 @@ +/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package spark.scheduler
+
+import java.io.PrintWriter
+import java.io.File
+import java.io.FileNotFoundException
+import java.text.SimpleDateFormat
+import java.util.{Date, Properties}
+import java.util.concurrent.LinkedBlockingQueue
+import scala.collection.mutable.{Map, HashMap, ListBuffer}
+import scala.io.Source
+import spark._
+import spark.executor.TaskMetrics
+import spark.scheduler.cluster.TaskInfo
+
+// Used to record runtime information for each job, including RDD graph
+// tasks' start/stop shuffle information and information from outside
+
+class JobLogger(val logDirName: String) extends SparkListener with Logging {
+ private val logDir =
+ if (System.getenv("SPARK_LOG_DIR") != null)
+ System.getenv("SPARK_LOG_DIR")
+ else
+ "/tmp/spark"
+ private val jobIDToPrintWriter = new HashMap[Int, PrintWriter]
+ private val stageIDToJobID = new HashMap[Int, Int]
+ private val jobIDToStages = new HashMap[Int, ListBuffer[Stage]]
+ private val DATE_FORMAT = new SimpleDateFormat("yyyy/MM/dd HH:mm:ss")
+ private val eventQueue = new LinkedBlockingQueue[SparkListenerEvents]
+
+ createLogDir()
+ def this() = this(String.valueOf(System.currentTimeMillis()))
+
+ def getLogDir = logDir
+ def getJobIDtoPrintWriter = jobIDToPrintWriter
+ def getStageIDToJobID = stageIDToJobID
+ def getJobIDToStages = jobIDToStages
+ def getEventQueue = eventQueue
+
+ new Thread("JobLogger") {
+ setDaemon(true)
+ override def run() {
+ while (true) {
+ val event = eventQueue.take
+ logDebug("Got event of type " + event.getClass.getName)
+ event match {
+ case SparkListenerJobStart(job, properties) =>
+ processJobStartEvent(job, properties)
+ case SparkListenerStageSubmitted(stage, taskSize) =>
+ processStageSubmittedEvent(stage, taskSize)
+ case StageCompleted(stageInfo) =>
+ processStageCompletedEvent(stageInfo)
+ case SparkListenerJobEnd(job, result) =>
+ processJobEndEvent(job, result)
+ case SparkListenerTaskStart(task, taskInfo) =>
+ processTaskStartEvent(task, taskInfo)
+ case SparkListenerTaskEnd(task, reason, taskInfo, taskMetrics) =>
+ processTaskEndEvent(task, reason, taskInfo, taskMetrics)
+ case _ =>
+ }
+ }
+ }
+ }.start()
+
+ // Create a folder for log files, the folder's name is the creation time of the jobLogger
+ protected def createLogDir() {
+ val dir = new File(logDir + "/" + logDirName + "/")
+ if (dir.exists()) {
+ return
+ }
+ if (dir.mkdirs() == false) {
+ logError("create log directory error:" + logDir + "/" + logDirName + "/")
+ }
+ }
+
+ // Create a log file for one job, the file name is the jobID
+ protected def createLogWriter(jobID: Int) {
+ try{
+ val fileWriter = new PrintWriter(logDir + "/" + logDirName + "/" + jobID)
+ jobIDToPrintWriter += (jobID -> fileWriter)
+ } catch {
+ case e: FileNotFoundException => e.printStackTrace()
+ }
+ }
+
+ // Close log file, and clean the stage relationship in stageIDToJobID
+ protected def closeLogWriter(jobID: Int) =
+ jobIDToPrintWriter.get(jobID).foreach { fileWriter =>
+ fileWriter.close()
+ jobIDToStages.get(jobID).foreach(_.foreach{ stage =>
+ stageIDToJobID -= stage.id
+ })
+ jobIDToPrintWriter -= jobID
+ jobIDToStages -= jobID
+ }
+
+ // Write log information to log file, withTime parameter controls whether to recored
+ // time stamp for the information
+ protected def jobLogInfo(jobID: Int, info: String, withTime: Boolean = true) {
+ var writeInfo = info
+ if (withTime) {
+ val date = new Date(System.currentTimeMillis())
+ writeInfo = DATE_FORMAT.format(date) + ": " +info
+ }
+ jobIDToPrintWriter.get(jobID).foreach(_.println(writeInfo))
+ }
+
+ protected def stageLogInfo(stageID: Int, info: String, withTime: Boolean = true) =
+ stageIDToJobID.get(stageID).foreach(jobID => jobLogInfo(jobID, info, withTime))
+
+ protected def buildJobDep(jobID: Int, stage: Stage) {
+ if (stage.priority == jobID) {
+ jobIDToStages.get(jobID) match {
+ case Some(stageList) => stageList += stage
+ case None => val stageList = new ListBuffer[Stage]
+ stageList += stage
+ jobIDToStages += (jobID -> stageList)
+ }
+ stageIDToJobID += (stage.id -> jobID)
+ stage.parents.foreach(buildJobDep(jobID, _))
+ }
+ }
+
+ protected def recordStageDep(jobID: Int) {
+ def getRddsInStage(rdd: RDD[_]): ListBuffer[RDD[_]] = {
+ var rddList = new ListBuffer[RDD[_]]
+ rddList += rdd
+ rdd.dependencies.foreach{ dep => dep match {
+ case shufDep: ShuffleDependency[_,_] =>
+ case _ => rddList ++= getRddsInStage(dep.rdd)
+ }
+ }
+ rddList
+ }
+ jobIDToStages.get(jobID).foreach {_.foreach { stage =>
+ var depRddDesc: String = ""
+ getRddsInStage(stage.rdd).foreach { rdd =>
+ depRddDesc += rdd.id + ","
+ }
+ var depStageDesc: String = ""
+ stage.parents.foreach { stage =>
+ depStageDesc += "(" + stage.id + "," + stage.shuffleDep.get.shuffleId + ")"
+ }
+ jobLogInfo(jobID, "STAGE_ID=" + stage.id + " RDD_DEP=(" +
+ depRddDesc.substring(0, depRddDesc.length - 1) + ")" +
+ " STAGE_DEP=" + depStageDesc, false)
+ }
+ }
+ }
+
+ // Generate indents and convert to String
+ protected def indentString(indent: Int) = {
+ val sb = new StringBuilder()
+ for (i <- 1 to indent) {
+ sb.append(" ")
+ }
+ sb.toString()
+ }
+
+ protected def getRddName(rdd: RDD[_]) = {
+ var rddName = rdd.getClass.getName
+ if (rdd.name != null) {
+ rddName = rdd.name
+ }
+ rddName
+ }
+
+ protected def recordRddInStageGraph(jobID: Int, rdd: RDD[_], indent: Int) {
+ val rddInfo = "RDD_ID=" + rdd.id + "(" + getRddName(rdd) + "," + rdd.generator + ")"
+ jobLogInfo(jobID, indentString(indent) + rddInfo, false)
+ rdd.dependencies.foreach{ dep => dep match {
+ case shufDep: ShuffleDependency[_,_] =>
+ val depInfo = "SHUFFLE_ID=" + shufDep.shuffleId
+ jobLogInfo(jobID, indentString(indent + 1) + depInfo, false)
+ case _ => recordRddInStageGraph(jobID, dep.rdd, indent + 1)
+ }
+ }
+ }
+
+ protected def recordStageDepGraph(jobID: Int, stage: Stage, indent: Int = 0) {
+ var stageInfo: String = ""
+ if (stage.isShuffleMap) {
+ stageInfo = "STAGE_ID=" + stage.id + " MAP_STAGE SHUFFLE_ID=" +
+ stage.shuffleDep.get.shuffleId
+ }else{
+ stageInfo = "STAGE_ID=" + stage.id + " RESULT_STAGE"
+ }
+ if (stage.priority == jobID) {
+ jobLogInfo(jobID, indentString(indent) + stageInfo, false)
+ recordRddInStageGraph(jobID, stage.rdd, indent)
+ stage.parents.foreach(recordStageDepGraph(jobID, _, indent + 2))
+ } else
+ jobLogInfo(jobID, indentString(indent) + stageInfo + " JOB_ID=" + stage.priority, false)
+ }
+
+ // Record task metrics into job log files
+ protected def recordTaskMetrics(stageID: Int, status: String,
+ taskInfo: TaskInfo, taskMetrics: TaskMetrics) {
+ val info = " TID=" + taskInfo.taskId + " STAGE_ID=" + stageID +
+ " START_TIME=" + taskInfo.launchTime + " FINISH_TIME=" + taskInfo.finishTime +
+ " EXECUTOR_ID=" + taskInfo.executorId + " HOST=" + taskMetrics.hostname
+ val executorRunTime = " EXECUTOR_RUN_TIME=" + taskMetrics.executorRunTime
+ val readMetrics =
+ taskMetrics.shuffleReadMetrics match {
+ case Some(metrics) =>
+ " SHUFFLE_FINISH_TIME=" + metrics.shuffleFinishTime +
+ " BLOCK_FETCHED_TOTAL=" + metrics.totalBlocksFetched +
+ " BLOCK_FETCHED_LOCAL=" + metrics.localBlocksFetched +
+ " BLOCK_FETCHED_REMOTE=" + metrics.remoteBlocksFetched +
+ " REMOTE_FETCH_WAIT_TIME=" + metrics.fetchWaitTime +
+ " REMOTE_FETCH_TIME=" + metrics.remoteFetchTime +
+ " REMOTE_BYTES_READ=" + metrics.remoteBytesRead
+ case None => ""
+ }
+ val writeMetrics =
+ taskMetrics.shuffleWriteMetrics match {
+ case Some(metrics) =>
+ " SHUFFLE_BYTES_WRITTEN=" + metrics.shuffleBytesWritten
+ case None => ""
+ }
+ stageLogInfo(stageID, status + info + executorRunTime + readMetrics + writeMetrics)
+ }
+
+ override def onStageSubmitted(stageSubmitted: SparkListenerStageSubmitted) {
+ eventQueue.put(stageSubmitted)
+ }
+
+ protected def processStageSubmittedEvent(stage: Stage, taskSize: Int) {
+ stageLogInfo(stage.id, "STAGE_ID=" + stage.id + " STATUS=SUBMITTED" + " TASK_SIZE=" + taskSize)
+ }
+
+ override def onStageCompleted(stageCompleted: StageCompleted) {
+ eventQueue.put(stageCompleted)
+ }
+
+ protected def processStageCompletedEvent(stageInfo: StageInfo) {
+ stageLogInfo(stageInfo.stage.id, "STAGE_ID=" +
+ stageInfo.stage.id + " STATUS=COMPLETED")
+
+ }
+
+ override def onTaskStart(taskStart: SparkListenerTaskStart) {
+ eventQueue.put(taskStart)
+ }
+
+ protected def processTaskStartEvent(task: Task[_], taskInfo: TaskInfo) {
+ var taskStatus = ""
+ task match {
+ case resultTask: ResultTask[_, _] => taskStatus = "TASK_TYPE=RESULT_TASK"
+ case shuffleMapTask: ShuffleMapTask => taskStatus = "TASK_TYPE=SHUFFLE_MAP_TASK"
+ }
+ }
+
+ override def onTaskEnd(taskEnd: SparkListenerTaskEnd) {
+ eventQueue.put(taskEnd)
+ }
+
+ protected def processTaskEndEvent(task: Task[_], reason: TaskEndReason,
+ taskInfo: TaskInfo, taskMetrics: TaskMetrics) {
+ var taskStatus = ""
+ task match {
+ case resultTask: ResultTask[_, _] => taskStatus = "TASK_TYPE=RESULT_TASK"
+ case shuffleMapTask: ShuffleMapTask => taskStatus = "TASK_TYPE=SHUFFLE_MAP_TASK"
+ }
+ reason match {
+ case Success => taskStatus += " STATUS=SUCCESS"
+ recordTaskMetrics(task.stageId, taskStatus, taskInfo, taskMetrics)
+ case Resubmitted =>
+ taskStatus += " STATUS=RESUBMITTED TID=" + taskInfo.taskId +
+ " STAGE_ID=" + task.stageId
+ stageLogInfo(task.stageId, taskStatus)
+ case FetchFailed(bmAddress, shuffleId, mapId, reduceId) =>
+ taskStatus += " STATUS=FETCHFAILED TID=" + taskInfo.taskId + " STAGE_ID=" +
+ task.stageId + " SHUFFLE_ID=" + shuffleId + " MAP_ID=" +
+ mapId + " REDUCE_ID=" + reduceId
+ stageLogInfo(task.stageId, taskStatus)
+ case OtherFailure(message) =>
+ taskStatus += " STATUS=FAILURE TID=" + taskInfo.taskId +
+ " STAGE_ID=" + task.stageId + " INFO=" + message
+ stageLogInfo(task.stageId, taskStatus)
+ case _ =>
+ }
+ }
+
+ override def onJobEnd(jobEnd: SparkListenerJobEnd) {
+ eventQueue.put(jobEnd)
+ }
+
+ protected def processJobEndEvent(job: ActiveJob, reason: JobResult) {
+ var info = "JOB_ID=" + job.runId
+ reason match {
+ case JobSucceeded => info += " STATUS=SUCCESS"
+ case JobFailed(exception, _) =>
+ info += " STATUS=FAILED REASON="
+ exception.getMessage.split("\\s+").foreach(info += _ + "_")
+ case _ =>
+ }
+ jobLogInfo(job.runId, info.substring(0, info.length - 1).toUpperCase)
+ closeLogWriter(job.runId)
+ }
+
+ protected def recordJobProperties(jobID: Int, properties: Properties) {
+ if(properties != null) {
+ val annotation = properties.getProperty("spark.job.annotation", "")
+ jobLogInfo(jobID, annotation, false)
+ }
+ }
+
+ override def onJobStart(jobStart: SparkListenerJobStart) {
+ eventQueue.put(jobStart)
+ }
+
+ protected def processJobStartEvent(job: ActiveJob, properties: Properties) {
+ createLogWriter(job.runId)
+ recordJobProperties(job.runId, properties)
+ buildJobDep(job.runId, job.finalStage)
+ recordStageDep(job.runId)
+ recordStageDepGraph(job.runId, job.finalStage)
+ jobLogInfo(job.runId, "JOB_ID=" + job.runId + " STATUS=STARTED")
+ }
+}
diff --git a/core/src/main/scala/spark/scheduler/JobResult.scala b/core/src/main/scala/spark/scheduler/JobResult.scala index 654131ee84..a61b335152 100644 --- a/core/src/main/scala/spark/scheduler/JobResult.scala +++ b/core/src/main/scala/spark/scheduler/JobResult.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler /** @@ -6,4 +23,4 @@ package spark.scheduler private[spark] sealed trait JobResult private[spark] case object JobSucceeded extends JobResult -private[spark] case class JobFailed(exception: Exception) extends JobResult +private[spark] case class JobFailed(exception: Exception, failedStage: Option[Stage]) extends JobResult diff --git a/core/src/main/scala/spark/scheduler/JobWaiter.scala b/core/src/main/scala/spark/scheduler/JobWaiter.scala index 3cc6a86345..69cd161c1f 100644 --- a/core/src/main/scala/spark/scheduler/JobWaiter.scala +++ b/core/src/main/scala/spark/scheduler/JobWaiter.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import scala.collection.mutable.ArrayBuffer @@ -35,7 +52,7 @@ private[spark] class JobWaiter[T](totalTasks: Int, resultHandler: (Int, T) => Un throw new UnsupportedOperationException("jobFailed() called on a finished JobWaiter") } jobFinished = true - jobResult = JobFailed(exception) + jobResult = JobFailed(exception, None) this.notifyAll() } } diff --git a/core/src/main/scala/spark/scheduler/MapStatus.scala b/core/src/main/scala/spark/scheduler/MapStatus.scala index 203abb917b..2f6a68ee85 100644 --- a/core/src/main/scala/spark/scheduler/MapStatus.scala +++ b/core/src/main/scala/spark/scheduler/MapStatus.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import spark.storage.BlockManagerId diff --git a/core/src/main/scala/spark/scheduler/ResultTask.scala b/core/src/main/scala/spark/scheduler/ResultTask.scala index 89dc6640b2..361b1e6b91 100644 --- a/core/src/main/scala/spark/scheduler/ResultTask.scala +++ b/core/src/main/scala/spark/scheduler/ResultTask.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import spark._ @@ -70,12 +87,11 @@ private[spark] class ResultTask[T, U]( rdd.partitions(partition) } - // data locality is on a per host basis, not hyper specific to container (host:port). Unique on set of hosts. - val preferredLocs: Seq[String] = if (locs == null) Nil else locs.map(loc => Utils.parseHostPort(loc)._1).toSet.toSeq + private val preferredLocs: Seq[String] = if (locs == null) Nil else locs.toSet.toSeq { // DEBUG code - preferredLocs.foreach (host => Utils.checkHost(host, "preferredLocs : " + preferredLocs)) + preferredLocs.foreach (hostPort => Utils.checkHost(Utils.parseHostPort(hostPort)._1, "preferredLocs : " + preferredLocs)) } override def run(attemptId: Long): U = { diff --git a/core/src/main/scala/spark/scheduler/ShuffleMapTask.scala b/core/src/main/scala/spark/scheduler/ShuffleMapTask.scala index 7dc6da4573..1c25605f75 100644 --- a/core/src/main/scala/spark/scheduler/ShuffleMapTask.scala +++ b/core/src/main/scala/spark/scheduler/ShuffleMapTask.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import java.io._ @@ -13,9 +30,10 @@ import com.ning.compress.lzf.LZFInputStream import com.ning.compress.lzf.LZFOutputStream import spark._ -import executor.ShuffleWriteMetrics +import spark.executor.ShuffleWriteMetrics import spark.storage._ -import util.{TimeStampedHashMap, MetadataCleaner} +import spark.util.{TimeStampedHashMap, MetadataCleaner} + private[spark] object ShuffleMapTask { @@ -84,12 +102,11 @@ private[spark] class ShuffleMapTask( protected def this() = this(0, null, null, 0, null) - // data locality is on a per host basis, not hyper specific to container (host:port). Unique on set of hosts. - private val preferredLocs: Seq[String] = if (locs == null) Nil else locs.map(loc => Utils.parseHostPort(loc)._1).toSet.toSeq + @transient private val preferredLocs: Seq[String] = if (locs == null) Nil else locs.toSet.toSeq { // DEBUG code - preferredLocs.foreach (host => Utils.checkHost(host, "preferredLocs : " + preferredLocs)) + preferredLocs.foreach (hostPort => Utils.checkHost(Utils.parseHostPort(hostPort)._1, "preferredLocs : " + preferredLocs)) } var split = if (rdd == null) { @@ -129,34 +146,52 @@ private[spark] class ShuffleMapTask( val taskContext = new TaskContext(stageId, partition, attemptId) metrics = Some(taskContext.taskMetrics) + + val blockManager = SparkEnv.get.blockManager + var shuffle: ShuffleBlocks = null + var buckets: ShuffleWriterGroup = null + try { - // Partition the map output. - val buckets = Array.fill(numOutputSplits)(new ArrayBuffer[(Any, Any)]) + // Obtain all the block writers for shuffle blocks. + val ser = SparkEnv.get.serializerManager.get(dep.serializerClass) + shuffle = blockManager.shuffleBlockManager.forShuffle(dep.shuffleId, numOutputSplits, ser) + buckets = shuffle.acquireWriters(partition) + + // Write the map output to its associated buckets. for (elem <- rdd.iterator(split, taskContext)) { val pair = elem.asInstanceOf[(Any, Any)] val bucketId = dep.partitioner.getPartition(pair._1) - buckets(bucketId) += pair + buckets.writers(bucketId).write(pair) } - val compressedSizes = new Array[Byte](numOutputSplits) - - var totalBytes = 0l - - val blockManager = SparkEnv.get.blockManager - for (i <- 0 until numOutputSplits) { - val blockId = "shuffle_" + dep.shuffleId + "_" + partition + "_" + i - // Get a Scala iterator from Java map - val iter: Iterator[(Any, Any)] = buckets(i).iterator - val size = blockManager.put(blockId, iter, StorageLevel.DISK_ONLY, false) + // Commit the writes. Get the size of each bucket block (total block size). + var totalBytes = 0L + val compressedSizes: Array[Byte] = buckets.writers.map { writer: BlockObjectWriter => + writer.commit() + writer.close() + val size = writer.size() totalBytes += size - compressedSizes(i) = MapOutputTracker.compressSize(size) + MapOutputTracker.compressSize(size) } + + // Update shuffle metrics. val shuffleMetrics = new ShuffleWriteMetrics shuffleMetrics.shuffleBytesWritten = totalBytes metrics.get.shuffleWriteMetrics = Some(shuffleMetrics) return new MapStatus(blockManager.blockManagerId, compressedSizes) + } catch { case e: Exception => + // If there is an exception from running the task, revert the partial writes + // and throw the exception upstream to Spark. + if (buckets != null) { + buckets.writers.foreach(_.revertPartialWrites()) + } + throw e } finally { + // Release the writers back to the shuffle block manager. + if (shuffle != null && buckets != null) { + shuffle.releaseWriters(buckets) + } // Execute the callbacks on task completion. taskContext.executeOnCompleteCallbacks() } diff --git a/core/src/main/scala/spark/scheduler/SparkListener.scala b/core/src/main/scala/spark/scheduler/SparkListener.scala index a65140b145..4eb7e4e6a5 100644 --- a/core/src/main/scala/spark/scheduler/SparkListener.scala +++ b/core/src/main/scala/spark/scheduler/SparkListener.scala @@ -1,27 +1,83 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler +import java.util.Properties import spark.scheduler.cluster.TaskInfo import spark.util.Distribution -import spark.{Utils, Logging} +import spark.{Logging, SparkContext, TaskEndReason, Utils} import spark.executor.TaskMetrics +sealed trait SparkListenerEvents + +case class SparkListenerStageSubmitted(stage: Stage, taskSize: Int) extends SparkListenerEvents + +case class StageCompleted(val stageInfo: StageInfo) extends SparkListenerEvents + +case class SparkListenerTaskStart(task: Task[_], taskInfo: TaskInfo) extends SparkListenerEvents + +case class SparkListenerTaskEnd(task: Task[_], reason: TaskEndReason, taskInfo: TaskInfo, + taskMetrics: TaskMetrics) extends SparkListenerEvents + +case class SparkListenerJobStart(job: ActiveJob, properties: Properties = null) + extends SparkListenerEvents + +case class SparkListenerJobEnd(job: ActiveJob, jobResult: JobResult) + extends SparkListenerEvents + trait SparkListener { /** - * called when a stage is completed, with information on the completed stage + * Called when a stage is completed, with information on the completed stage */ - def onStageCompleted(stageCompleted: StageCompleted) -} + def onStageCompleted(stageCompleted: StageCompleted) { } + + /** + * Called when a stage is submitted + */ + def onStageSubmitted(stageSubmitted: SparkListenerStageSubmitted) { } -sealed trait SparkListenerEvents + /** + * Called when a task starts + */ + def onTaskStart(taskEnd: SparkListenerTaskStart) { } -case class StageCompleted(val stageInfo: StageInfo) extends SparkListenerEvents + /** + * Called when a task ends + */ + def onTaskEnd(taskEnd: SparkListenerTaskEnd) { } + /** + * Called when a job starts + */ + def onJobStart(jobStart: SparkListenerJobStart) { } + + /** + * Called when a job ends + */ + def onJobEnd(jobEnd: SparkListenerJobEnd) { } + +} /** * Simple SparkListener that logs a few summary statistics when each stage completes */ class StatsReportListener extends SparkListener with Logging { - def onStageCompleted(stageCompleted: StageCompleted) { + override def onStageCompleted(stageCompleted: StageCompleted) { import spark.scheduler.StatsReportListener._ implicit val sc = stageCompleted this.logInfo("Finished stage: " + stageCompleted.stageInfo) @@ -36,6 +92,7 @@ class StatsReportListener extends SparkListener with Logging { showBytesDistribution("task result size:", (_, metric) => Some(metric.resultSize)) //runtime breakdown + val runtimePcts = stageCompleted.stageInfo.taskInfos.map{ case (info, metrics) => RuntimePercentage(info.duration, metrics) } diff --git a/core/src/main/scala/spark/scheduler/SplitInfo.scala b/core/src/main/scala/spark/scheduler/SplitInfo.scala index 6abfb7a1f7..4e3661ec5d 100644 --- a/core/src/main/scala/spark/scheduler/SplitInfo.scala +++ b/core/src/main/scala/spark/scheduler/SplitInfo.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import collection.mutable.ArrayBuffer diff --git a/core/src/main/scala/spark/scheduler/Stage.scala b/core/src/main/scala/spark/scheduler/Stage.scala index 552061e46b..5428daeb94 100644 --- a/core/src/main/scala/spark/scheduler/Stage.scala +++ b/core/src/main/scala/spark/scheduler/Stage.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import java.net.URI @@ -24,9 +41,10 @@ private[spark] class Stage( val rdd: RDD[_], val shuffleDep: Option[ShuffleDependency[_,_]], // Output shuffle if stage is a map stage val parents: List[Stage], - val priority: Int) + val priority: Int, + callSite: Option[String]) extends Logging { - + val isShuffleMap = shuffleDep != None val numPartitions = rdd.partitions.size val outputLocs = Array.fill[List[MapStatus]](numPartitions)(Nil) @@ -34,6 +52,7 @@ private[spark] class Stage( /** When first task was submitted to scheduler. */ var submissionTime: Option[Long] = None + var completionTime: Option[Long] = None private var nextAttemptId = 0 @@ -60,7 +79,7 @@ private[spark] class Stage( numAvailableOutputs -= 1 } } - + def removeOutputsOnExecutor(execId: String) { var becameUnavailable = false for (partition <- 0 until numPartitions) { @@ -84,7 +103,7 @@ private[spark] class Stage( return id } - def origin: String = rdd.origin + val name = callSite.getOrElse(rdd.origin) override def toString = "Stage " + id diff --git a/core/src/main/scala/spark/scheduler/StageInfo.scala b/core/src/main/scala/spark/scheduler/StageInfo.scala index 8d83ff10c4..c4026f995a 100644 --- a/core/src/main/scala/spark/scheduler/StageInfo.scala +++ b/core/src/main/scala/spark/scheduler/StageInfo.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import spark.scheduler.cluster.TaskInfo @@ -9,4 +26,4 @@ case class StageInfo( val taskInfos: mutable.Buffer[(TaskInfo, TaskMetrics)] = mutable.Buffer[(TaskInfo, TaskMetrics)]() ) { override def toString = stage.rdd.toString -}
\ No newline at end of file +} diff --git a/core/src/main/scala/spark/scheduler/Task.scala b/core/src/main/scala/spark/scheduler/Task.scala index a6462c6968..50768d43e0 100644 --- a/core/src/main/scala/spark/scheduler/Task.scala +++ b/core/src/main/scala/spark/scheduler/Task.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import spark.serializer.SerializerInstance diff --git a/core/src/main/scala/spark/scheduler/TaskResult.scala b/core/src/main/scala/spark/scheduler/TaskResult.scala index 6de0aa7adf..dc0621ea7b 100644 --- a/core/src/main/scala/spark/scheduler/TaskResult.scala +++ b/core/src/main/scala/spark/scheduler/TaskResult.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import java.io._ diff --git a/core/src/main/scala/spark/scheduler/TaskScheduler.scala b/core/src/main/scala/spark/scheduler/TaskScheduler.scala index 7787b54762..5188308006 100644 --- a/core/src/main/scala/spark/scheduler/TaskScheduler.scala +++ b/core/src/main/scala/spark/scheduler/TaskScheduler.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler /** diff --git a/core/src/main/scala/spark/scheduler/TaskSchedulerListener.scala b/core/src/main/scala/spark/scheduler/TaskSchedulerListener.scala index b75d3736cf..2cdeb1c8c0 100644 --- a/core/src/main/scala/spark/scheduler/TaskSchedulerListener.scala +++ b/core/src/main/scala/spark/scheduler/TaskSchedulerListener.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import spark.scheduler.cluster.TaskInfo @@ -10,6 +27,9 @@ import spark.executor.TaskMetrics * Interface for getting events back from the TaskScheduler. */ private[spark] trait TaskSchedulerListener { + // A task has started. + def taskStarted(task: Task[_], taskInfo: TaskInfo) + // A task has finished or failed. def taskEnded(task: Task[_], reason: TaskEndReason, result: Any, accumUpdates: Map[Long, Any], taskInfo: TaskInfo, taskMetrics: TaskMetrics): Unit diff --git a/core/src/main/scala/spark/scheduler/TaskSet.scala b/core/src/main/scala/spark/scheduler/TaskSet.scala index a3002ca477..dc3550dd0b 100644 --- a/core/src/main/scala/spark/scheduler/TaskSet.scala +++ b/core/src/main/scala/spark/scheduler/TaskSet.scala @@ -1,11 +1,35 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler +import java.util.Properties + /** * A set of tasks submitted together to the low-level TaskScheduler, usually representing * missing partitions of a particular stage. */ -private[spark] class TaskSet(val tasks: Array[Task[_]], val stageId: Int, val attempt: Int, val priority: Int) { - val id: String = stageId + "." + attempt +private[spark] class TaskSet( + val tasks: Array[Task[_]], + val stageId: Int, + val attempt: Int, + val priority: Int, + val properties: Properties) { + val id: String = stageId + "." + attempt override def toString: String = "TaskSet " + id } diff --git a/core/src/main/scala/spark/scheduler/cluster/ClusterScheduler.scala b/core/src/main/scala/spark/scheduler/cluster/ClusterScheduler.scala index a9d9c5e44c..7c10074dc7 100644 --- a/core/src/main/scala/spark/scheduler/cluster/ClusterScheduler.scala +++ b/core/src/main/scala/spark/scheduler/cluster/ClusterScheduler.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster import java.lang.{Boolean => JBoolean} @@ -32,31 +49,30 @@ private[spark] class ClusterScheduler(val sc: SparkContext) val TASK_REVIVAL_INTERVAL = System.getProperty("spark.tasks.revive.interval", "0").toLong /* - This property controls how aggressive we should be to modulate waiting for host local task scheduling. - To elaborate, currently there is a time limit (3 sec def) to ensure that spark attempts to wait for host locality of tasks before + This property controls how aggressive we should be to modulate waiting for node local task scheduling. + To elaborate, currently there is a time limit (3 sec def) to ensure that spark attempts to wait for node locality of tasks before scheduling on other nodes. We have modified this in yarn branch such that offers to task set happen in prioritized order : - host-local, rack-local and then others - But once all available host local (and no pref) tasks are scheduled, instead of waiting for 3 sec before + node-local, rack-local and then others + But once all available node local (and no pref) tasks are scheduled, instead of waiting for 3 sec before scheduling to other nodes (which degrades performance for time sensitive tasks and on larger clusters), we can modulate that : to also allow rack local nodes or any node. The default is still set to HOST - so that previous behavior is maintained. This is to allow tuning the tension between pulling rdd data off node and scheduling computation asap. TODO: rename property ? The value is one of - - HOST_LOCAL (default, no change w.r.t current behavior), + - NODE_LOCAL (default, no change w.r.t current behavior), - RACK_LOCAL and - ANY Note that this property makes more sense when used in conjugation with spark.tasks.revive.interval > 0 : else it is not very effective. Additional Note: For non trivial clusters, there is a 4x - 5x reduction in running time (in some of our experiments) based on whether - it is left at default HOST_LOCAL, RACK_LOCAL (if cluster is configured to be rack aware) or ANY. + it is left at default NODE_LOCAL, RACK_LOCAL (if cluster is configured to be rack aware) or ANY. If cluster is rack aware, then setting it to RACK_LOCAL gives best tradeoff and a 3x - 4x performance improvement while minimizing IO impact. Also, it brings down the variance in running time drastically. */ - val TASK_SCHEDULING_AGGRESSION = TaskLocality.parse(System.getProperty("spark.tasks.schedule.aggression", "HOST_LOCAL")) + val TASK_SCHEDULING_AGGRESSION = TaskLocality.parse(System.getProperty("spark.tasks.schedule.aggression", "NODE_LOCAL")) val activeTaskSets = new HashMap[String, TaskSetManager] - var activeTaskSetsQueue = new ArrayBuffer[TaskSetManager] val taskIdToTaskSetId = new HashMap[Long, String] val taskIdToExecutorId = new HashMap[Long, String] @@ -73,15 +89,15 @@ private[spark] class ClusterScheduler(val sc: SparkContext) val activeExecutorIds = new HashSet[String] // TODO: We might want to remove this and merge it with execId datastructures - but later. - // Which hosts in the cluster are alive (contains hostPort's) - used for hyper local and local task locality. + // Which hosts in the cluster are alive (contains hostPort's) - used for process local and node local task locality. private val hostPortsAlive = new HashSet[String] private val hostToAliveHostPorts = new HashMap[String, HashSet[String]] // The set of executors we have on each host; this is used to compute hostsAlive, which // in turn is used to decide when we can attain data locality on a given host - val executorsByHostPort = new HashMap[String, HashSet[String]] + private val executorsByHostPort = new HashMap[String, HashSet[String]] - val executorIdToHostPort = new HashMap[String, String] + private val executorIdToHostPort = new HashMap[String, String] // JAR server, if any JARs were added by the user to the SparkContext var jarServer: HttpServer = null @@ -96,14 +112,39 @@ private[spark] class ClusterScheduler(val sc: SparkContext) val mapOutputTracker = SparkEnv.get.mapOutputTracker + var schedulableBuilder: SchedulableBuilder = null + var rootPool: Pool = null + override def setListener(listener: TaskSchedulerListener) { this.listener = listener } def initialize(context: SchedulerBackend) { backend = context + //default scheduler is FIFO + val schedulingMode = System.getProperty("spark.cluster.schedulingmode", "FIFO") + //temporarily set rootPool name to empty + rootPool = new Pool("", SchedulingMode.withName(schedulingMode), 0, 0) + schedulableBuilder = { + schedulingMode match { + case "FIFO" => + new FIFOSchedulableBuilder(rootPool) + case "FAIR" => + new FairSchedulableBuilder(rootPool) + } + } + schedulableBuilder.buildPools() + // resolve executorId to hostPort mapping. + def executorToHostPort(executorId: String, defaultHostPort: String): String = { + executorIdToHostPort.getOrElse(executorId, defaultHostPort) + } + + // Unfortunately, this means that SparkEnv is indirectly referencing ClusterScheduler + // Will that be a design violation ? + SparkEnv.get.executorIdToHostPort = Some(executorToHostPort) } + def newTaskId(): Long = nextTaskId.getAndIncrement() override def start() { @@ -153,9 +194,9 @@ private[spark] class ClusterScheduler(val sc: SparkContext) val tasks = taskSet.tasks logInfo("Adding task set " + taskSet.id + " with " + tasks.length + " tasks") this.synchronized { - val manager = new TaskSetManager(this, taskSet) + val manager = new ClusterTaskSetManager(this, taskSet) activeTaskSets(taskSet.id) = manager - activeTaskSetsQueue += manager + schedulableBuilder.addTaskSetManager(manager, manager.taskSet.properties) taskSetTaskIds(taskSet.id) = new HashSet[Long]() if (hasReceivedTask == false) { @@ -178,7 +219,8 @@ private[spark] class ClusterScheduler(val sc: SparkContext) def taskSetFinished(manager: TaskSetManager) { this.synchronized { activeTaskSets -= manager.taskSet.id - activeTaskSetsQueue -= manager + manager.parent.removeSchedulable(manager) + logInfo("Remove TaskSet %s from pool %s".format(manager.taskSet.id, manager.parent.name)) taskIdToTaskSetId --= taskSetTaskIds(manager.taskSet.id) taskIdToExecutorId --= taskSetTaskIds(manager.taskSet.id) taskSetTaskIds.remove(manager.taskSet.id) @@ -209,14 +251,34 @@ private[spark] class ClusterScheduler(val sc: SparkContext) } // Build a list of tasks to assign to each slave val tasks = offers.map(o => new ArrayBuffer[TaskDescription](o.cores)) + // merge availableCpus into nodeToAvailableCpus block ? val availableCpus = offers.map(o => o.cores).toArray - var launchedTask = false + val nodeToAvailableCpus = { + val map = new HashMap[String, Int]() + for (offer <- offers) { + val hostPort = offer.hostPort + val cores = offer.cores + // DEBUG code + Utils.checkHostPort(hostPort) + + val host = Utils.parseHostPort(hostPort)._1 + map.put(host, map.getOrElse(host, 0) + cores) + } - for (manager <- activeTaskSetsQueue.sortBy(m => (m.taskSet.priority, m.taskSet.stageId))) { + map + } + var launchedTask = false + val sortedTaskSetQueue = rootPool.getSortedTaskSetQueue() + for (manager <- sortedTaskSetQueue) + { + logInfo("parentName:%s,name:%s,runningTasks:%s".format(manager.parent.name, manager.name, manager.runningTasks)) + } + for (manager <- sortedTaskSetQueue) { - // Split offers based on host local, rack local and off-rack tasks. - val hostLocalOffers = new HashMap[String, ArrayBuffer[Int]]() + // Split offers based on node local, rack local and off-rack tasks. + val processLocalOffers = new HashMap[String, ArrayBuffer[Int]]() + val nodeLocalOffers = new HashMap[String, ArrayBuffer[Int]]() val rackLocalOffers = new HashMap[String, ArrayBuffer[Int]]() val otherOffers = new HashMap[String, ArrayBuffer[Int]]() @@ -224,21 +286,30 @@ private[spark] class ClusterScheduler(val sc: SparkContext) val hostPort = offers(i).hostPort // DEBUG code Utils.checkHostPort(hostPort) + + val numProcessLocalTasks = math.max(0, math.min(manager.numPendingTasksForHostPort(hostPort), availableCpus(i))) + if (numProcessLocalTasks > 0){ + val list = processLocalOffers.getOrElseUpdate(hostPort, new ArrayBuffer[Int]) + for (j <- 0 until numProcessLocalTasks) list += i + } + val host = Utils.parseHostPort(hostPort)._1 - val numHostLocalTasks = math.max(0, math.min(manager.numPendingTasksForHost(hostPort), availableCpus(i))) - if (numHostLocalTasks > 0){ - val list = hostLocalOffers.getOrElseUpdate(host, new ArrayBuffer[Int]) - for (j <- 0 until numHostLocalTasks) list += i + val numNodeLocalTasks = math.max(0, + // Remove process local tasks (which are also host local btw !) from this + math.min(manager.numPendingTasksForHost(hostPort) - numProcessLocalTasks, nodeToAvailableCpus(host))) + if (numNodeLocalTasks > 0){ + val list = nodeLocalOffers.getOrElseUpdate(host, new ArrayBuffer[Int]) + for (j <- 0 until numNodeLocalTasks) list += i } val numRackLocalTasks = math.max(0, - // Remove host local tasks (which are also rack local btw !) from this - math.min(manager.numRackLocalPendingTasksForHost(hostPort) - numHostLocalTasks, availableCpus(i))) + // Remove node local tasks (which are also rack local btw !) from this + math.min(manager.numRackLocalPendingTasksForHost(hostPort) - numProcessLocalTasks - numNodeLocalTasks, nodeToAvailableCpus(host))) if (numRackLocalTasks > 0){ val list = rackLocalOffers.getOrElseUpdate(host, new ArrayBuffer[Int]) for (j <- 0 until numRackLocalTasks) list += i } - if (numHostLocalTasks <= 0 && numRackLocalTasks <= 0){ + if (numNodeLocalTasks <= 0 && numRackLocalTasks <= 0){ // add to others list - spread even this across cluster. val list = otherOffers.getOrElseUpdate(host, new ArrayBuffer[Int]) list += i @@ -246,12 +317,19 @@ private[spark] class ClusterScheduler(val sc: SparkContext) } val offersPriorityList = new ArrayBuffer[Int]( - hostLocalOffers.size + rackLocalOffers.size + otherOffers.size) - // First host local, then rack, then others - val numHostLocalOffers = { - val hostLocalPriorityList = ClusterScheduler.prioritizeContainers(hostLocalOffers) - offersPriorityList ++= hostLocalPriorityList - hostLocalPriorityList.size + processLocalOffers.size + nodeLocalOffers.size + rackLocalOffers.size + otherOffers.size) + + // First process local, then host local, then rack, then others + + // numNodeLocalOffers contains count of both process local and host offers. + val numNodeLocalOffers = { + val processLocalPriorityList = ClusterScheduler.prioritizeContainers(processLocalOffers) + offersPriorityList ++= processLocalPriorityList + + val nodeLocalPriorityList = ClusterScheduler.prioritizeContainers(nodeLocalOffers) + offersPriorityList ++= nodeLocalPriorityList + + processLocalPriorityList.size + nodeLocalPriorityList.size } val numRackLocalOffers = { val rackLocalPriorityList = ClusterScheduler.prioritizeContainers(rackLocalOffers) @@ -262,8 +340,8 @@ private[spark] class ClusterScheduler(val sc: SparkContext) var lastLoop = false val lastLoopIndex = TASK_SCHEDULING_AGGRESSION match { - case TaskLocality.HOST_LOCAL => numHostLocalOffers - case TaskLocality.RACK_LOCAL => numRackLocalOffers + numHostLocalOffers + case TaskLocality.NODE_LOCAL => numNodeLocalOffers + case TaskLocality.RACK_LOCAL => numRackLocalOffers + numNodeLocalOffers case TaskLocality.ANY => offersPriorityList.size } @@ -291,10 +369,10 @@ private[spark] class ClusterScheduler(val sc: SparkContext) executorsByHostPort(hostPort) += execId availableCpus(i) -= 1 launchedTask = true - + case None => {} + } } - } // Loop once more - when lastLoop = true, then we try to schedule task on all nodes irrespective of // data locality (we still go in order of priority : but that would not change anything since // if data local tasks had been available, we would have scheduled them already) @@ -302,8 +380,8 @@ private[spark] class ClusterScheduler(val sc: SparkContext) // prevent more looping launchedTask = false } else if (!lastLoop && !launchedTask) { - // Do this only if TASK_SCHEDULING_AGGRESSION != HOST_LOCAL - if (TASK_SCHEDULING_AGGRESSION != TaskLocality.HOST_LOCAL) { + // Do this only if TASK_SCHEDULING_AGGRESSION != NODE_LOCAL + if (TASK_SCHEDULING_AGGRESSION != TaskLocality.NODE_LOCAL) { // fudge launchedTask to ensure we loop once more launchedTask = true // dont loop anymore @@ -312,7 +390,7 @@ private[spark] class ClusterScheduler(val sc: SparkContext) } } while (launchedTask) } - + if (tasks.size > 0) { hasLaunchedTask = true } @@ -365,6 +443,7 @@ private[spark] class ClusterScheduler(val sc: SparkContext) backend.reviveOffers() } if (taskFailed) { + // Also revive offers if a task had failed for some reason other than host lost backend.reviveOffers() } @@ -411,9 +490,7 @@ private[spark] class ClusterScheduler(val sc: SparkContext) def checkSpeculatableTasks() { var shouldRevive = false synchronized { - for (ts <- activeTaskSetsQueue) { - shouldRevive |= ts.checkSpeculatableTasks() - } + shouldRevive = rootPool.checkSpeculatableTasks() } if (shouldRevive) { backend.reviveOffers() @@ -423,7 +500,7 @@ private[spark] class ClusterScheduler(val sc: SparkContext) // Check for pending tasks in all our active jobs. def hasPendingTasks(): Boolean = { synchronized { - activeTaskSetsQueue.exists( _.hasPendingTasks() ) + rootPool.hasPendingTasks() } } @@ -462,14 +539,14 @@ private[spark] class ClusterScheduler(val sc: SparkContext) hostPortsAlive -= hostPort hostToAliveHostPorts.getOrElseUpdate(Utils.parseHostPort(hostPort)._1, new HashSet[String]).remove(hostPort) } - + val execs = executorsByHostPort.getOrElse(hostPort, new HashSet) execs -= executorId if (execs.isEmpty) { executorsByHostPort -= hostPort } executorIdToHostPort -= executorId - activeTaskSetsQueue.foreach(_.executorLost(executorId, hostPort)) + rootPool.executorLost(executorId, hostPort) } def executorGained(execId: String, hostPort: String) { @@ -477,6 +554,8 @@ private[spark] class ClusterScheduler(val sc: SparkContext) } def getExecutorsAliveOnHost(host: String): Option[Set[String]] = { + Utils.checkHost(host) + val retval = hostToAliveHostPorts.get(host) if (retval.isDefined) { return Some(retval.get.toSet) @@ -485,6 +564,13 @@ private[spark] class ClusterScheduler(val sc: SparkContext) None } + def isExecutorAliveOnHostPort(hostPort: String): Boolean = { + // Even if hostPort is a host, it does not matter - it is just a specific check. + // But we do have to ensure that only hostPort get into hostPortsAlive ! + // So no check against Utils.checkHostPort + hostPortsAlive.contains(hostPort) + } + // By default, rack is unknown def getRackForHost(value: String): Option[String] = None diff --git a/core/src/main/scala/spark/scheduler/cluster/ClusterTaskSetManager.scala b/core/src/main/scala/spark/scheduler/cluster/ClusterTaskSetManager.scala new file mode 100644 index 0000000000..860a38e9f8 --- /dev/null +++ b/core/src/main/scala/spark/scheduler/cluster/ClusterTaskSetManager.scala @@ -0,0 +1,803 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler.cluster + +import java.nio.ByteBuffer +import java.util.{Arrays, NoSuchElementException} + +import scala.collection.mutable.ArrayBuffer +import scala.collection.mutable.HashMap +import scala.collection.mutable.HashSet +import scala.math.max +import scala.math.min + +import spark.{FetchFailed, Logging, Resubmitted, SparkEnv, Success, TaskEndReason, TaskState, Utils} +import spark.{ExceptionFailure, SparkException, TaskResultTooBigFailure} +import spark.TaskState.TaskState +import spark.scheduler.{ShuffleMapTask, Task, TaskResult, TaskSet} + + +private[spark] object TaskLocality + extends Enumeration("PROCESS_LOCAL", "NODE_LOCAL", "RACK_LOCAL", "ANY") with Logging { + + // process local is expected to be used ONLY within tasksetmanager for now. + val PROCESS_LOCAL, NODE_LOCAL, RACK_LOCAL, ANY = Value + + type TaskLocality = Value + + def isAllowed(constraint: TaskLocality, condition: TaskLocality): Boolean = { + + // Must not be the constraint. + assert (constraint != TaskLocality.PROCESS_LOCAL) + + constraint match { + case TaskLocality.NODE_LOCAL => + condition == TaskLocality.NODE_LOCAL + case TaskLocality.RACK_LOCAL => + condition == TaskLocality.NODE_LOCAL || condition == TaskLocality.RACK_LOCAL + // For anything else, allow + case _ => true + } + } + + def parse(str: String): TaskLocality = { + // better way to do this ? + try { + val retval = TaskLocality.withName(str) + // Must not specify PROCESS_LOCAL ! + assert (retval != TaskLocality.PROCESS_LOCAL) + retval + } catch { + case nEx: NoSuchElementException => { + logWarning("Invalid task locality specified '" + str + "', defaulting to NODE_LOCAL") + // default to preserve earlier behavior + NODE_LOCAL + } + } + } +} + +/** + * Schedules the tasks within a single TaskSet in the ClusterScheduler. + */ +private[spark] class ClusterTaskSetManager(sched: ClusterScheduler, val taskSet: TaskSet) + extends TaskSetManager with Logging { + + // Maximum time to wait to run a task in a preferred location (in ms) + val LOCALITY_WAIT = System.getProperty("spark.locality.wait", "3000").toLong + + // CPUs to request per task + val CPUS_PER_TASK = System.getProperty("spark.task.cpus", "1").toDouble + + // Maximum times a task is allowed to fail before failing the job + val MAX_TASK_FAILURES = 4 + + // Quantile of tasks at which to start speculation + val SPECULATION_QUANTILE = System.getProperty("spark.speculation.quantile", "0.75").toDouble + val SPECULATION_MULTIPLIER = System.getProperty("spark.speculation.multiplier", "1.5").toDouble + + // Serializer for closures and tasks. + val ser = SparkEnv.get.closureSerializer.newInstance() + + val tasks = taskSet.tasks + val numTasks = tasks.length + val copiesRunning = new Array[Int](numTasks) + val finished = new Array[Boolean](numTasks) + val numFailures = new Array[Int](numTasks) + val taskAttempts = Array.fill[List[TaskInfo]](numTasks)(Nil) + var tasksFinished = 0 + + var weight = 1 + var minShare = 0 + var runningTasks = 0 + var priority = taskSet.priority + var stageId = taskSet.stageId + var name = "TaskSet_" + taskSet.stageId.toString + var parent: Schedulable = null + + // Last time when we launched a preferred task (for delay scheduling) + var lastPreferredLaunchTime = System.currentTimeMillis + + // List of pending tasks for each node (process local to container). + // These collections are actually + // treated as stacks, in which new tasks are added to the end of the + // ArrayBuffer and removed from the end. This makes it faster to detect + // tasks that repeatedly fail because whenever a task failed, it is put + // back at the head of the stack. They are also only cleaned up lazily; + // when a task is launched, it remains in all the pending lists except + // the one that it was launched from, but gets removed from them later. + private val pendingTasksForHostPort = new HashMap[String, ArrayBuffer[Int]] + + // List of pending tasks for each node. + // Essentially, similar to pendingTasksForHostPort, except at host level + private val pendingTasksForHost = new HashMap[String, ArrayBuffer[Int]] + + // List of pending tasks for each node based on rack locality. + // Essentially, similar to pendingTasksForHost, except at rack level + private val pendingRackLocalTasksForHost = new HashMap[String, ArrayBuffer[Int]] + + // List containing pending tasks with no locality preferences + val pendingTasksWithNoPrefs = new ArrayBuffer[Int] + + // List containing all pending tasks (also used as a stack, as above) + val allPendingTasks = new ArrayBuffer[Int] + + // Tasks that can be speculated. Since these will be a small fraction of total + // tasks, we'll just hold them in a HashSet. + val speculatableTasks = new HashSet[Int] + + // Task index, start and finish time for each task attempt (indexed by task ID) + val taskInfos = new HashMap[Long, TaskInfo] + + // Did the job fail? + var failed = false + var causeOfFailure = "" + + // How frequently to reprint duplicate exceptions in full, in milliseconds + val EXCEPTION_PRINT_INTERVAL = + System.getProperty("spark.logging.exceptionPrintInterval", "10000").toLong + // Map of recent exceptions (identified by string representation and + // top stack frame) to duplicate count (how many times the same + // exception has appeared) and time the full exception was + // printed. This should ideally be an LRU map that can drop old + // exceptions automatically. + val recentExceptions = HashMap[String, (Int, Long)]() + + // Figure out the current map output tracker generation and set it on all tasks + val generation = sched.mapOutputTracker.getGeneration + logDebug("Generation for " + taskSet.id + ": " + generation) + for (t <- tasks) { + t.generation = generation + } + + // Add all our tasks to the pending lists. We do this in reverse order + // of task index so that tasks with low indices get launched first. + for (i <- (0 until numTasks).reverse) { + addPendingTask(i) + } + + // Note that it follows the hierarchy. + // if we search for NODE_LOCAL, the output will include PROCESS_LOCAL and + // if we search for RACK_LOCAL, it will include PROCESS_LOCAL & NODE_LOCAL + private def findPreferredLocations( + _taskPreferredLocations: Seq[String], + scheduler: ClusterScheduler, + taskLocality: TaskLocality.TaskLocality): HashSet[String] = + { + if (TaskLocality.PROCESS_LOCAL == taskLocality) { + // straight forward comparison ! Special case it. + val retval = new HashSet[String]() + scheduler.synchronized { + for (location <- _taskPreferredLocations) { + if (scheduler.isExecutorAliveOnHostPort(location)) { + retval += location + } + } + } + + return retval + } + + val taskPreferredLocations = { + if (TaskLocality.NODE_LOCAL == taskLocality) { + _taskPreferredLocations + } else { + assert (TaskLocality.RACK_LOCAL == taskLocality) + // Expand set to include all 'seen' rack local hosts. + // This works since container allocation/management happens within master - + // so any rack locality information is updated in msater. + // Best case effort, and maybe sort of kludge for now ... rework it later ? + val hosts = new HashSet[String] + _taskPreferredLocations.foreach(h => { + val rackOpt = scheduler.getRackForHost(h) + if (rackOpt.isDefined) { + val hostsOpt = scheduler.getCachedHostsForRack(rackOpt.get) + if (hostsOpt.isDefined) { + hosts ++= hostsOpt.get + } + } + + // Ensure that irrespective of what scheduler says, host is always added ! + hosts += h + }) + + hosts + } + } + + val retval = new HashSet[String] + scheduler.synchronized { + for (prefLocation <- taskPreferredLocations) { + val aliveLocationsOpt = scheduler.getExecutorsAliveOnHost(Utils.parseHostPort(prefLocation)._1) + if (aliveLocationsOpt.isDefined) { + retval ++= aliveLocationsOpt.get + } + } + } + + retval + } + + // Add a task to all the pending-task lists that it should be on. + private def addPendingTask(index: Int) { + // We can infer hostLocalLocations from rackLocalLocations by joining it against + // tasks(index).preferredLocations (with appropriate hostPort <-> host conversion). + // But not doing it for simplicity sake. If this becomes a performance issue, modify it. + val locs = tasks(index).preferredLocations + val processLocalLocations = findPreferredLocations(locs, sched, TaskLocality.PROCESS_LOCAL) + val hostLocalLocations = findPreferredLocations(locs, sched, TaskLocality.NODE_LOCAL) + val rackLocalLocations = findPreferredLocations(locs, sched, TaskLocality.RACK_LOCAL) + + if (rackLocalLocations.size == 0) { + // Current impl ensures this. + assert (processLocalLocations.size == 0) + assert (hostLocalLocations.size == 0) + pendingTasksWithNoPrefs += index + } else { + + // process local locality + for (hostPort <- processLocalLocations) { + // DEBUG Code + Utils.checkHostPort(hostPort) + + val hostPortList = pendingTasksForHostPort.getOrElseUpdate(hostPort, ArrayBuffer()) + hostPortList += index + } + + // host locality (includes process local) + for (hostPort <- hostLocalLocations) { + // DEBUG Code + Utils.checkHostPort(hostPort) + + val host = Utils.parseHostPort(hostPort)._1 + val hostList = pendingTasksForHost.getOrElseUpdate(host, ArrayBuffer()) + hostList += index + } + + // rack locality (includes process local and host local) + for (rackLocalHostPort <- rackLocalLocations) { + // DEBUG Code + Utils.checkHostPort(rackLocalHostPort) + + val rackLocalHost = Utils.parseHostPort(rackLocalHostPort)._1 + val list = pendingRackLocalTasksForHost.getOrElseUpdate(rackLocalHost, ArrayBuffer()) + list += index + } + } + + allPendingTasks += index + } + + // Return the pending tasks list for a given host port (process local), or an empty list if + // there is no map entry for that host + private def getPendingTasksForHostPort(hostPort: String): ArrayBuffer[Int] = { + // DEBUG Code + Utils.checkHostPort(hostPort) + pendingTasksForHostPort.getOrElse(hostPort, ArrayBuffer()) + } + + // Return the pending tasks list for a given host, or an empty list if + // there is no map entry for that host + private def getPendingTasksForHost(hostPort: String): ArrayBuffer[Int] = { + val host = Utils.parseHostPort(hostPort)._1 + pendingTasksForHost.getOrElse(host, ArrayBuffer()) + } + + // Return the pending tasks (rack level) list for a given host, or an empty list if + // there is no map entry for that host + private def getRackLocalPendingTasksForHost(hostPort: String): ArrayBuffer[Int] = { + val host = Utils.parseHostPort(hostPort)._1 + pendingRackLocalTasksForHost.getOrElse(host, ArrayBuffer()) + } + + // Number of pending tasks for a given host Port (which would be process local) + override def numPendingTasksForHostPort(hostPort: String): Int = { + getPendingTasksForHostPort(hostPort).count { index => + copiesRunning(index) == 0 && !finished(index) + } + } + + // Number of pending tasks for a given host (which would be data local) + override def numPendingTasksForHost(hostPort: String): Int = { + getPendingTasksForHost(hostPort).count { index => + copiesRunning(index) == 0 && !finished(index) + } + } + + // Number of pending rack local tasks for a given host + override def numRackLocalPendingTasksForHost(hostPort: String): Int = { + getRackLocalPendingTasksForHost(hostPort).count { index => + copiesRunning(index) == 0 && !finished(index) + } + } + + + // Dequeue a pending task from the given list and return its index. + // Return None if the list is empty. + // This method also cleans up any tasks in the list that have already + // been launched, since we want that to happen lazily. + private def findTaskFromList(list: ArrayBuffer[Int]): Option[Int] = { + while (!list.isEmpty) { + val index = list.last + list.trimEnd(1) + if (copiesRunning(index) == 0 && !finished(index)) { + return Some(index) + } + } + return None + } + + // Return a speculative task for a given host if any are available. The task should not have an + // attempt running on this host, in case the host is slow. In addition, if locality is set, the + // task must have a preference for this host/rack/no preferred locations at all. + private def findSpeculativeTask(hostPort: String, locality: TaskLocality.TaskLocality): Option[Int] = { + + assert (TaskLocality.isAllowed(locality, TaskLocality.NODE_LOCAL)) + speculatableTasks.retain(index => !finished(index)) // Remove finished tasks from set + + if (speculatableTasks.size > 0) { + val localTask = speculatableTasks.find { index => + val locations = findPreferredLocations(tasks(index).preferredLocations, sched, + TaskLocality.NODE_LOCAL) + val attemptLocs = taskAttempts(index).map(_.hostPort) + (locations.size == 0 || locations.contains(hostPort)) && !attemptLocs.contains(hostPort) + } + + if (localTask != None) { + speculatableTasks -= localTask.get + return localTask + } + + // check for rack locality + if (TaskLocality.isAllowed(locality, TaskLocality.RACK_LOCAL)) { + val rackTask = speculatableTasks.find { index => + val locations = findPreferredLocations(tasks(index).preferredLocations, sched, + TaskLocality.RACK_LOCAL) + val attemptLocs = taskAttempts(index).map(_.hostPort) + locations.contains(hostPort) && !attemptLocs.contains(hostPort) + } + + if (rackTask != None) { + speculatableTasks -= rackTask.get + return rackTask + } + } + + // Any task ... + if (TaskLocality.isAllowed(locality, TaskLocality.ANY)) { + // Check for attemptLocs also ? + val nonLocalTask = speculatableTasks.find { i => + !taskAttempts(i).map(_.hostPort).contains(hostPort) + } + if (nonLocalTask != None) { + speculatableTasks -= nonLocalTask.get + return nonLocalTask + } + } + } + return None + } + + // Dequeue a pending task for a given node and return its index. + // If localOnly is set to false, allow non-local tasks as well. + private def findTask(hostPort: String, locality: TaskLocality.TaskLocality): Option[Int] = { + val processLocalTask = findTaskFromList(getPendingTasksForHostPort(hostPort)) + if (processLocalTask != None) { + return processLocalTask + } + + val localTask = findTaskFromList(getPendingTasksForHost(hostPort)) + if (localTask != None) { + return localTask + } + + if (TaskLocality.isAllowed(locality, TaskLocality.RACK_LOCAL)) { + val rackLocalTask = findTaskFromList(getRackLocalPendingTasksForHost(hostPort)) + if (rackLocalTask != None) { + return rackLocalTask + } + } + + // Look for no pref tasks AFTER rack local tasks - this has side effect that we will get to + // failed tasks later rather than sooner. + // TODO: That code path needs to be revisited (adding to no prefs list when host:port goes down). + val noPrefTask = findTaskFromList(pendingTasksWithNoPrefs) + if (noPrefTask != None) { + return noPrefTask + } + + if (TaskLocality.isAllowed(locality, TaskLocality.ANY)) { + val nonLocalTask = findTaskFromList(allPendingTasks) + if (nonLocalTask != None) { + return nonLocalTask + } + } + + // Finally, if all else has failed, find a speculative task + return findSpeculativeTask(hostPort, locality) + } + + private def isProcessLocalLocation(task: Task[_], hostPort: String): Boolean = { + Utils.checkHostPort(hostPort) + + val locs = task.preferredLocations + + locs.contains(hostPort) + } + + private def isHostLocalLocation(task: Task[_], hostPort: String): Boolean = { + val locs = task.preferredLocations + + // If no preference, consider it as host local + if (locs.isEmpty) return true + + val host = Utils.parseHostPort(hostPort)._1 + locs.find(h => Utils.parseHostPort(h)._1 == host).isDefined + } + + // Does a host count as a rack local preferred location for a task? + // (assumes host is NOT preferred location). + // This is true if either the task has preferred locations and this host is one, or it has + // no preferred locations (in which we still count the launch as preferred). + private def isRackLocalLocation(task: Task[_], hostPort: String): Boolean = { + + val locs = task.preferredLocations + + val preferredRacks = new HashSet[String]() + for (preferredHost <- locs) { + val rack = sched.getRackForHost(preferredHost) + if (None != rack) preferredRacks += rack.get + } + + if (preferredRacks.isEmpty) return false + + val hostRack = sched.getRackForHost(hostPort) + + return None != hostRack && preferredRacks.contains(hostRack.get) + } + + // Respond to an offer of a single slave from the scheduler by finding a task + override def slaveOffer( + execId: String, + hostPort: String, + availableCpus: Double, + overrideLocality: TaskLocality.TaskLocality = null): Option[TaskDescription] = + { + if (tasksFinished < numTasks && availableCpus >= CPUS_PER_TASK) { + // If explicitly specified, use that + val locality = if (overrideLocality != null) overrideLocality else { + // expand only if we have waited for more than LOCALITY_WAIT for a host local task ... + val time = System.currentTimeMillis + if (time - lastPreferredLaunchTime < LOCALITY_WAIT) { + TaskLocality.NODE_LOCAL + } else { + TaskLocality.ANY + } + } + + findTask(hostPort, locality) match { + case Some(index) => { + // Found a task; do some bookkeeping and return a Mesos task for it + val task = tasks(index) + val taskId = sched.newTaskId() + // Figure out whether this should count as a preferred launch + val taskLocality = + if (isProcessLocalLocation(task, hostPort)) TaskLocality.PROCESS_LOCAL + else if (isHostLocalLocation(task, hostPort)) TaskLocality.NODE_LOCAL + else if (isRackLocalLocation(task, hostPort)) TaskLocality.RACK_LOCAL + else TaskLocality.ANY + val prefStr = taskLocality.toString + logInfo("Starting task %s:%d as TID %s on slave %s: %s (%s)".format( + taskSet.id, index, taskId, execId, hostPort, prefStr)) + // Do various bookkeeping + copiesRunning(index) += 1 + val time = System.currentTimeMillis + val info = new TaskInfo(taskId, index, time, execId, hostPort, taskLocality) + taskInfos(taskId) = info + taskAttempts(index) = info :: taskAttempts(index) + if (taskLocality == TaskLocality.PROCESS_LOCAL || taskLocality == TaskLocality.NODE_LOCAL) { + lastPreferredLaunchTime = time + } + // Serialize and return the task + val startTime = System.currentTimeMillis + // We rely on the DAGScheduler to catch non-serializable closures and RDDs, so in here + // we assume the task can be serialized without exceptions. + val serializedTask = Task.serializeWithDependencies( + task, sched.sc.addedFiles, sched.sc.addedJars, ser) + val timeTaken = System.currentTimeMillis - startTime + increaseRunningTasks(1) + logInfo("Serialized task %s:%d as %d bytes in %d ms".format( + taskSet.id, index, serializedTask.limit, timeTaken)) + val taskName = "task %s:%d".format(taskSet.id, index) + if (taskAttempts(index).size == 1) + taskStarted(task,info) + return Some(new TaskDescription(taskId, execId, taskName, serializedTask)) + } + case _ => + } + } + return None + } + + override def statusUpdate(tid: Long, state: TaskState, serializedData: ByteBuffer) { + state match { + case TaskState.FINISHED => + taskFinished(tid, state, serializedData) + case TaskState.LOST => + taskLost(tid, state, serializedData) + case TaskState.FAILED => + taskLost(tid, state, serializedData) + case TaskState.KILLED => + taskLost(tid, state, serializedData) + case _ => + } + } + + def taskStarted(task: Task[_], info: TaskInfo) { + sched.listener.taskStarted(task, info) + } + + def taskFinished(tid: Long, state: TaskState, serializedData: ByteBuffer) { + val info = taskInfos(tid) + if (info.failed) { + // We might get two task-lost messages for the same task in coarse-grained Mesos mode, + // or even from Mesos itself when acks get delayed. + return + } + val index = info.index + info.markSuccessful() + decreaseRunningTasks(1) + if (!finished(index)) { + tasksFinished += 1 + logInfo("Finished TID %s in %d ms (progress: %d/%d)".format( + tid, info.duration, tasksFinished, numTasks)) + // Deserialize task result and pass it to the scheduler + try { + val result = ser.deserialize[TaskResult[_]](serializedData) + result.metrics.resultSize = serializedData.limit() + sched.listener.taskEnded( + tasks(index), Success, result.value, result.accumUpdates, info, result.metrics) + } catch { + case cnf: ClassNotFoundException => + val loader = Thread.currentThread().getContextClassLoader + throw new SparkException("ClassNotFound with classloader: " + loader, cnf) + case ex => throw ex + } + // Mark finished and stop if we've finished all the tasks + finished(index) = true + if (tasksFinished == numTasks) { + sched.taskSetFinished(this) + } + } else { + logInfo("Ignoring task-finished event for TID " + tid + + " because task " + index + " is already finished") + } + } + + def taskLost(tid: Long, state: TaskState, serializedData: ByteBuffer) { + val info = taskInfos(tid) + if (info.failed) { + // We might get two task-lost messages for the same task in coarse-grained Mesos mode, + // or even from Mesos itself when acks get delayed. + return + } + val index = info.index + info.markFailed() + decreaseRunningTasks(1) + if (!finished(index)) { + logInfo("Lost TID %s (task %s:%d)".format(tid, taskSet.id, index)) + copiesRunning(index) -= 1 + // Check if the problem is a map output fetch failure. In that case, this + // task will never succeed on any node, so tell the scheduler about it. + if (serializedData != null && serializedData.limit() > 0) { + val reason = ser.deserialize[TaskEndReason](serializedData, getClass.getClassLoader) + reason match { + case fetchFailed: FetchFailed => + logInfo("Loss was due to fetch failure from " + fetchFailed.bmAddress) + sched.listener.taskEnded(tasks(index), fetchFailed, null, null, info, null) + finished(index) = true + tasksFinished += 1 + sched.taskSetFinished(this) + decreaseRunningTasks(runningTasks) + return + + case taskResultTooBig: TaskResultTooBigFailure => + logInfo("Loss was due to task %s result exceeding Akka frame size; aborting job".format( + tid)) + abort("Task %s result exceeded Akka frame size".format(tid)) + return + + case ef: ExceptionFailure => + sched.listener.taskEnded(tasks(index), ef, null, null, info, ef.metrics.getOrElse(null)) + val key = ef.description + val now = System.currentTimeMillis + val (printFull, dupCount) = { + if (recentExceptions.contains(key)) { + val (dupCount, printTime) = recentExceptions(key) + if (now - printTime > EXCEPTION_PRINT_INTERVAL) { + recentExceptions(key) = (0, now) + (true, 0) + } else { + recentExceptions(key) = (dupCount + 1, printTime) + (false, dupCount + 1) + } + } else { + recentExceptions(key) = (0, now) + (true, 0) + } + } + if (printFull) { + val locs = ef.stackTrace.map(loc => "\tat %s".format(loc.toString)) + logInfo("Loss was due to %s\n%s\n%s".format( + ef.className, ef.description, locs.mkString("\n"))) + } else { + logInfo("Loss was due to %s [duplicate %d]".format(ef.description, dupCount)) + } + + case _ => {} + } + } + // On non-fetch failures, re-enqueue the task as pending for a max number of retries + addPendingTask(index) + // Count failed attempts only on FAILED and LOST state (not on KILLED) + if (state == TaskState.FAILED || state == TaskState.LOST) { + numFailures(index) += 1 + if (numFailures(index) > MAX_TASK_FAILURES) { + logError("Task %s:%d failed more than %d times; aborting job".format( + taskSet.id, index, MAX_TASK_FAILURES)) + abort("Task %s:%d failed more than %d times".format(taskSet.id, index, MAX_TASK_FAILURES)) + } + } + } else { + logInfo("Ignoring task-lost event for TID " + tid + + " because task " + index + " is already finished") + } + } + + override def error(message: String) { + // Save the error message + abort("Error: " + message) + } + + def abort(message: String) { + failed = true + causeOfFailure = message + // TODO: Kill running tasks if we were not terminated due to a Mesos error + sched.listener.taskSetFailed(taskSet, message) + decreaseRunningTasks(runningTasks) + sched.taskSetFinished(this) + } + + override def increaseRunningTasks(taskNum: Int) { + runningTasks += taskNum + if (parent != null) { + parent.increaseRunningTasks(taskNum) + } + } + + override def decreaseRunningTasks(taskNum: Int) { + runningTasks -= taskNum + if (parent != null) { + parent.decreaseRunningTasks(taskNum) + } + } + + // TODO: for now we just find Pool not TaskSetManager, + // we can extend this function in future if needed + override def getSchedulableByName(name: String): Schedulable = { + return null + } + + override def addSchedulable(schedulable:Schedulable) { + //nothing + } + + override def removeSchedulable(schedulable:Schedulable) { + //nothing + } + + override def getSortedTaskSetQueue(): ArrayBuffer[TaskSetManager] = { + var sortedTaskSetQueue = new ArrayBuffer[TaskSetManager] + sortedTaskSetQueue += this + return sortedTaskSetQueue + } + + override def executorLost(execId: String, hostPort: String) { + logInfo("Re-queueing tasks for " + execId + " from TaskSet " + taskSet.id) + + // If some task has preferred locations only on hostname, and there are no more executors there, + // put it in the no-prefs list to avoid the wait from delay scheduling + + // host local tasks - should we push this to rack local or no pref list ? For now, preserving + // behavior and moving to no prefs list. Note, this was done due to impliations related to + // 'waiting' for data local tasks, etc. + // Note: NOT checking process local list - since host local list is super set of that. We need + // to ad to no prefs only if there is no host local node for the task (not if there is no + // process local node for the task) + for (index <- getPendingTasksForHost(Utils.parseHostPort(hostPort)._1)) { + val newLocs = findPreferredLocations( + tasks(index).preferredLocations, sched, TaskLocality.NODE_LOCAL) + if (newLocs.isEmpty) { + pendingTasksWithNoPrefs += index + } + } + + // Re-enqueue any tasks that ran on the failed executor if this is a shuffle map stage + if (tasks(0).isInstanceOf[ShuffleMapTask]) { + for ((tid, info) <- taskInfos if info.executorId == execId) { + val index = taskInfos(tid).index + if (finished(index)) { + finished(index) = false + copiesRunning(index) -= 1 + tasksFinished -= 1 + addPendingTask(index) + // Tell the DAGScheduler that this task was resubmitted so that it doesn't think our + // stage finishes when a total of tasks.size tasks finish. + sched.listener.taskEnded(tasks(index), Resubmitted, null, null, info, null) + } + } + } + // Also re-enqueue any tasks that were running on the node + for ((tid, info) <- taskInfos if info.running && info.executorId == execId) { + taskLost(tid, TaskState.KILLED, null) + } + } + + /** + * Check for tasks to be speculated and return true if there are any. This is called periodically + * by the ClusterScheduler. + * + * TODO: To make this scale to large jobs, we need to maintain a list of running tasks, so that + * we don't scan the whole task set. It might also help to make this sorted by launch time. + */ + override def checkSpeculatableTasks(): Boolean = { + // Can't speculate if we only have one task, or if all tasks have finished. + if (numTasks == 1 || tasksFinished == numTasks) { + return false + } + var foundTasks = false + val minFinishedForSpeculation = (SPECULATION_QUANTILE * numTasks).floor.toInt + logDebug("Checking for speculative tasks: minFinished = " + minFinishedForSpeculation) + if (tasksFinished >= minFinishedForSpeculation) { + val time = System.currentTimeMillis() + val durations = taskInfos.values.filter(_.successful).map(_.duration).toArray + Arrays.sort(durations) + val medianDuration = durations(min((0.5 * numTasks).round.toInt, durations.size - 1)) + val threshold = max(SPECULATION_MULTIPLIER * medianDuration, 100) + // TODO: Threshold should also look at standard deviation of task durations and have a lower + // bound based on that. + logDebug("Task length threshold for speculation: " + threshold) + for ((tid, info) <- taskInfos) { + val index = info.index + if (!finished(index) && copiesRunning(index) == 1 && info.timeRunning(time) > threshold && + !speculatableTasks.contains(index)) { + logInfo( + "Marking task %s:%d (on %s) as speculatable because it ran more than %.0f ms".format( + taskSet.id, index, info.hostPort, threshold)) + speculatableTasks += index + foundTasks = true + } + } + } + return foundTasks + } + + override def hasPendingTasks(): Boolean = { + numTasks > 0 && tasksFinished < numTasks + } +} diff --git a/core/src/main/scala/spark/scheduler/cluster/ExecutorLossReason.scala b/core/src/main/scala/spark/scheduler/cluster/ExecutorLossReason.scala index 8bf838209f..8825f2dd24 100644 --- a/core/src/main/scala/spark/scheduler/cluster/ExecutorLossReason.scala +++ b/core/src/main/scala/spark/scheduler/cluster/ExecutorLossReason.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster import spark.executor.ExecutorExitCode diff --git a/core/src/main/scala/spark/scheduler/cluster/Pool.scala b/core/src/main/scala/spark/scheduler/cluster/Pool.scala new file mode 100644 index 0000000000..83708f07e1 --- /dev/null +++ b/core/src/main/scala/spark/scheduler/cluster/Pool.scala @@ -0,0 +1,121 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler.cluster + +import scala.collection.mutable.ArrayBuffer +import scala.collection.mutable.HashMap + +import spark.Logging +import spark.scheduler.cluster.SchedulingMode.SchedulingMode + +/** + * An Schedulable entity that represent collection of Pools or TaskSetManagers + */ + +private[spark] class Pool( + val poolName: String, + val schedulingMode: SchedulingMode, + initMinShare: Int, + initWeight: Int) + extends Schedulable + with Logging { + + var schedulableQueue = new ArrayBuffer[Schedulable] + var schedulableNameToSchedulable = new HashMap[String, Schedulable] + + var weight = initWeight + var minShare = initMinShare + var runningTasks = 0 + + var priority = 0 + var stageId = 0 + var name = poolName + var parent:Schedulable = null + + var taskSetSchedulingAlgorithm: SchedulingAlgorithm = { + schedulingMode match { + case SchedulingMode.FAIR => + new FairSchedulingAlgorithm() + case SchedulingMode.FIFO => + new FIFOSchedulingAlgorithm() + } + } + + override def addSchedulable(schedulable: Schedulable) { + schedulableQueue += schedulable + schedulableNameToSchedulable(schedulable.name) = schedulable + schedulable.parent= this + } + + override def removeSchedulable(schedulable: Schedulable) { + schedulableQueue -= schedulable + schedulableNameToSchedulable -= schedulable.name + } + + override def getSchedulableByName(schedulableName: String): Schedulable = { + if (schedulableNameToSchedulable.contains(schedulableName)) { + return schedulableNameToSchedulable(schedulableName) + } + for (schedulable <- schedulableQueue) { + var sched = schedulable.getSchedulableByName(schedulableName) + if (sched != null) { + return sched + } + } + return null + } + + override def executorLost(executorId: String, host: String) { + schedulableQueue.foreach(_.executorLost(executorId, host)) + } + + override def checkSpeculatableTasks(): Boolean = { + var shouldRevive = false + for (schedulable <- schedulableQueue) { + shouldRevive |= schedulable.checkSpeculatableTasks() + } + return shouldRevive + } + + override def getSortedTaskSetQueue(): ArrayBuffer[TaskSetManager] = { + var sortedTaskSetQueue = new ArrayBuffer[TaskSetManager] + val sortedSchedulableQueue = schedulableQueue.sortWith(taskSetSchedulingAlgorithm.comparator) + for (schedulable <- sortedSchedulableQueue) { + sortedTaskSetQueue ++= schedulable.getSortedTaskSetQueue() + } + return sortedTaskSetQueue + } + + override def increaseRunningTasks(taskNum: Int) { + runningTasks += taskNum + if (parent != null) { + parent.increaseRunningTasks(taskNum) + } + } + + override def decreaseRunningTasks(taskNum: Int) { + runningTasks -= taskNum + if (parent != null) { + parent.decreaseRunningTasks(taskNum) + } + } + + override def hasPendingTasks(): Boolean = { + schedulableQueue.exists(_.hasPendingTasks()) + } +} diff --git a/core/src/main/scala/spark/scheduler/cluster/Schedulable.scala b/core/src/main/scala/spark/scheduler/cluster/Schedulable.scala new file mode 100644 index 0000000000..f557b142c4 --- /dev/null +++ b/core/src/main/scala/spark/scheduler/cluster/Schedulable.scala @@ -0,0 +1,44 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler.cluster + +import scala.collection.mutable.ArrayBuffer + +/** + * An interface for schedulable entities. + * there are two type of Schedulable entities(Pools and TaskSetManagers) + */ +private[spark] trait Schedulable { + var parent: Schedulable + def weight: Int + def minShare: Int + def runningTasks: Int + def priority: Int + def stageId: Int + def name: String + + def increaseRunningTasks(taskNum: Int): Unit + def decreaseRunningTasks(taskNum: Int): Unit + def addSchedulable(schedulable: Schedulable): Unit + def removeSchedulable(schedulable: Schedulable): Unit + def getSchedulableByName(name: String): Schedulable + def executorLost(executorId: String, host: String): Unit + def checkSpeculatableTasks(): Boolean + def getSortedTaskSetQueue(): ArrayBuffer[TaskSetManager] + def hasPendingTasks(): Boolean +} diff --git a/core/src/main/scala/spark/scheduler/cluster/SchedulableBuilder.scala b/core/src/main/scala/spark/scheduler/cluster/SchedulableBuilder.scala new file mode 100644 index 0000000000..95554023c0 --- /dev/null +++ b/core/src/main/scala/spark/scheduler/cluster/SchedulableBuilder.scala @@ -0,0 +1,132 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler.cluster + +import java.io.{File, FileInputStream, FileOutputStream} + +import scala.collection.mutable.ArrayBuffer +import scala.collection.mutable.ArrayBuffer +import scala.collection.mutable.HashMap +import scala.collection.mutable.HashSet +import scala.util.control.Breaks._ +import scala.xml._ + +import spark.Logging +import spark.scheduler.cluster.SchedulingMode.SchedulingMode + +import java.util.Properties + +/** + * An interface to build Schedulable tree + * buildPools: build the tree nodes(pools) + * addTaskSetManager: build the leaf nodes(TaskSetManagers) + */ +private[spark] trait SchedulableBuilder { + def buildPools() + def addTaskSetManager(manager: Schedulable, properties: Properties) +} + +private[spark] class FIFOSchedulableBuilder(val rootPool: Pool) extends SchedulableBuilder with Logging { + + override def buildPools() { + //nothing + } + + override def addTaskSetManager(manager: Schedulable, properties: Properties) { + rootPool.addSchedulable(manager) + } +} + +private[spark] class FairSchedulableBuilder(val rootPool: Pool) extends SchedulableBuilder with Logging { + + val schedulerAllocFile = System.getProperty("spark.fairscheduler.allocation.file","unspecified") + val FAIR_SCHEDULER_PROPERTIES = "spark.scheduler.cluster.fair.pool" + val DEFAULT_POOL_NAME = "default" + val MINIMUM_SHARES_PROPERTY = "minShare" + val SCHEDULING_MODE_PROPERTY = "schedulingMode" + val WEIGHT_PROPERTY = "weight" + val POOL_NAME_PROPERTY = "@name" + val POOLS_PROPERTY = "pool" + val DEFAULT_SCHEDULING_MODE = SchedulingMode.FIFO + val DEFAULT_MINIMUM_SHARE = 2 + val DEFAULT_WEIGHT = 1 + + override def buildPools() { + val file = new File(schedulerAllocFile) + if (file.exists()) { + val xml = XML.loadFile(file) + for (poolNode <- (xml \\ POOLS_PROPERTY)) { + + val poolName = (poolNode \ POOL_NAME_PROPERTY).text + var schedulingMode = DEFAULT_SCHEDULING_MODE + var minShare = DEFAULT_MINIMUM_SHARE + var weight = DEFAULT_WEIGHT + + val xmlSchedulingMode = (poolNode \ SCHEDULING_MODE_PROPERTY).text + if (xmlSchedulingMode != "") { + try { + schedulingMode = SchedulingMode.withName(xmlSchedulingMode) + } catch { + case e: Exception => logInfo("Error xml schedulingMode, using default schedulingMode") + } + } + + val xmlMinShare = (poolNode \ MINIMUM_SHARES_PROPERTY).text + if (xmlMinShare != "") { + minShare = xmlMinShare.toInt + } + + val xmlWeight = (poolNode \ WEIGHT_PROPERTY).text + if (xmlWeight != "") { + weight = xmlWeight.toInt + } + + val pool = new Pool(poolName, schedulingMode, minShare, weight) + rootPool.addSchedulable(pool) + logInfo("Create new pool with name:%s,schedulingMode:%s,minShare:%d,weight:%d".format( + poolName, schedulingMode, minShare, weight)) + } + } + + //finally create "default" pool + if (rootPool.getSchedulableByName(DEFAULT_POOL_NAME) == null) { + val pool = new Pool(DEFAULT_POOL_NAME, DEFAULT_SCHEDULING_MODE, DEFAULT_MINIMUM_SHARE, DEFAULT_WEIGHT) + rootPool.addSchedulable(pool) + logInfo("Create default pool with name:%s,schedulingMode:%s,minShare:%d,weight:%d".format( + DEFAULT_POOL_NAME, DEFAULT_SCHEDULING_MODE, DEFAULT_MINIMUM_SHARE, DEFAULT_WEIGHT)) + } + } + + override def addTaskSetManager(manager: Schedulable, properties: Properties) { + var poolName = DEFAULT_POOL_NAME + var parentPool = rootPool.getSchedulableByName(poolName) + if (properties != null) { + poolName = properties.getProperty(FAIR_SCHEDULER_PROPERTIES, DEFAULT_POOL_NAME) + parentPool = rootPool.getSchedulableByName(poolName) + if (parentPool == null) { + //we will create a new pool that user has configured in app instead of being defined in xml file + parentPool = new Pool(poolName,DEFAULT_SCHEDULING_MODE, DEFAULT_MINIMUM_SHARE, DEFAULT_WEIGHT) + rootPool.addSchedulable(parentPool) + logInfo("Create pool with name:%s,schedulingMode:%s,minShare:%d,weight:%d".format( + poolName, DEFAULT_SCHEDULING_MODE, DEFAULT_MINIMUM_SHARE, DEFAULT_WEIGHT)) + } + } + parentPool.addSchedulable(manager) + logInfo("Added task set " + manager.name + " tasks to pool "+poolName) + } +} diff --git a/core/src/main/scala/spark/scheduler/cluster/SchedulerBackend.scala b/core/src/main/scala/spark/scheduler/cluster/SchedulerBackend.scala index 9ac875de3a..4431744ec3 100644 --- a/core/src/main/scala/spark/scheduler/cluster/SchedulerBackend.scala +++ b/core/src/main/scala/spark/scheduler/cluster/SchedulerBackend.scala @@ -1,6 +1,23 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster -import spark.Utils +import spark.{SparkContext, Utils} /** * A backend interface for cluster scheduling systems that allows plugging in different ones under @@ -14,14 +31,7 @@ private[spark] trait SchedulerBackend { def defaultParallelism(): Int // Memory used by each executor (in megabytes) - protected val executorMemory = { - // TODO: Might need to add some extra memory for the non-heap parts of the JVM - Option(System.getProperty("spark.executor.memory")) - .orElse(Option(System.getenv("SPARK_MEM"))) - .map(Utils.memoryStringToMb) - .getOrElse(512) - } - + protected val executorMemory: Int = SparkContext.executorMemoryRequested // TODO: Probably want to add a killTask too } diff --git a/core/src/main/scala/spark/scheduler/cluster/SchedulingAlgorithm.scala b/core/src/main/scala/spark/scheduler/cluster/SchedulingAlgorithm.scala new file mode 100644 index 0000000000..69e0ac2a6b --- /dev/null +++ b/core/src/main/scala/spark/scheduler/cluster/SchedulingAlgorithm.scala @@ -0,0 +1,81 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler.cluster + +/** + * An interface for sort algorithm + * FIFO: FIFO algorithm between TaskSetManagers + * FS: FS algorithm between Pools, and FIFO or FS within Pools + */ +private[spark] trait SchedulingAlgorithm { + def comparator(s1: Schedulable, s2: Schedulable): Boolean +} + +private[spark] class FIFOSchedulingAlgorithm extends SchedulingAlgorithm { + override def comparator(s1: Schedulable, s2: Schedulable): Boolean = { + val priority1 = s1.priority + val priority2 = s2.priority + var res = math.signum(priority1 - priority2) + if (res == 0) { + val stageId1 = s1.stageId + val stageId2 = s2.stageId + res = math.signum(stageId1 - stageId2) + } + if (res < 0) { + return true + } else { + return false + } + } +} + +private[spark] class FairSchedulingAlgorithm extends SchedulingAlgorithm { + override def comparator(s1: Schedulable, s2: Schedulable): Boolean = { + val minShare1 = s1.minShare + val minShare2 = s2.minShare + val runningTasks1 = s1.runningTasks + val runningTasks2 = s2.runningTasks + val s1Needy = runningTasks1 < minShare1 + val s2Needy = runningTasks2 < minShare2 + val minShareRatio1 = runningTasks1.toDouble / math.max(minShare1, 1.0).toDouble + val minShareRatio2 = runningTasks2.toDouble / math.max(minShare2, 1.0).toDouble + val taskToWeightRatio1 = runningTasks1.toDouble / s1.weight.toDouble + val taskToWeightRatio2 = runningTasks2.toDouble / s2.weight.toDouble + var res:Boolean = true + var compare:Int = 0 + + if (s1Needy && !s2Needy) { + return true + } else if (!s1Needy && s2Needy) { + return false + } else if (s1Needy && s2Needy) { + compare = minShareRatio1.compareTo(minShareRatio2) + } else { + compare = taskToWeightRatio1.compareTo(taskToWeightRatio2) + } + + if (compare < 0) { + return true + } else if (compare > 0) { + return false + } else { + return s1.name < s2.name + } + } +} + diff --git a/core/src/main/scala/spark/scheduler/cluster/SchedulingMode.scala b/core/src/main/scala/spark/scheduler/cluster/SchedulingMode.scala new file mode 100644 index 0000000000..4b3e3e50e1 --- /dev/null +++ b/core/src/main/scala/spark/scheduler/cluster/SchedulingMode.scala @@ -0,0 +1,24 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler.cluster + +object SchedulingMode extends Enumeration("FAIR","FIFO"){ + + type SchedulingMode = Value + val FAIR,FIFO = Value +} diff --git a/core/src/main/scala/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala b/core/src/main/scala/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala index 0b8922d139..55d6c0a47e 100644 --- a/core/src/main/scala/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala +++ b/core/src/main/scala/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster import spark.{Utils, Logging, SparkContext} @@ -31,7 +48,8 @@ private[spark] class SparkDeploySchedulerBackend( val command = Command("spark.executor.StandaloneExecutorBackend", args, sc.executorEnvs) val sparkHome = sc.getSparkHome().getOrElse( throw new IllegalArgumentException("must supply spark home for spark standalone")) - val appDesc = new ApplicationDescription(appName, maxCores, executorMemory, command, sparkHome) + val appDesc = new ApplicationDescription(appName, maxCores, executorMemory, command, sparkHome, + sc.ui.appUIAddress) client = new Client(sc.env.actorSystem, master, appDesc, this) client.start() diff --git a/core/src/main/scala/spark/scheduler/cluster/StandaloneClusterMessage.scala b/core/src/main/scala/spark/scheduler/cluster/StandaloneClusterMessage.scala index 3335294844..ac9e5ef94d 100644 --- a/core/src/main/scala/spark/scheduler/cluster/StandaloneClusterMessage.scala +++ b/core/src/main/scala/spark/scheduler/cluster/StandaloneClusterMessage.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster import spark.TaskState.TaskState diff --git a/core/src/main/scala/spark/scheduler/cluster/StandaloneSchedulerBackend.scala b/core/src/main/scala/spark/scheduler/cluster/StandaloneSchedulerBackend.scala index 004592a540..03a64e0192 100644 --- a/core/src/main/scala/spark/scheduler/cluster/StandaloneSchedulerBackend.scala +++ b/core/src/main/scala/spark/scheduler/cluster/StandaloneSchedulerBackend.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} diff --git a/core/src/main/scala/spark/scheduler/cluster/TaskDescription.scala b/core/src/main/scala/spark/scheduler/cluster/TaskDescription.scala index b41e951be9..761fdf6919 100644 --- a/core/src/main/scala/spark/scheduler/cluster/TaskDescription.scala +++ b/core/src/main/scala/spark/scheduler/cluster/TaskDescription.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/scheduler/cluster/TaskInfo.scala b/core/src/main/scala/spark/scheduler/cluster/TaskInfo.scala index 718f26bfbd..c693b722ac 100644 --- a/core/src/main/scala/spark/scheduler/cluster/TaskInfo.scala +++ b/core/src/main/scala/spark/scheduler/cluster/TaskInfo.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster import spark.Utils @@ -34,6 +51,17 @@ class TaskInfo( def running: Boolean = !finished + def status: String = { + if (running) + "RUNNING" + else if (failed) + "FAILED" + else if (successful) + "SUCCESS" + else + "UNKNOWN" + } + def duration: Long = { if (!finished) { throw new UnsupportedOperationException("duration() called on unfinished tasks") diff --git a/core/src/main/scala/spark/scheduler/cluster/TaskSetManager.scala b/core/src/main/scala/spark/scheduler/cluster/TaskSetManager.scala index 27e713e2c4..7978a5df74 100644 --- a/core/src/main/scala/spark/scheduler/cluster/TaskSetManager.scala +++ b/core/src/main/scala/spark/scheduler/cluster/TaskSetManager.scala @@ -1,627 +1,44 @@ -package spark.scheduler.cluster - -import java.util.{HashMap => JHashMap, NoSuchElementException, Arrays} - -import scala.collection.mutable.ArrayBuffer -import scala.collection.mutable.HashMap -import scala.collection.mutable.HashSet -import scala.math.max -import scala.math.min - -import spark._ -import spark.scheduler._ -import spark.TaskState.TaskState -import java.nio.ByteBuffer - -private[spark] object TaskLocality extends Enumeration("HOST_LOCAL", "RACK_LOCAL", "ANY") with Logging { - - val HOST_LOCAL, RACK_LOCAL, ANY = Value - - type TaskLocality = Value - - def isAllowed(constraint: TaskLocality, condition: TaskLocality): Boolean = { - - constraint match { - case TaskLocality.HOST_LOCAL => condition == TaskLocality.HOST_LOCAL - case TaskLocality.RACK_LOCAL => condition == TaskLocality.HOST_LOCAL || condition == TaskLocality.RACK_LOCAL - // For anything else, allow - case _ => true - } - } - - def parse(str: String): TaskLocality = { - // better way to do this ? - try { - TaskLocality.withName(str) - } catch { - case nEx: NoSuchElementException => { - logWarning("Invalid task locality specified '" + str + "', defaulting to HOST_LOCAL"); - // default to preserve earlier behavior - HOST_LOCAL - } - } - } -} - -/** - * Schedules the tasks within a single TaskSet in the ClusterScheduler. +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. */ -private[spark] class TaskSetManager(sched: ClusterScheduler, val taskSet: TaskSet) extends Logging { - - // Maximum time to wait to run a task in a preferred location (in ms) - val LOCALITY_WAIT = System.getProperty("spark.locality.wait", "3000").toLong - - // CPUs to request per task - val CPUS_PER_TASK = System.getProperty("spark.task.cpus", "1").toDouble - - // Maximum times a task is allowed to fail before failing the job - val MAX_TASK_FAILURES = 4 - - // Quantile of tasks at which to start speculation - val SPECULATION_QUANTILE = System.getProperty("spark.speculation.quantile", "0.75").toDouble - val SPECULATION_MULTIPLIER = System.getProperty("spark.speculation.multiplier", "1.5").toDouble - - // Serializer for closures and tasks. - val ser = SparkEnv.get.closureSerializer.newInstance() - - val priority = taskSet.priority - val tasks = taskSet.tasks - val numTasks = tasks.length - val copiesRunning = new Array[Int](numTasks) - val finished = new Array[Boolean](numTasks) - val numFailures = new Array[Int](numTasks) - val taskAttempts = Array.fill[List[TaskInfo]](numTasks)(Nil) - var tasksFinished = 0 - - // Last time when we launched a preferred task (for delay scheduling) - var lastPreferredLaunchTime = System.currentTimeMillis - - // List of pending tasks for each node (hyper local to container). These collections are actually - // treated as stacks, in which new tasks are added to the end of the - // ArrayBuffer and removed from the end. This makes it faster to detect - // tasks that repeatedly fail because whenever a task failed, it is put - // back at the head of the stack. They are also only cleaned up lazily; - // when a task is launched, it remains in all the pending lists except - // the one that it was launched from, but gets removed from them later. - private val pendingTasksForHostPort = new HashMap[String, ArrayBuffer[Int]] - - // List of pending tasks for each node. - // Essentially, similar to pendingTasksForHostPort, except at host level - private val pendingTasksForHost = new HashMap[String, ArrayBuffer[Int]] - - // List of pending tasks for each node based on rack locality. - // Essentially, similar to pendingTasksForHost, except at rack level - private val pendingRackLocalTasksForHost = new HashMap[String, ArrayBuffer[Int]] - - // List containing pending tasks with no locality preferences - val pendingTasksWithNoPrefs = new ArrayBuffer[Int] - - // List containing all pending tasks (also used as a stack, as above) - val allPendingTasks = new ArrayBuffer[Int] - - // Tasks that can be speculated. Since these will be a small fraction of total - // tasks, we'll just hold them in a HashSet. - val speculatableTasks = new HashSet[Int] - - // Task index, start and finish time for each task attempt (indexed by task ID) - val taskInfos = new HashMap[Long, TaskInfo] - - // Did the job fail? - var failed = false - var causeOfFailure = "" - - // How frequently to reprint duplicate exceptions in full, in milliseconds - val EXCEPTION_PRINT_INTERVAL = - System.getProperty("spark.logging.exceptionPrintInterval", "10000").toLong - // Map of recent exceptions (identified by string representation and - // top stack frame) to duplicate count (how many times the same - // exception has appeared) and time the full exception was - // printed. This should ideally be an LRU map that can drop old - // exceptions automatically. - val recentExceptions = HashMap[String, (Int, Long)]() - - // Figure out the current map output tracker generation and set it on all tasks - val generation = sched.mapOutputTracker.getGeneration - logDebug("Generation for " + taskSet.id + ": " + generation) - for (t <- tasks) { - t.generation = generation - } - - // Add all our tasks to the pending lists. We do this in reverse order - // of task index so that tasks with low indices get launched first. - for (i <- (0 until numTasks).reverse) { - addPendingTask(i) - } - - private def findPreferredLocations(_taskPreferredLocations: Seq[String], scheduler: ClusterScheduler, rackLocal: Boolean = false): ArrayBuffer[String] = { - // DEBUG code - _taskPreferredLocations.foreach(h => Utils.checkHost(h, "taskPreferredLocation " + _taskPreferredLocations)) - - val taskPreferredLocations = if (! rackLocal) _taskPreferredLocations else { - // Expand set to include all 'seen' rack local hosts. - // This works since container allocation/management happens within master - so any rack locality information is updated in msater. - // Best case effort, and maybe sort of kludge for now ... rework it later ? - val hosts = new HashSet[String] - _taskPreferredLocations.foreach(h => { - val rackOpt = scheduler.getRackForHost(h) - if (rackOpt.isDefined) { - val hostsOpt = scheduler.getCachedHostsForRack(rackOpt.get) - if (hostsOpt.isDefined) { - hosts ++= hostsOpt.get - } - } - - // Ensure that irrespective of what scheduler says, host is always added ! - hosts += h - }) - - hosts - } - - val retval = new ArrayBuffer[String] - scheduler.synchronized { - for (prefLocation <- taskPreferredLocations) { - val aliveLocationsOpt = scheduler.getExecutorsAliveOnHost(prefLocation) - if (aliveLocationsOpt.isDefined) { - retval ++= aliveLocationsOpt.get - } - } - } - - retval - } - - // Add a task to all the pending-task lists that it should be on. - private def addPendingTask(index: Int) { - // We can infer hostLocalLocations from rackLocalLocations by joining it against tasks(index).preferredLocations (with appropriate - // hostPort <-> host conversion). But not doing it for simplicity sake. If this becomes a performance issue, modify it. - val hostLocalLocations = findPreferredLocations(tasks(index).preferredLocations, sched) - val rackLocalLocations = findPreferredLocations(tasks(index).preferredLocations, sched, true) - - if (rackLocalLocations.size == 0) { - // Current impl ensures this. - assert (hostLocalLocations.size == 0) - pendingTasksWithNoPrefs += index - } else { - // host locality - for (hostPort <- hostLocalLocations) { - // DEBUG Code - Utils.checkHostPort(hostPort) - - val hostPortList = pendingTasksForHostPort.getOrElseUpdate(hostPort, ArrayBuffer()) - hostPortList += index - - val host = Utils.parseHostPort(hostPort)._1 - val hostList = pendingTasksForHost.getOrElseUpdate(host, ArrayBuffer()) - hostList += index - } - - // rack locality - for (rackLocalHostPort <- rackLocalLocations) { - // DEBUG Code - Utils.checkHostPort(rackLocalHostPort) - - val rackLocalHost = Utils.parseHostPort(rackLocalHostPort)._1 - val list = pendingRackLocalTasksForHost.getOrElseUpdate(rackLocalHost, ArrayBuffer()) - list += index - } - } - - allPendingTasks += index - } - - // Return the pending tasks list for a given host port (hyper local), or an empty list if - // there is no map entry for that host - private def getPendingTasksForHostPort(hostPort: String): ArrayBuffer[Int] = { - // DEBUG Code - Utils.checkHostPort(hostPort) - pendingTasksForHostPort.getOrElse(hostPort, ArrayBuffer()) - } - - // Return the pending tasks list for a given host, or an empty list if - // there is no map entry for that host - private def getPendingTasksForHost(hostPort: String): ArrayBuffer[Int] = { - val host = Utils.parseHostPort(hostPort)._1 - pendingTasksForHost.getOrElse(host, ArrayBuffer()) - } - - // Return the pending tasks (rack level) list for a given host, or an empty list if - // there is no map entry for that host - private def getRackLocalPendingTasksForHost(hostPort: String): ArrayBuffer[Int] = { - val host = Utils.parseHostPort(hostPort)._1 - pendingRackLocalTasksForHost.getOrElse(host, ArrayBuffer()) - } - - // Number of pending tasks for a given host (which would be data local) - def numPendingTasksForHost(hostPort: String): Int = { - getPendingTasksForHost(hostPort).count( index => copiesRunning(index) == 0 && !finished(index) ) - } - - // Number of pending rack local tasks for a given host - def numRackLocalPendingTasksForHost(hostPort: String): Int = { - getRackLocalPendingTasksForHost(hostPort).count( index => copiesRunning(index) == 0 && !finished(index) ) - } - - - // Dequeue a pending task from the given list and return its index. - // Return None if the list is empty. - // This method also cleans up any tasks in the list that have already - // been launched, since we want that to happen lazily. - private def findTaskFromList(list: ArrayBuffer[Int]): Option[Int] = { - while (!list.isEmpty) { - val index = list.last - list.trimEnd(1) - if (copiesRunning(index) == 0 && !finished(index)) { - return Some(index) - } - } - return None - } - - // Return a speculative task for a given host if any are available. The task should not have an - // attempt running on this host, in case the host is slow. In addition, if locality is set, the - // task must have a preference for this host/rack/no preferred locations at all. - private def findSpeculativeTask(hostPort: String, locality: TaskLocality.TaskLocality): Option[Int] = { - - assert (TaskLocality.isAllowed(locality, TaskLocality.HOST_LOCAL)) - speculatableTasks.retain(index => !finished(index)) // Remove finished tasks from set - - if (speculatableTasks.size > 0) { - val localTask = speculatableTasks.find { - index => - val locations = findPreferredLocations(tasks(index).preferredLocations, sched) - val attemptLocs = taskAttempts(index).map(_.hostPort) - (locations.size == 0 || locations.contains(hostPort)) && !attemptLocs.contains(hostPort) - } - - if (localTask != None) { - speculatableTasks -= localTask.get - return localTask - } - - // check for rack locality - if (TaskLocality.isAllowed(locality, TaskLocality.RACK_LOCAL)) { - val rackTask = speculatableTasks.find { - index => - val locations = findPreferredLocations(tasks(index).preferredLocations, sched, true) - val attemptLocs = taskAttempts(index).map(_.hostPort) - locations.contains(hostPort) && !attemptLocs.contains(hostPort) - } - - if (rackTask != None) { - speculatableTasks -= rackTask.get - return rackTask - } - } - - // Any task ... - if (TaskLocality.isAllowed(locality, TaskLocality.ANY)) { - // Check for attemptLocs also ? - val nonLocalTask = speculatableTasks.find(i => !taskAttempts(i).map(_.hostPort).contains(hostPort)) - if (nonLocalTask != None) { - speculatableTasks -= nonLocalTask.get - return nonLocalTask - } - } - } - return None - } - - // Dequeue a pending task for a given node and return its index. - // If localOnly is set to false, allow non-local tasks as well. - private def findTask(hostPort: String, locality: TaskLocality.TaskLocality): Option[Int] = { - val localTask = findTaskFromList(getPendingTasksForHost(hostPort)) - if (localTask != None) { - return localTask - } - - if (TaskLocality.isAllowed(locality, TaskLocality.RACK_LOCAL)) { - val rackLocalTask = findTaskFromList(getRackLocalPendingTasksForHost(hostPort)) - if (rackLocalTask != None) { - return rackLocalTask - } - } - - // Look for no pref tasks AFTER rack local tasks - this has side effect that we will get to failed tasks later rather than sooner. - // TODO: That code path needs to be revisited (adding to no prefs list when host:port goes down). - val noPrefTask = findTaskFromList(pendingTasksWithNoPrefs) - if (noPrefTask != None) { - return noPrefTask - } - - if (TaskLocality.isAllowed(locality, TaskLocality.ANY)) { - val nonLocalTask = findTaskFromList(allPendingTasks) - if (nonLocalTask != None) { - return nonLocalTask - } - } - - // Finally, if all else has failed, find a speculative task - return findSpeculativeTask(hostPort, locality) - } - - // Does a host count as a preferred location for a task? This is true if - // either the task has preferred locations and this host is one, or it has - // no preferred locations (in which we still count the launch as preferred). - private def isPreferredLocation(task: Task[_], hostPort: String): Boolean = { - val locs = task.preferredLocations - // DEBUG code - locs.foreach(h => Utils.checkHost(h, "preferredLocation " + locs)) - - if (locs.contains(hostPort) || locs.isEmpty) return true - - val host = Utils.parseHostPort(hostPort)._1 - locs.contains(host) - } - - // Does a host count as a rack local preferred location for a task? (assumes host is NOT preferred location). - // This is true if either the task has preferred locations and this host is one, or it has - // no preferred locations (in which we still count the launch as preferred). - def isRackLocalLocation(task: Task[_], hostPort: String): Boolean = { - - val locs = task.preferredLocations - - // DEBUG code - locs.foreach(h => Utils.checkHost(h, "preferredLocation " + locs)) - - val preferredRacks = new HashSet[String]() - for (preferredHost <- locs) { - val rack = sched.getRackForHost(preferredHost) - if (None != rack) preferredRacks += rack.get - } - - if (preferredRacks.isEmpty) return false - - val hostRack = sched.getRackForHost(hostPort) - - return None != hostRack && preferredRacks.contains(hostRack.get) - } - - // Respond to an offer of a single slave from the scheduler by finding a task - def slaveOffer(execId: String, hostPort: String, availableCpus: Double, overrideLocality: TaskLocality.TaskLocality = null): Option[TaskDescription] = { - - if (tasksFinished < numTasks && availableCpus >= CPUS_PER_TASK) { - // If explicitly specified, use that - val locality = if (overrideLocality != null) overrideLocality else { - // expand only if we have waited for more than LOCALITY_WAIT for a host local task ... - val time = System.currentTimeMillis - if (time - lastPreferredLaunchTime < LOCALITY_WAIT) TaskLocality.HOST_LOCAL else TaskLocality.ANY - } - - findTask(hostPort, locality) match { - case Some(index) => { - // Found a task; do some bookkeeping and return a Mesos task for it - val task = tasks(index) - val taskId = sched.newTaskId() - // Figure out whether this should count as a preferred launch - val taskLocality = if (isPreferredLocation(task, hostPort)) TaskLocality.HOST_LOCAL else - if (isRackLocalLocation(task, hostPort)) TaskLocality.RACK_LOCAL else TaskLocality.ANY - val prefStr = taskLocality.toString - logInfo("Starting task %s:%d as TID %s on slave %s: %s (%s)".format( - taskSet.id, index, taskId, execId, hostPort, prefStr)) - // Do various bookkeeping - copiesRunning(index) += 1 - val time = System.currentTimeMillis - val info = new TaskInfo(taskId, index, time, execId, hostPort, taskLocality) - taskInfos(taskId) = info - taskAttempts(index) = info :: taskAttempts(index) - if (TaskLocality.HOST_LOCAL == taskLocality) { - lastPreferredLaunchTime = time - } - // Serialize and return the task - val startTime = System.currentTimeMillis - val serializedTask = Task.serializeWithDependencies( - task, sched.sc.addedFiles, sched.sc.addedJars, ser) - val timeTaken = System.currentTimeMillis - startTime - logInfo("Serialized task %s:%d as %d bytes in %d ms".format( - taskSet.id, index, serializedTask.limit, timeTaken)) - val taskName = "task %s:%d".format(taskSet.id, index) - return Some(new TaskDescription(taskId, execId, taskName, serializedTask)) - } - case _ => - } - } - return None - } +package spark.scheduler.cluster - def statusUpdate(tid: Long, state: TaskState, serializedData: ByteBuffer) { - state match { - case TaskState.FINISHED => - taskFinished(tid, state, serializedData) - case TaskState.LOST => - taskLost(tid, state, serializedData) - case TaskState.FAILED => - taskLost(tid, state, serializedData) - case TaskState.KILLED => - taskLost(tid, state, serializedData) - case _ => - } - } +import java.nio.ByteBuffer - def taskFinished(tid: Long, state: TaskState, serializedData: ByteBuffer) { - val info = taskInfos(tid) - if (info.failed) { - // We might get two task-lost messages for the same task in coarse-grained Mesos mode, - // or even from Mesos itself when acks get delayed. - return - } - val index = info.index - info.markSuccessful() - if (!finished(index)) { - tasksFinished += 1 - logInfo("Finished TID %s in %d ms (progress: %d/%d)".format( - tid, info.duration, tasksFinished, numTasks)) - // Deserialize task result and pass it to the scheduler - val result = ser.deserialize[TaskResult[_]](serializedData, getClass.getClassLoader) - result.metrics.resultSize = serializedData.limit() - sched.listener.taskEnded(tasks(index), Success, result.value, result.accumUpdates, info, result.metrics) - // Mark finished and stop if we've finished all the tasks - finished(index) = true - if (tasksFinished == numTasks) { - sched.taskSetFinished(this) - } - } else { - logInfo("Ignoring task-finished event for TID " + tid + - " because task " + index + " is already finished") - } - } +import spark.TaskState.TaskState +import spark.scheduler.TaskSet - def taskLost(tid: Long, state: TaskState, serializedData: ByteBuffer) { - val info = taskInfos(tid) - if (info.failed) { - // We might get two task-lost messages for the same task in coarse-grained Mesos mode, - // or even from Mesos itself when acks get delayed. - return - } - val index = info.index - info.markFailed() - if (!finished(index)) { - logInfo("Lost TID %s (task %s:%d)".format(tid, taskSet.id, index)) - copiesRunning(index) -= 1 - // Check if the problem is a map output fetch failure. In that case, this - // task will never succeed on any node, so tell the scheduler about it. - if (serializedData != null && serializedData.limit() > 0) { - val reason = ser.deserialize[TaskEndReason](serializedData, getClass.getClassLoader) - reason match { - case fetchFailed: FetchFailed => - logInfo("Loss was due to fetch failure from " + fetchFailed.bmAddress) - sched.listener.taskEnded(tasks(index), fetchFailed, null, null, info, null) - finished(index) = true - tasksFinished += 1 - sched.taskSetFinished(this) - return +private[spark] trait TaskSetManager extends Schedulable { - case ef: ExceptionFailure => - val key = ef.exception.toString - val now = System.currentTimeMillis - val (printFull, dupCount) = { - if (recentExceptions.contains(key)) { - val (dupCount, printTime) = recentExceptions(key) - if (now - printTime > EXCEPTION_PRINT_INTERVAL) { - recentExceptions(key) = (0, now) - (true, 0) - } else { - recentExceptions(key) = (dupCount + 1, printTime) - (false, dupCount + 1) - } - } else { - recentExceptions(key) = (0, now) - (true, 0) - } - } - if (printFull) { - val locs = ef.exception.getStackTrace.map(loc => "\tat %s".format(loc.toString)) - logInfo("Loss was due to %s\n%s".format(ef.exception.toString, locs.mkString("\n"))) - } else { - logInfo("Loss was due to %s [duplicate %d]".format(ef.exception.toString, dupCount)) - } + def taskSet: TaskSet - case _ => {} - } - } - // On non-fetch failures, re-enqueue the task as pending for a max number of retries - addPendingTask(index) - // Count failed attempts only on FAILED and LOST state (not on KILLED) - if (state == TaskState.FAILED || state == TaskState.LOST) { - numFailures(index) += 1 - if (numFailures(index) > MAX_TASK_FAILURES) { - logError("Task %s:%d failed more than %d times; aborting job".format( - taskSet.id, index, MAX_TASK_FAILURES)) - abort("Task %s:%d failed more than %d times".format(taskSet.id, index, MAX_TASK_FAILURES)) - } - } - } else { - logInfo("Ignoring task-lost event for TID " + tid + - " because task " + index + " is already finished") - } - } + def slaveOffer( + execId: String, + hostPort: String, + availableCpus: Double, + overrideLocality: TaskLocality.TaskLocality = null): Option[TaskDescription] - def error(message: String) { - // Save the error message - abort("Error: " + message) - } + def numPendingTasksForHostPort(hostPort: String): Int - def abort(message: String) { - failed = true - causeOfFailure = message - // TODO: Kill running tasks if we were not terminated due to a Mesos error - sched.listener.taskSetFailed(taskSet, message) - sched.taskSetFinished(this) - } + def numRackLocalPendingTasksForHost(hostPort: String): Int - def executorLost(execId: String, hostPort: String) { - logInfo("Re-queueing tasks for " + execId + " from TaskSet " + taskSet.id) - // If some task has preferred locations only on hostname, and there are no more executors there, - // put it in the no-prefs list to avoid the wait from delay scheduling - for (index <- getPendingTasksForHostPort(hostPort)) { - val newLocs = findPreferredLocations(tasks(index).preferredLocations, sched, true) - if (newLocs.isEmpty) { - assert (findPreferredLocations(tasks(index).preferredLocations, sched).isEmpty) - pendingTasksWithNoPrefs += index - } - } - // Re-enqueue any tasks that ran on the failed executor if this is a shuffle map stage - if (tasks(0).isInstanceOf[ShuffleMapTask]) { - for ((tid, info) <- taskInfos if info.executorId == execId) { - val index = taskInfos(tid).index - if (finished(index)) { - finished(index) = false - copiesRunning(index) -= 1 - tasksFinished -= 1 - addPendingTask(index) - // Tell the DAGScheduler that this task was resubmitted so that it doesn't think our - // stage finishes when a total of tasks.size tasks finish. - sched.listener.taskEnded(tasks(index), Resubmitted, null, null, info, null) - } - } - } - // Also re-enqueue any tasks that were running on the node - for ((tid, info) <- taskInfos if info.running && info.executorId == execId) { - taskLost(tid, TaskState.KILLED, null) - } - } + def numPendingTasksForHost(hostPort: String): Int - /** - * Check for tasks to be speculated and return true if there are any. This is called periodically - * by the ClusterScheduler. - * - * TODO: To make this scale to large jobs, we need to maintain a list of running tasks, so that - * we don't scan the whole task set. It might also help to make this sorted by launch time. - */ - def checkSpeculatableTasks(): Boolean = { - // Can't speculate if we only have one task, or if all tasks have finished. - if (numTasks == 1 || tasksFinished == numTasks) { - return false - } - var foundTasks = false - val minFinishedForSpeculation = (SPECULATION_QUANTILE * numTasks).floor.toInt - logDebug("Checking for speculative tasks: minFinished = " + minFinishedForSpeculation) - if (tasksFinished >= minFinishedForSpeculation) { - val time = System.currentTimeMillis() - val durations = taskInfos.values.filter(_.successful).map(_.duration).toArray - Arrays.sort(durations) - val medianDuration = durations(min((0.5 * numTasks).round.toInt, durations.size - 1)) - val threshold = max(SPECULATION_MULTIPLIER * medianDuration, 100) - // TODO: Threshold should also look at standard deviation of task durations and have a lower - // bound based on that. - logDebug("Task length threshold for speculation: " + threshold) - for ((tid, info) <- taskInfos) { - val index = info.index - if (!finished(index) && copiesRunning(index) == 1 && info.timeRunning(time) > threshold && - !speculatableTasks.contains(index)) { - logInfo( - "Marking task %s:%d (on %s) as speculatable because it ran more than %.0f ms".format( - taskSet.id, index, info.hostPort, threshold)) - speculatableTasks += index - foundTasks = true - } - } - } - return foundTasks - } + def statusUpdate(tid: Long, state: TaskState, serializedData: ByteBuffer) - def hasPendingTasks(): Boolean = { - numTasks > 0 && tasksFinished < numTasks - } + def error(message: String) } diff --git a/core/src/main/scala/spark/scheduler/cluster/WorkerOffer.scala b/core/src/main/scala/spark/scheduler/cluster/WorkerOffer.scala index c47824315c..06d1203f70 100644 --- a/core/src/main/scala/spark/scheduler/cluster/WorkerOffer.scala +++ b/core/src/main/scala/spark/scheduler/cluster/WorkerOffer.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.cluster /** diff --git a/core/src/main/scala/spark/scheduler/local/LocalScheduler.scala b/core/src/main/scala/spark/scheduler/local/LocalScheduler.scala index f060a940a9..edd83d4cb4 100644 --- a/core/src/main/scala/spark/scheduler/local/LocalScheduler.scala +++ b/core/src/main/scala/spark/scheduler/local/LocalScheduler.scala @@ -1,20 +1,73 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.local import java.io.File import java.util.concurrent.atomic.AtomicInteger +import java.nio.ByteBuffer +import scala.collection.mutable.ArrayBuffer import scala.collection.mutable.HashMap +import scala.collection.mutable.HashSet import spark._ +import spark.TaskState.TaskState import spark.executor.ExecutorURLClassLoader import spark.scheduler._ -import spark.scheduler.cluster.{TaskLocality, TaskInfo} +import spark.scheduler.cluster._ +import akka.actor._ /** - * A simple TaskScheduler implementation that runs tasks locally in a thread pool. Optionally + * A FIFO or Fair TaskScheduler implementation that runs tasks locally in a thread pool. Optionally * the scheduler also allows each task to fail up to maxFailures times, which is useful for * testing fault recovery. */ -private[spark] class LocalScheduler(threads: Int, maxFailures: Int, sc: SparkContext) + +private[spark] +case class LocalReviveOffers() + +private[spark] +case class LocalStatusUpdate(taskId: Long, state: TaskState, serializedData: ByteBuffer) + +private[spark] +class LocalActor(localScheduler: LocalScheduler, var freeCores: Int) extends Actor with Logging { + + def receive = { + case LocalReviveOffers => + launchTask(localScheduler.resourceOffer(freeCores)) + case LocalStatusUpdate(taskId, state, serializeData) => + freeCores += 1 + localScheduler.statusUpdate(taskId, state, serializeData) + launchTask(localScheduler.resourceOffer(freeCores)) + } + + def launchTask(tasks : Seq[TaskDescription]) { + for (task <- tasks) { + freeCores -= 1 + localScheduler.threadPool.submit(new Runnable { + def run() { + localScheduler.runTask(task.taskId, task.serializedTask) + } + }) + } + } +} + +private[spark] class LocalScheduler(threads: Int, val maxFailures: Int, val sc: SparkContext) extends TaskScheduler with Logging { @@ -30,87 +83,135 @@ private[spark] class LocalScheduler(threads: Int, maxFailures: Int, sc: SparkCon val classLoader = new ExecutorURLClassLoader(Array(), Thread.currentThread.getContextClassLoader) - // TODO: Need to take into account stage priority in scheduling + var schedulableBuilder: SchedulableBuilder = null + var rootPool: Pool = null + val activeTaskSets = new HashMap[String, TaskSetManager] + val taskIdToTaskSetId = new HashMap[Long, String] + val taskSetTaskIds = new HashMap[String, HashSet[Long]] - override def start() { } + var localActor: ActorRef = null + + override def start() { + //default scheduler is FIFO + val schedulingMode = System.getProperty("spark.cluster.schedulingmode", "FIFO") + //temporarily set rootPool name to empty + rootPool = new Pool("", SchedulingMode.withName(schedulingMode), 0, 0) + schedulableBuilder = { + schedulingMode match { + case "FIFO" => + new FIFOSchedulableBuilder(rootPool) + case "FAIR" => + new FairSchedulableBuilder(rootPool) + } + } + schedulableBuilder.buildPools() + + localActor = env.actorSystem.actorOf(Props(new LocalActor(this, threads)), "Test") + } override def setListener(listener: TaskSchedulerListener) { this.listener = listener } override def submitTasks(taskSet: TaskSet) { - val tasks = taskSet.tasks - val failCount = new Array[Int](tasks.size) - - def submitTask(task: Task[_], idInJob: Int) { - val myAttemptId = attemptId.getAndIncrement() - threadPool.submit(new Runnable { - def run() { - runTask(task, idInJob, myAttemptId) - } - }) + synchronized { + val manager = new LocalTaskSetManager(this, taskSet) + schedulableBuilder.addTaskSetManager(manager, manager.taskSet.properties) + activeTaskSets(taskSet.id) = manager + taskSetTaskIds(taskSet.id) = new HashSet[Long]() + localActor ! LocalReviveOffers } + } - def runTask(task: Task[_], idInJob: Int, attemptId: Int) { - logInfo("Running " + task) - val info = new TaskInfo(attemptId, idInJob, System.currentTimeMillis(), "local", "local:1", TaskLocality.HOST_LOCAL) - // Set the Spark execution environment for the worker thread - SparkEnv.set(env) - try { - Accumulators.clear() - Thread.currentThread().setContextClassLoader(classLoader) - - // Serialize and deserialize the task so that accumulators are changed to thread-local ones; - // this adds a bit of unnecessary overhead but matches how the Mesos Executor works. - val ser = SparkEnv.get.closureSerializer.newInstance() - val bytes = Task.serializeWithDependencies(task, sc.addedFiles, sc.addedJars, ser) - logInfo("Size of task " + idInJob + " is " + bytes.limit + " bytes") - val (taskFiles, taskJars, taskBytes) = Task.deserializeWithDependencies(bytes) - updateDependencies(taskFiles, taskJars) // Download any files added with addFile - val deserStart = System.currentTimeMillis() - val deserializedTask = ser.deserialize[Task[_]]( - taskBytes, Thread.currentThread.getContextClassLoader) - val deserTime = System.currentTimeMillis() - deserStart - - // Run it - val result: Any = deserializedTask.run(attemptId) - - // Serialize and deserialize the result to emulate what the Mesos - // executor does. This is useful to catch serialization errors early - // on in development (so when users move their local Spark programs - // to the cluster, they don't get surprised by serialization errors). - val serResult = ser.serialize(result) - deserializedTask.metrics.get.resultSize = serResult.limit() - val resultToReturn = ser.deserialize[Any](serResult) - val accumUpdates = ser.deserialize[collection.mutable.Map[Long, Any]]( - ser.serialize(Accumulators.values)) - logInfo("Finished " + task) - info.markSuccessful() - deserializedTask.metrics.get.executorRunTime = info.duration.toInt //close enough - deserializedTask.metrics.get.executorDeserializeTime = deserTime.toInt - - // If the threadpool has not already been shutdown, notify DAGScheduler - if (!Thread.currentThread().isInterrupted) - listener.taskEnded(task, Success, resultToReturn, accumUpdates, info, deserializedTask.metrics.getOrElse(null)) - } catch { - case t: Throwable => { - logError("Exception in task " + idInJob, t) - failCount.synchronized { - failCount(idInJob) += 1 - if (failCount(idInJob) <= maxFailures) { - submitTask(task, idInJob) - } else { - // TODO: Do something nicer here to return all the way to the user - if (!Thread.currentThread().isInterrupted) - listener.taskEnded(task, new ExceptionFailure(t), null, null, info, null) - } + def resourceOffer(freeCores: Int): Seq[TaskDescription] = { + synchronized { + var freeCpuCores = freeCores + val tasks = new ArrayBuffer[TaskDescription](freeCores) + val sortedTaskSetQueue = rootPool.getSortedTaskSetQueue() + for (manager <- sortedTaskSetQueue) { + logDebug("parentName:%s,name:%s,runningTasks:%s".format( + manager.parent.name, manager.name, manager.runningTasks)) + } + + var launchTask = false + for (manager <- sortedTaskSetQueue) { + do { + launchTask = false + manager.slaveOffer(null, null, freeCpuCores) match { + case Some(task) => + tasks += task + taskIdToTaskSetId(task.taskId) = manager.taskSet.id + taskSetTaskIds(manager.taskSet.id) += task.taskId + freeCpuCores -= 1 + launchTask = true + case None => {} } - } + } while(launchTask) } + return tasks + } + } + + def taskSetFinished(manager: TaskSetManager) { + synchronized { + activeTaskSets -= manager.taskSet.id + manager.parent.removeSchedulable(manager) + logInfo("Remove TaskSet %s from pool %s".format(manager.taskSet.id, manager.parent.name)) + taskIdToTaskSetId --= taskSetTaskIds(manager.taskSet.id) + taskSetTaskIds -= manager.taskSet.id } + } + + def runTask(taskId: Long, bytes: ByteBuffer) { + logInfo("Running " + taskId) + val info = new TaskInfo(taskId, 0, System.currentTimeMillis(), "local", "local:1", TaskLocality.NODE_LOCAL) + // Set the Spark execution environment for the worker thread + SparkEnv.set(env) + val ser = SparkEnv.get.closureSerializer.newInstance() + var attemptedTask: Option[Task[_]] = None + val start = System.currentTimeMillis() + var taskStart: Long = 0 + try { + Accumulators.clear() + Thread.currentThread().setContextClassLoader(classLoader) - for ((task, i) <- tasks.zipWithIndex) { - submitTask(task, i) + // Serialize and deserialize the task so that accumulators are changed to thread-local ones; + // this adds a bit of unnecessary overhead but matches how the Mesos Executor works. + val (taskFiles, taskJars, taskBytes) = Task.deserializeWithDependencies(bytes) + updateDependencies(taskFiles, taskJars) // Download any files added with addFile + val deserializedTask = ser.deserialize[Task[_]]( + taskBytes, Thread.currentThread.getContextClassLoader) + attemptedTask = Some(deserializedTask) + val deserTime = System.currentTimeMillis() - start + taskStart = System.currentTimeMillis() + + // Run it + val result: Any = deserializedTask.run(taskId) + + // Serialize and deserialize the result to emulate what the Mesos + // executor does. This is useful to catch serialization errors early + // on in development (so when users move their local Spark programs + // to the cluster, they don't get surprised by serialization errors). + val serResult = ser.serialize(result) + deserializedTask.metrics.get.resultSize = serResult.limit() + val resultToReturn = ser.deserialize[Any](serResult) + val accumUpdates = ser.deserialize[collection.mutable.Map[Long, Any]]( + ser.serialize(Accumulators.values)) + val serviceTime = System.currentTimeMillis() - taskStart + logInfo("Finished " + taskId) + deserializedTask.metrics.get.executorRunTime = serviceTime.toInt + deserializedTask.metrics.get.executorDeserializeTime = deserTime.toInt + val taskResult = new TaskResult(result, accumUpdates, deserializedTask.metrics.getOrElse(null)) + val serializedResult = ser.serialize(taskResult) + localActor ! LocalStatusUpdate(taskId, TaskState.FINISHED, serializedResult) + } catch { + case t: Throwable => { + val serviceTime = System.currentTimeMillis() - taskStart + val metrics = attemptedTask.flatMap(t => t.metrics) + metrics.foreach{m => m.executorRunTime = serviceTime.toInt} + val failure = new ExceptionFailure(t.getClass.getName, t.toString, t.getStackTrace, metrics) + localActor ! LocalStatusUpdate(taskId, TaskState.FAILED, ser.serialize(failure)) + } } } @@ -126,6 +227,7 @@ private[spark] class LocalScheduler(threads: Int, maxFailures: Int, sc: SparkCon Utils.fetchFile(name, new File(SparkFiles.getRootDirectory)) currentFiles(name) = timestamp } + for ((name, timestamp) <- newJars if currentJars.getOrElse(name, -1L) < timestamp) { logInfo("Fetching " + name + " with timestamp " + timestamp) Utils.fetchFile(name, new File(SparkFiles.getRootDirectory)) @@ -141,7 +243,16 @@ private[spark] class LocalScheduler(threads: Int, maxFailures: Int, sc: SparkCon } } - override def stop() { + def statusUpdate(taskId :Long, state: TaskState, serializedData: ByteBuffer) { + synchronized { + val taskSetId = taskIdToTaskSetId(taskId) + val taskSetManager = activeTaskSets(taskSetId) + taskSetTaskIds(taskSetId) -= taskId + taskSetManager.statusUpdate(taskId, state, serializedData) + } + } + + override def stop() { threadPool.shutdownNow() } diff --git a/core/src/main/scala/spark/scheduler/local/LocalTaskSetManager.scala b/core/src/main/scala/spark/scheduler/local/LocalTaskSetManager.scala new file mode 100644 index 0000000000..b29740c886 --- /dev/null +++ b/core/src/main/scala/spark/scheduler/local/LocalTaskSetManager.scala @@ -0,0 +1,203 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler.local + +import java.nio.ByteBuffer +import scala.collection.mutable.ArrayBuffer +import scala.collection.mutable.HashMap + +import spark.{ExceptionFailure, Logging, SparkEnv, Success, TaskState} +import spark.TaskState.TaskState +import spark.scheduler.{Task, TaskResult, TaskSet} +import spark.scheduler.cluster.{Schedulable, TaskDescription, TaskInfo, TaskLocality, TaskSetManager} + + +private[spark] class LocalTaskSetManager(sched: LocalScheduler, val taskSet: TaskSet) + extends TaskSetManager with Logging { + + var parent: Schedulable = null + var weight: Int = 1 + var minShare: Int = 0 + var runningTasks: Int = 0 + var priority: Int = taskSet.priority + var stageId: Int = taskSet.stageId + var name: String = "TaskSet_" + taskSet.stageId.toString + + var failCount = new Array[Int](taskSet.tasks.size) + val taskInfos = new HashMap[Long, TaskInfo] + val numTasks = taskSet.tasks.size + var numFinished = 0 + val ser = SparkEnv.get.closureSerializer.newInstance() + val copiesRunning = new Array[Int](numTasks) + val finished = new Array[Boolean](numTasks) + val numFailures = new Array[Int](numTasks) + val MAX_TASK_FAILURES = sched.maxFailures + + override def increaseRunningTasks(taskNum: Int): Unit = { + runningTasks += taskNum + if (parent != null) { + parent.increaseRunningTasks(taskNum) + } + } + + override def decreaseRunningTasks(taskNum: Int): Unit = { + runningTasks -= taskNum + if (parent != null) { + parent.decreaseRunningTasks(taskNum) + } + } + + override def addSchedulable(schedulable: Schedulable): Unit = { + //nothing + } + + override def removeSchedulable(schedulable: Schedulable): Unit = { + //nothing + } + + override def getSchedulableByName(name: String): Schedulable = { + return null + } + + override def executorLost(executorId: String, host: String): Unit = { + //nothing + } + + override def checkSpeculatableTasks() = true + + override def getSortedTaskSetQueue(): ArrayBuffer[TaskSetManager] = { + var sortedTaskSetQueue = new ArrayBuffer[TaskSetManager] + sortedTaskSetQueue += this + return sortedTaskSetQueue + } + + override def hasPendingTasks() = true + + def findTask(): Option[Int] = { + for (i <- 0 to numTasks-1) { + if (copiesRunning(i) == 0 && !finished(i)) { + return Some(i) + } + } + return None + } + + override def slaveOffer( + execId: String, + hostPort: String, + availableCpus: Double, + overrideLocality: TaskLocality.TaskLocality = null): Option[TaskDescription] = + { + SparkEnv.set(sched.env) + logDebug("availableCpus:%d,numFinished:%d,numTasks:%d".format( + availableCpus.toInt, numFinished, numTasks)) + if (availableCpus > 0 && numFinished < numTasks) { + findTask() match { + case Some(index) => + val taskId = sched.attemptId.getAndIncrement() + val task = taskSet.tasks(index) + val info = new TaskInfo(taskId, index, System.currentTimeMillis(), "local", "local:1", + TaskLocality.NODE_LOCAL) + taskInfos(taskId) = info + // We rely on the DAGScheduler to catch non-serializable closures and RDDs, so in here + // we assume the task can be serialized without exceptions. + val bytes = Task.serializeWithDependencies( + task, sched.sc.addedFiles, sched.sc.addedJars, ser) + logInfo("Size of task " + taskId + " is " + bytes.limit + " bytes") + val taskName = "task %s:%d".format(taskSet.id, index) + copiesRunning(index) += 1 + increaseRunningTasks(1) + taskStarted(task, info) + return Some(new TaskDescription(taskId, null, taskName, bytes)) + case None => {} + } + } + return None + } + + override def numPendingTasksForHostPort(hostPort: String): Int = { + return 0 + } + + override def numRackLocalPendingTasksForHost(hostPort :String): Int = { + return 0 + } + + override def numPendingTasksForHost(hostPort: String): Int = { + return 0 + } + + override def statusUpdate(tid: Long, state: TaskState, serializedData: ByteBuffer) { + state match { + case TaskState.FINISHED => + taskEnded(tid, state, serializedData) + case TaskState.FAILED => + taskFailed(tid, state, serializedData) + case _ => {} + } + } + + def taskStarted(task: Task[_], info: TaskInfo) { + sched.listener.taskStarted(task, info) + } + + def taskEnded(tid: Long, state: TaskState, serializedData: ByteBuffer) { + val info = taskInfos(tid) + val index = info.index + val task = taskSet.tasks(index) + info.markSuccessful() + val result = ser.deserialize[TaskResult[_]](serializedData, getClass.getClassLoader) + result.metrics.resultSize = serializedData.limit() + sched.listener.taskEnded(task, Success, result.value, result.accumUpdates, info, result.metrics) + numFinished += 1 + decreaseRunningTasks(1) + finished(index) = true + if (numFinished == numTasks) { + sched.taskSetFinished(this) + } + } + + def taskFailed(tid: Long, state: TaskState, serializedData: ByteBuffer) { + val info = taskInfos(tid) + val index = info.index + val task = taskSet.tasks(index) + info.markFailed() + decreaseRunningTasks(1) + val reason: ExceptionFailure = ser.deserialize[ExceptionFailure]( + serializedData, getClass.getClassLoader) + sched.listener.taskEnded(task, reason, null, null, info, reason.metrics.getOrElse(null)) + if (!finished(index)) { + copiesRunning(index) -= 1 + numFailures(index) += 1 + val locs = reason.stackTrace.map(loc => "\tat %s".format(loc.toString)) + logInfo("Loss was due to %s\n%s\n%s".format( + reason.className, reason.description, locs.mkString("\n"))) + if (numFailures(index) > MAX_TASK_FAILURES) { + val errorMessage = "Task %s:%d failed more than %d times; aborting job %s".format( + taskSet.id, index, 4, reason.description) + decreaseRunningTasks(runningTasks) + sched.listener.taskSetFailed(taskSet, errorMessage) + // need to delete failed Taskset from schedule queue + sched.taskSetFinished(this) + } + } + } + + override def error(message: String) { + } +} diff --git a/core/src/main/scala/spark/scheduler/mesos/CoarseMesosSchedulerBackend.scala b/core/src/main/scala/spark/scheduler/mesos/CoarseMesosSchedulerBackend.scala index f4a2994b6d..7bc6040544 100644 --- a/core/src/main/scala/spark/scheduler/mesos/CoarseMesosSchedulerBackend.scala +++ b/core/src/main/scala/spark/scheduler/mesos/CoarseMesosSchedulerBackend.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.mesos import com.google.protobuf.ByteString diff --git a/core/src/main/scala/spark/scheduler/mesos/MesosSchedulerBackend.scala b/core/src/main/scala/spark/scheduler/mesos/MesosSchedulerBackend.scala index ca7fab4cc5..75b8268b55 100644 --- a/core/src/main/scala/spark/scheduler/mesos/MesosSchedulerBackend.scala +++ b/core/src/main/scala/spark/scheduler/mesos/MesosSchedulerBackend.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler.mesos import com.google.protobuf.ByteString @@ -43,8 +60,12 @@ private[spark] class MesosSchedulerBackend( // An ExecutorInfo for our tasks var execArgs: Array[Byte] = null + var classLoader: ClassLoader = null + override def start() { synchronized { + classLoader = Thread.currentThread.getContextClassLoader + new Thread("MesosSchedulerBackend driver") { setDaemon(true) override def run() { @@ -114,13 +135,28 @@ private[spark] class MesosSchedulerBackend( return execArgs } + private def setClassLoader(): ClassLoader = { + val oldClassLoader = Thread.currentThread.getContextClassLoader + Thread.currentThread.setContextClassLoader(classLoader) + return oldClassLoader + } + + private def restoreClassLoader(oldClassLoader: ClassLoader) { + Thread.currentThread.setContextClassLoader(oldClassLoader) + } + override def offerRescinded(d: SchedulerDriver, o: OfferID) {} override def registered(d: SchedulerDriver, frameworkId: FrameworkID, masterInfo: MasterInfo) { - logInfo("Registered as framework ID " + frameworkId.getValue) - registeredLock.synchronized { - isRegistered = true - registeredLock.notifyAll() + val oldClassLoader = setClassLoader() + try { + logInfo("Registered as framework ID " + frameworkId.getValue) + registeredLock.synchronized { + isRegistered = true + registeredLock.notifyAll() + } + } finally { + restoreClassLoader(oldClassLoader) } } @@ -142,49 +178,54 @@ private[spark] class MesosSchedulerBackend( * tasks are balanced across the cluster. */ override def resourceOffers(d: SchedulerDriver, offers: JList[Offer]) { - synchronized { - // Build a big list of the offerable workers, and remember their indices so that we can - // figure out which Offer to reply to for each worker - val offerableIndices = new ArrayBuffer[Int] - val offerableWorkers = new ArrayBuffer[WorkerOffer] - - def enoughMemory(o: Offer) = { - val mem = getResource(o.getResourcesList, "mem") - val slaveId = o.getSlaveId.getValue - mem >= executorMemory || slaveIdsWithExecutors.contains(slaveId) - } + val oldClassLoader = setClassLoader() + try { + synchronized { + // Build a big list of the offerable workers, and remember their indices so that we can + // figure out which Offer to reply to for each worker + val offerableIndices = new ArrayBuffer[Int] + val offerableWorkers = new ArrayBuffer[WorkerOffer] + + def enoughMemory(o: Offer) = { + val mem = getResource(o.getResourcesList, "mem") + val slaveId = o.getSlaveId.getValue + mem >= executorMemory || slaveIdsWithExecutors.contains(slaveId) + } - for ((offer, index) <- offers.zipWithIndex if enoughMemory(offer)) { - offerableIndices += index - offerableWorkers += new WorkerOffer( - offer.getSlaveId.getValue, - offer.getHostname, - getResource(offer.getResourcesList, "cpus").toInt) - } + for ((offer, index) <- offers.zipWithIndex if enoughMemory(offer)) { + offerableIndices += index + offerableWorkers += new WorkerOffer( + offer.getSlaveId.getValue, + offer.getHostname, + getResource(offer.getResourcesList, "cpus").toInt) + } - // Call into the ClusterScheduler - val taskLists = scheduler.resourceOffers(offerableWorkers) - - // Build a list of Mesos tasks for each slave - val mesosTasks = offers.map(o => Collections.emptyList[MesosTaskInfo]()) - for ((taskList, index) <- taskLists.zipWithIndex) { - if (!taskList.isEmpty) { - val offerNum = offerableIndices(index) - val slaveId = offers(offerNum).getSlaveId.getValue - slaveIdsWithExecutors += slaveId - mesosTasks(offerNum) = new JArrayList[MesosTaskInfo](taskList.size) - for (taskDesc <- taskList) { - taskIdToSlaveId(taskDesc.taskId) = slaveId - mesosTasks(offerNum).add(createMesosTask(taskDesc, slaveId)) + // Call into the ClusterScheduler + val taskLists = scheduler.resourceOffers(offerableWorkers) + + // Build a list of Mesos tasks for each slave + val mesosTasks = offers.map(o => Collections.emptyList[MesosTaskInfo]()) + for ((taskList, index) <- taskLists.zipWithIndex) { + if (!taskList.isEmpty) { + val offerNum = offerableIndices(index) + val slaveId = offers(offerNum).getSlaveId.getValue + slaveIdsWithExecutors += slaveId + mesosTasks(offerNum) = new JArrayList[MesosTaskInfo](taskList.size) + for (taskDesc <- taskList) { + taskIdToSlaveId(taskDesc.taskId) = slaveId + mesosTasks(offerNum).add(createMesosTask(taskDesc, slaveId)) + } } } - } - // Reply to the offers - val filters = Filters.newBuilder().setRefuseSeconds(1).build() // TODO: lower timeout? - for (i <- 0 until offers.size) { - d.launchTasks(offers(i).getId, mesosTasks(i), filters) + // Reply to the offers + val filters = Filters.newBuilder().setRefuseSeconds(1).build() // TODO: lower timeout? + for (i <- 0 until offers.size) { + d.launchTasks(offers(i).getId, mesosTasks(i), filters) + } } + } finally { + restoreClassLoader(oldClassLoader) } } @@ -224,23 +265,33 @@ private[spark] class MesosSchedulerBackend( } override def statusUpdate(d: SchedulerDriver, status: TaskStatus) { - val tid = status.getTaskId.getValue.toLong - val state = TaskState.fromMesos(status.getState) - synchronized { - if (status.getState == MesosTaskState.TASK_LOST && taskIdToSlaveId.contains(tid)) { - // We lost the executor on this slave, so remember that it's gone - slaveIdsWithExecutors -= taskIdToSlaveId(tid) - } - if (isFinished(status.getState)) { - taskIdToSlaveId.remove(tid) + val oldClassLoader = setClassLoader() + try { + val tid = status.getTaskId.getValue.toLong + val state = TaskState.fromMesos(status.getState) + synchronized { + if (status.getState == MesosTaskState.TASK_LOST && taskIdToSlaveId.contains(tid)) { + // We lost the executor on this slave, so remember that it's gone + slaveIdsWithExecutors -= taskIdToSlaveId(tid) + } + if (isFinished(status.getState)) { + taskIdToSlaveId.remove(tid) + } } + scheduler.statusUpdate(tid, state, status.getData.asReadOnlyByteBuffer) + } finally { + restoreClassLoader(oldClassLoader) } - scheduler.statusUpdate(tid, state, status.getData.asReadOnlyByteBuffer) } override def error(d: SchedulerDriver, message: String) { - logError("Mesos error: " + message) - scheduler.error(message) + val oldClassLoader = setClassLoader() + try { + logError("Mesos error: " + message) + scheduler.error(message) + } finally { + restoreClassLoader(oldClassLoader) + } } override def stop() { @@ -256,11 +307,16 @@ private[spark] class MesosSchedulerBackend( override def frameworkMessage(d: SchedulerDriver, e: ExecutorID, s: SlaveID, b: Array[Byte]) {} private def recordSlaveLost(d: SchedulerDriver, slaveId: SlaveID, reason: ExecutorLossReason) { - logInfo("Mesos slave lost: " + slaveId.getValue) - synchronized { - slaveIdsWithExecutors -= slaveId.getValue + val oldClassLoader = setClassLoader() + try { + logInfo("Mesos slave lost: " + slaveId.getValue) + synchronized { + slaveIdsWithExecutors -= slaveId.getValue + } + scheduler.executorLost(slaveId.getValue, reason) + } finally { + restoreClassLoader(oldClassLoader) } - scheduler.executorLost(slaveId.getValue, reason) } override def slaveLost(d: SchedulerDriver, slaveId: SlaveID) { diff --git a/core/src/main/scala/spark/serializer/Serializer.scala b/core/src/main/scala/spark/serializer/Serializer.scala index aca86ab6f0..dc94d42bb6 100644 --- a/core/src/main/scala/spark/serializer/Serializer.scala +++ b/core/src/main/scala/spark/serializer/Serializer.scala @@ -1,10 +1,30 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.serializer -import java.nio.ByteBuffer import java.io.{EOFException, InputStream, OutputStream} +import java.nio.ByteBuffer + import it.unimi.dsi.fastutil.io.FastByteArrayOutputStream + import spark.util.ByteBufferInputStream + /** * A serializer. Because some serialization libraries are not thread safe, this class is used to * create [[spark.serializer.SerializerInstance]] objects that do the actual serialization and are @@ -14,6 +34,7 @@ trait Serializer { def newInstance(): SerializerInstance } + /** * An instance of a serializer, for use by one thread at a time. */ @@ -45,6 +66,7 @@ trait SerializerInstance { } } + /** * A stream for writing serialized objects. */ @@ -61,6 +83,7 @@ trait SerializationStream { } } + /** * A stream for reading serialized objects. */ diff --git a/core/src/main/scala/spark/serializer/SerializerManager.scala b/core/src/main/scala/spark/serializer/SerializerManager.scala new file mode 100644 index 0000000000..b7b24705a2 --- /dev/null +++ b/core/src/main/scala/spark/serializer/SerializerManager.scala @@ -0,0 +1,62 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.serializer + +import java.util.concurrent.ConcurrentHashMap + + +/** + * A service that returns a serializer object given the serializer's class name. If a previous + * instance of the serializer object has been created, the get method returns that instead of + * creating a new one. + */ +private[spark] class SerializerManager { + + private val serializers = new ConcurrentHashMap[String, Serializer] + private var _default: Serializer = _ + + def default = _default + + def setDefault(clsName: String): Serializer = { + _default = get(clsName) + _default + } + + def get(clsName: String): Serializer = { + if (clsName == null) { + default + } else { + var serializer = serializers.get(clsName) + if (serializer != null) { + // If the serializer has been created previously, reuse that. + serializer + } else this.synchronized { + // Otherwise, create a new one. But make sure no other thread has attempted + // to create another new one at the same time. + serializer = serializers.get(clsName) + if (serializer == null) { + val clsLoader = Thread.currentThread.getContextClassLoader + serializer = + Class.forName(clsName, true, clsLoader).newInstance().asInstanceOf[Serializer] + serializers.put(clsName, serializer) + } + serializer + } + } + } +} diff --git a/core/src/main/scala/spark/storage/BlockException.scala b/core/src/main/scala/spark/storage/BlockException.scala new file mode 100644 index 0000000000..8ebfaf3cbf --- /dev/null +++ b/core/src/main/scala/spark/storage/BlockException.scala @@ -0,0 +1,22 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.storage + +private[spark] +case class BlockException(blockId: String, message: String) extends Exception(message) + diff --git a/core/src/main/scala/spark/storage/BlockFetchTracker.scala b/core/src/main/scala/spark/storage/BlockFetchTracker.scala index 993aece1f7..265e554ad8 100644 --- a/core/src/main/scala/spark/storage/BlockFetchTracker.scala +++ b/core/src/main/scala/spark/storage/BlockFetchTracker.scala @@ -1,10 +1,27 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage private[spark] trait BlockFetchTracker { - def totalBlocks : Int - def numLocalBlocks: Int - def numRemoteBlocks: Int - def remoteFetchTime : Long - def fetchWaitTime: Long - def remoteBytesRead : Long + def totalBlocks : Int + def numLocalBlocks: Int + def numRemoteBlocks: Int + def remoteFetchTime : Long + def fetchWaitTime: Long + def remoteBytesRead : Long } diff --git a/core/src/main/scala/spark/storage/BlockFetcherIterator.scala b/core/src/main/scala/spark/storage/BlockFetcherIterator.scala new file mode 100644 index 0000000000..1965c5bc19 --- /dev/null +++ b/core/src/main/scala/spark/storage/BlockFetcherIterator.scala @@ -0,0 +1,347 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.storage + +import java.nio.ByteBuffer +import java.util.concurrent.LinkedBlockingQueue + +import scala.collection.mutable.ArrayBuffer +import scala.collection.mutable.HashSet +import scala.collection.mutable.Queue + +import io.netty.buffer.ByteBuf + +import spark.Logging +import spark.Utils +import spark.SparkException +import spark.network.BufferMessage +import spark.network.ConnectionManagerId +import spark.network.netty.ShuffleCopier +import spark.serializer.Serializer + + +/** + * A block fetcher iterator interface. There are two implementations: + * + * BasicBlockFetcherIterator: uses a custom-built NIO communication layer. + * NettyBlockFetcherIterator: uses Netty (OIO) as the communication layer. + * + * Eventually we would like the two to converge and use a single NIO-based communication layer, + * but extensive tests show that under some circumstances (e.g. large shuffles with lots of cores), + * NIO would perform poorly and thus the need for the Netty OIO one. + */ + +private[storage] +trait BlockFetcherIterator extends Iterator[(String, Option[Iterator[Any]])] + with Logging with BlockFetchTracker { + def initialize() +} + + +private[storage] +object BlockFetcherIterator { + + // A request to fetch one or more blocks, complete with their sizes + class FetchRequest(val address: BlockManagerId, val blocks: Seq[(String, Long)]) { + val size = blocks.map(_._2).sum + } + + // A result of a fetch. Includes the block ID, size in bytes, and a function to deserialize + // the block (since we want all deserializaton to happen in the calling thread); can also + // represent a fetch failure if size == -1. + class FetchResult(val blockId: String, val size: Long, val deserialize: () => Iterator[Any]) { + def failed: Boolean = size == -1 + } + + class BasicBlockFetcherIterator( + private val blockManager: BlockManager, + val blocksByAddress: Seq[(BlockManagerId, Seq[(String, Long)])], + serializer: Serializer) + extends BlockFetcherIterator { + + import blockManager._ + + private var _remoteBytesRead = 0l + private var _remoteFetchTime = 0l + private var _fetchWaitTime = 0l + + if (blocksByAddress == null) { + throw new IllegalArgumentException("BlocksByAddress is null") + } + + // Total number blocks fetched (local + remote). Also number of FetchResults expected + protected var _numBlocksToFetch = 0 + + protected var startTime = System.currentTimeMillis + + // This represents the number of local blocks, also counting zero-sized blocks + private var numLocal = 0 + // BlockIds for local blocks that need to be fetched. Excludes zero-sized blocks + protected val localBlocksToFetch = new ArrayBuffer[String]() + + // This represents the number of remote blocks, also counting zero-sized blocks + private var numRemote = 0 + // BlockIds for remote blocks that need to be fetched. Excludes zero-sized blocks + protected val remoteBlocksToFetch = new HashSet[String]() + + // A queue to hold our results. + protected val results = new LinkedBlockingQueue[FetchResult] + + // Queue of fetch requests to issue; we'll pull requests off this gradually to make sure that + // the number of bytes in flight is limited to maxBytesInFlight + private val fetchRequests = new Queue[FetchRequest] + + // Current bytes in flight from our requests + private var bytesInFlight = 0L + + protected def sendRequest(req: FetchRequest) { + logDebug("Sending request for %d blocks (%s) from %s".format( + req.blocks.size, Utils.memoryBytesToString(req.size), req.address.hostPort)) + val cmId = new ConnectionManagerId(req.address.host, req.address.port) + val blockMessageArray = new BlockMessageArray(req.blocks.map { + case (blockId, size) => BlockMessage.fromGetBlock(GetBlock(blockId)) + }) + bytesInFlight += req.size + val sizeMap = req.blocks.toMap // so we can look up the size of each blockID + val fetchStart = System.currentTimeMillis() + val future = connectionManager.sendMessageReliably(cmId, blockMessageArray.toBufferMessage) + future.onSuccess { + case Some(message) => { + val fetchDone = System.currentTimeMillis() + _remoteFetchTime += fetchDone - fetchStart + val bufferMessage = message.asInstanceOf[BufferMessage] + val blockMessageArray = BlockMessageArray.fromBufferMessage(bufferMessage) + for (blockMessage <- blockMessageArray) { + if (blockMessage.getType != BlockMessage.TYPE_GOT_BLOCK) { + throw new SparkException( + "Unexpected message " + blockMessage.getType + " received from " + cmId) + } + val blockId = blockMessage.getId + results.put(new FetchResult(blockId, sizeMap(blockId), + () => dataDeserialize(blockId, blockMessage.getData, serializer))) + _remoteBytesRead += req.size + logDebug("Got remote block " + blockId + " after " + Utils.getUsedTimeMs(startTime)) + } + } + case None => { + logError("Could not get block(s) from " + cmId) + for ((blockId, size) <- req.blocks) { + results.put(new FetchResult(blockId, -1, null)) + } + } + } + } + + protected def splitLocalRemoteBlocks(): ArrayBuffer[FetchRequest] = { + // Split local and remote blocks. Remote blocks are further split into FetchRequests of size + // at most maxBytesInFlight in order to limit the amount of data in flight. + val remoteRequests = new ArrayBuffer[FetchRequest] + for ((address, blockInfos) <- blocksByAddress) { + if (address == blockManagerId) { + numLocal = blockInfos.size + // Filter out zero-sized blocks + localBlocksToFetch ++= blockInfos.filter(_._2 != 0).map(_._1) + _numBlocksToFetch += localBlocksToFetch.size + } else { + numRemote += blockInfos.size + // Make our requests at least maxBytesInFlight / 5 in length; the reason to keep them + // smaller than maxBytesInFlight is to allow multiple, parallel fetches from up to 5 + // nodes, rather than blocking on reading output from one node. + val minRequestSize = math.max(maxBytesInFlight / 5, 1L) + logInfo("maxBytesInFlight: " + maxBytesInFlight + ", minRequest: " + minRequestSize) + val iterator = blockInfos.iterator + var curRequestSize = 0L + var curBlocks = new ArrayBuffer[(String, Long)] + while (iterator.hasNext) { + val (blockId, size) = iterator.next() + // Skip empty blocks + if (size > 0) { + curBlocks += ((blockId, size)) + remoteBlocksToFetch += blockId + _numBlocksToFetch += 1 + curRequestSize += size + } else if (size < 0) { + throw new BlockException(blockId, "Negative block size " + size) + } + if (curRequestSize >= minRequestSize) { + // Add this FetchRequest + remoteRequests += new FetchRequest(address, curBlocks) + curRequestSize = 0 + curBlocks = new ArrayBuffer[(String, Long)] + } + } + // Add in the final request + if (!curBlocks.isEmpty) { + remoteRequests += new FetchRequest(address, curBlocks) + } + } + } + logInfo("Getting " + _numBlocksToFetch + " non-zero-bytes blocks out of " + + totalBlocks + " blocks") + remoteRequests + } + + protected def getLocalBlocks() { + // Get the local blocks while remote blocks are being fetched. Note that it's okay to do + // these all at once because they will just memory-map some files, so they won't consume + // any memory that might exceed our maxBytesInFlight + for (id <- localBlocksToFetch) { + getLocalFromDisk(id, serializer) match { + case Some(iter) => { + // Pass 0 as size since it's not in flight + results.put(new FetchResult(id, 0, () => iter)) + logDebug("Got local block " + id) + } + case None => { + throw new BlockException(id, "Could not get block " + id + " from local machine") + } + } + } + } + + override def initialize() { + // Split local and remote blocks. + val remoteRequests = splitLocalRemoteBlocks() + // Add the remote requests into our queue in a random order + fetchRequests ++= Utils.randomize(remoteRequests) + + // Send out initial requests for blocks, up to our maxBytesInFlight + while (!fetchRequests.isEmpty && + (bytesInFlight == 0 || bytesInFlight + fetchRequests.front.size <= maxBytesInFlight)) { + sendRequest(fetchRequests.dequeue()) + } + + val numGets = remoteRequests.size - fetchRequests.size + logInfo("Started " + numGets + " remote gets in " + Utils.getUsedTimeMs(startTime)) + + // Get Local Blocks + startTime = System.currentTimeMillis + getLocalBlocks() + logDebug("Got local blocks in " + Utils.getUsedTimeMs(startTime) + " ms") + } + + //an iterator that will read fetched blocks off the queue as they arrive. + @volatile protected var resultsGotten = 0 + + override def hasNext: Boolean = resultsGotten < _numBlocksToFetch + + override def next(): (String, Option[Iterator[Any]]) = { + resultsGotten += 1 + val startFetchWait = System.currentTimeMillis() + val result = results.take() + val stopFetchWait = System.currentTimeMillis() + _fetchWaitTime += (stopFetchWait - startFetchWait) + if (! result.failed) bytesInFlight -= result.size + while (!fetchRequests.isEmpty && + (bytesInFlight == 0 || bytesInFlight + fetchRequests.front.size <= maxBytesInFlight)) { + sendRequest(fetchRequests.dequeue()) + } + (result.blockId, if (result.failed) None else Some(result.deserialize())) + } + + // Implementing BlockFetchTracker trait. + override def totalBlocks: Int = numLocal + numRemote + override def numLocalBlocks: Int = numLocal + override def numRemoteBlocks: Int = numRemote + override def remoteFetchTime: Long = _remoteFetchTime + override def fetchWaitTime: Long = _fetchWaitTime + override def remoteBytesRead: Long = _remoteBytesRead + } + // End of BasicBlockFetcherIterator + + class NettyBlockFetcherIterator( + blockManager: BlockManager, + blocksByAddress: Seq[(BlockManagerId, Seq[(String, Long)])], + serializer: Serializer) + extends BasicBlockFetcherIterator(blockManager, blocksByAddress, serializer) { + + import blockManager._ + + val fetchRequestsSync = new LinkedBlockingQueue[FetchRequest] + + private def startCopiers(numCopiers: Int): List[_ <: Thread] = { + (for ( i <- Range(0,numCopiers) ) yield { + val copier = new Thread { + override def run(){ + try { + while(!isInterrupted && !fetchRequestsSync.isEmpty) { + sendRequest(fetchRequestsSync.take()) + } + } catch { + case x: InterruptedException => logInfo("Copier Interrupted") + //case _ => throw new SparkException("Exception Throw in Shuffle Copier") + } + } + } + copier.start + copier + }).toList + } + + // keep this to interrupt the threads when necessary + private def stopCopiers() { + for (copier <- copiers) { + copier.interrupt() + } + } + + override protected def sendRequest(req: FetchRequest) { + + def putResult(blockId: String, blockSize: Long, blockData: ByteBuf) { + val fetchResult = new FetchResult(blockId, blockSize, + () => dataDeserialize(blockId, blockData.nioBuffer, serializer)) + results.put(fetchResult) + } + + logDebug("Sending request for %d blocks (%s) from %s".format( + req.blocks.size, Utils.memoryBytesToString(req.size), req.address.host)) + val cmId = new ConnectionManagerId(req.address.host, req.address.nettyPort) + val cpier = new ShuffleCopier + cpier.getBlocks(cmId, req.blocks, putResult) + logDebug("Sent request for remote blocks " + req.blocks + " from " + req.address.host ) + } + + private var copiers: List[_ <: Thread] = null + + override def initialize() { + // Split Local Remote Blocks and set numBlocksToFetch + val remoteRequests = splitLocalRemoteBlocks() + // Add the remote requests into our queue in a random order + for (request <- Utils.randomize(remoteRequests)) { + fetchRequestsSync.put(request) + } + + copiers = startCopiers(System.getProperty("spark.shuffle.copier.threads", "6").toInt) + logInfo("Started " + fetchRequestsSync.size + " remote gets in " + + Utils.getUsedTimeMs(startTime)) + + // Get Local Blocks + startTime = System.currentTimeMillis + getLocalBlocks() + logDebug("Got local blocks in " + Utils.getUsedTimeMs(startTime) + " ms") + } + + override def next(): (String, Option[Iterator[Any]]) = { + resultsGotten += 1 + val result = results.take() + // If all the results has been retrieved, copiers will exit automatically + (result.blockId, if (result.failed) None else Some(result.deserialize())) + } + } + // End of NettyBlockFetcherIterator +} diff --git a/core/src/main/scala/spark/storage/BlockManager.scala b/core/src/main/scala/spark/storage/BlockManager.scala index 6e861ac734..e4ffa57ad2 100644 --- a/core/src/main/scala/spark/storage/BlockManager.scala +++ b/core/src/main/scala/spark/storage/BlockManager.scala @@ -1,11 +1,26 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.io.{InputStream, OutputStream} import java.nio.{ByteBuffer, MappedByteBuffer} -import java.util.concurrent.{ConcurrentHashMap, LinkedBlockingQueue} -import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet, Queue} -import scala.collection.JavaConversions._ +import scala.collection.mutable.{HashMap, ArrayBuffer, HashSet} import akka.actor.{ActorSystem, Cancellable, Props} import akka.dispatch.{Await, Future} @@ -16,7 +31,7 @@ import com.ning.compress.lzf.{LZFInputStream, LZFOutputStream} import it.unimi.dsi.fastutil.io.FastByteArrayOutputStream -import spark.{Logging, SizeEstimator, SparkEnv, SparkException, Utils} +import spark.{Logging, SparkEnv, SparkException, Utils} import spark.network._ import spark.serializer.Serializer import spark.util.{ByteBufferInputStream, IdGenerator, MetadataCleaner, TimeStampedHashMap} @@ -24,16 +39,11 @@ import spark.util.{ByteBufferInputStream, IdGenerator, MetadataCleaner, TimeStam import sun.nio.ch.DirectBuffer -private[spark] -case class BlockException(blockId: String, message: String, ex: Exception = null) -extends Exception(message) - -private[spark] -class BlockManager( +private[spark] class BlockManager( executorId: String, actorSystem: ActorSystem, val master: BlockManagerMaster, - val serializer: Serializer, + val defaultSerializer: Serializer, maxMemory: Long) extends Logging { @@ -92,17 +102,26 @@ class BlockManager( } } + val shuffleBlockManager = new ShuffleBlockManager(this) + private val blockInfo = new TimeStampedHashMap[String, BlockInfo] private[storage] val memoryStore: BlockStore = new MemoryStore(this, maxMemory) - private[storage] val diskStore: BlockStore = + private[storage] val diskStore: DiskStore = new DiskStore(this, System.getProperty("spark.local.dir", System.getProperty("java.io.tmpdir"))) + // If we use Netty for shuffle, start a new Netty-based shuffle sender service. + private val nettyPort: Int = { + val useNetty = System.getProperty("spark.shuffle.use.netty", "false").toBoolean + val nettyPortConfig = System.getProperty("spark.shuffle.sender.port", "0").toInt + if (useNetty) diskStore.startShuffleBlockSender(nettyPortConfig) else 0 + } + val connectionManager = new ConnectionManager(0) implicit val futureExecContext = connectionManager.futureExecContext val blockManagerId = BlockManagerId( - executorId, connectionManager.id.host, connectionManager.id.port) + executorId, connectionManager.id.host, connectionManager.id.port, nettyPort) // Max megabytes of data to keep in flight per reducer (to avoid over-allocating memory // for receiving shuffle outputs) @@ -270,26 +289,24 @@ class BlockManager( } } - /** - * Get locations of the block. + * Get locations of an array of blocks. */ - def getLocations(blockId: String): Seq[String] = { + def getLocationBlockIds(blockIds: Array[String]): Array[Seq[BlockManagerId]] = { val startTimeMs = System.currentTimeMillis - var managers = master.getLocations(blockId) - val locations = managers.map(_.hostPort) - logDebug("Got block locations in " + Utils.getUsedTimeMs(startTimeMs)) - return locations + val locations = master.getLocations(blockIds).toArray + logDebug("Got multiple block location in " + Utils.getUsedTimeMs(startTimeMs)) + locations } /** - * Get locations of an array of blocks. + * A short-circuited method to get blocks directly from disk. This is used for getting + * shuffle blocks. It is safe to do so without a lock on block info since disk store + * never deletes (recent) items. */ - def getLocations(blockIds: Array[String]): Array[Seq[String]] = { - val startTimeMs = System.currentTimeMillis - val locations = master.getLocations(blockIds).map(_.map(_.hostPort).toSeq).toArray - logDebug("Got multiple block location in " + Utils.getUsedTimeMs(startTimeMs)) - return locations + def getLocalFromDisk(blockId: String, serializer: Serializer): Option[Iterator[Any]] = { + diskStore.getValues(blockId, serializer).orElse( + sys.error("Block " + blockId + " not found on disk, though it should be")) } /** @@ -297,18 +314,6 @@ class BlockManager( */ def getLocal(blockId: String): Option[Iterator[Any]] = { logDebug("Getting local block " + blockId) - - // As an optimization for map output fetches, if the block is for a shuffle, return it - // without acquiring a lock; the disk store never deletes (recent) items so this should work - if (blockId.startsWith("shuffle_")) { - return diskStore.getValues(blockId) match { - case Some(iterator) => - Some(iterator) - case None => - throw new Exception("Block " + blockId + " not found on disk, though it should be") - } - } - val info = blockInfo.get(blockId).orNull if (info != null) { info.synchronized { @@ -394,7 +399,7 @@ class BlockManager( // As an optimization for map output fetches, if the block is for a shuffle, return it // without acquiring a lock; the disk store never deletes (recent) items so this should work - if (blockId.startsWith("shuffle_")) { + if (ShuffleBlockManager.isShuffle(blockId)) { return diskStore.getBytes(blockId) match { case Some(bytes) => Some(bytes) @@ -496,9 +501,19 @@ class BlockManager( * fashion as they're received. Expects a size in bytes to be provided for each block fetched, * so that we can control the maxMegabytesInFlight for the fetch. */ - def getMultiple(blocksByAddress: Seq[(BlockManagerId, Seq[(String, Long)])]) + def getMultiple( + blocksByAddress: Seq[(BlockManagerId, Seq[(String, Long)])], serializer: Serializer) : BlockFetcherIterator = { - return new BlockFetcherIterator(this, blocksByAddress) + + val iter = + if (System.getProperty("spark.shuffle.use.netty", "false").toBoolean) { + new BlockFetcherIterator.NettyBlockFetcherIterator(this, blocksByAddress, serializer) + } else { + new BlockFetcherIterator.BasicBlockFetcherIterator(this, blocksByAddress, serializer) + } + + iter.initialize() + iter } def put(blockId: String, values: Iterator[Any], level: StorageLevel, tellMaster: Boolean) @@ -509,6 +524,22 @@ class BlockManager( } /** + * A short circuited method to get a block writer that can write data directly to disk. + * This is currently used for writing shuffle files out. Callers should handle error + * cases. + */ + def getDiskBlockWriter(blockId: String, serializer: Serializer, bufferSize: Int) + : BlockObjectWriter = { + val writer = diskStore.getBlockWriter(blockId, serializer, bufferSize) + writer.registerCloseEventHandler(() => { + val myInfo = new BlockInfo(StorageLevel.DISK_ONLY, false) + blockInfo.put(blockId, myInfo) + myInfo.markReady(writer.size()) + }) + writer + } + + /** * Put a new block of values to the block manager. Returns its (estimated) size in bytes. */ def put(blockId: String, values: ArrayBuffer[Any], level: StorageLevel, @@ -607,7 +638,6 @@ class BlockManager( } logDebug("Put block " + blockId + " locally took " + Utils.getUsedTimeMs(startTimeMs)) - // Replicate block if required if (level.replication > 1) { val remoteStartTime = System.currentTimeMillis @@ -813,9 +843,23 @@ class BlockManager( } /** + * Remove all blocks belonging to the given RDD. + * @return The number of blocks removed. + */ + def removeRdd(rddId: Int): Int = { + // TODO: Instead of doing a linear scan on the blockInfo map, create another map that maps + // from RDD.id to blocks. + logInfo("Removing RDD " + rddId) + val rddPrefix = "rdd_" + rddId + "_" + val blocksToRemove = blockInfo.filter(_._1.startsWith(rddPrefix)).map(_._1) + blocksToRemove.foreach(blockId => removeBlock(blockId, false)) + blocksToRemove.size + } + + /** * Remove a block from both memory and disk. */ - def removeBlock(blockId: String) { + def removeBlock(blockId: String, tellMaster: Boolean = true) { logInfo("Removing block " + blockId) val info = blockInfo.get(blockId).orNull if (info != null) info.synchronized { @@ -827,7 +871,7 @@ class BlockManager( "the disk or memory store") } blockInfo.remove(blockId) - if (info.tellMaster) { + if (tellMaster && info.tellMaster) { reportBlockStatus(blockId, info) } } else { @@ -860,7 +904,7 @@ class BlockManager( } def shouldCompress(blockId: String): Boolean = { - if (blockId.startsWith("shuffle_")) { + if (ShuffleBlockManager.isShuffle(blockId)) { compressShuffle } else if (blockId.startsWith("broadcast_")) { compressBroadcast @@ -875,7 +919,11 @@ class BlockManager( * Wrap an output stream for compression if block compression is enabled for its block type */ def wrapForCompression(blockId: String, s: OutputStream): OutputStream = { - if (shouldCompress(blockId)) new LZFOutputStream(s) else s + if (shouldCompress(blockId)) { + (new LZFOutputStream(s)).setFinishBlockOnFlush(true) + } else { + s + } } /** @@ -885,7 +933,10 @@ class BlockManager( if (shouldCompress(blockId)) new LZFInputStream(s) else s } - def dataSerialize(blockId: String, values: Iterator[Any]): ByteBuffer = { + def dataSerialize( + blockId: String, + values: Iterator[Any], + serializer: Serializer = defaultSerializer): ByteBuffer = { val byteStream = new FastByteArrayOutputStream(4096) val ser = serializer.newInstance() ser.serializeStream(wrapForCompression(blockId, byteStream)).writeAll(values).close() @@ -897,7 +948,10 @@ class BlockManager( * Deserializes a ByteBuffer into an iterator of values and disposes of it when the end of * the iterator is reached. */ - def dataDeserialize(blockId: String, bytes: ByteBuffer): Iterator[Any] = { + def dataDeserialize( + blockId: String, + bytes: ByteBuffer, + serializer: Serializer = defaultSerializer): Iterator[Any] = { bytes.rewind() val stream = wrapForCompression(blockId, new ByteBufferInputStream(bytes, true)) serializer.newInstance().deserializeStream(stream).asIterator @@ -917,8 +971,8 @@ class BlockManager( } } -private[spark] -object BlockManager extends Logging { + +private[spark] object BlockManager extends Logging { val ID_GENERATOR = new IdGenerator @@ -928,7 +982,7 @@ object BlockManager extends Logging { } def getHeartBeatFrequencyFromSystemProperties: Long = - System.getProperty("spark.storage.blockManagerHeartBeatMs", "5000").toLong + System.getProperty("spark.storage.blockManagerTimeoutIntervalMs", "60000").toLong / 4 def getDisableHeartBeatsForTesting: Boolean = System.getProperty("spark.test.disableBlockManagerHeartBeat", "false").toBoolean @@ -947,177 +1001,44 @@ object BlockManager extends Logging { } } } -} - -class BlockFetcherIterator( - private val blockManager: BlockManager, - val blocksByAddress: Seq[(BlockManagerId, Seq[(String, Long)])] -) extends Iterator[(String, Option[Iterator[Any]])] with Logging with BlockFetchTracker { - - import blockManager._ - - private var _remoteBytesRead = 0l - private var _remoteFetchTime = 0l - private var _fetchWaitTime = 0l - - if (blocksByAddress == null) { - throw new IllegalArgumentException("BlocksByAddress is null") - } - val totalBlocks = blocksByAddress.map(_._2.size).sum - logDebug("Getting " + totalBlocks + " blocks") - var startTime = System.currentTimeMillis - val localBlockIds = new ArrayBuffer[String]() - val remoteBlockIds = new HashSet[String]() - - // A result of a fetch. Includes the block ID, size in bytes, and a function to deserialize - // the block (since we want all deserializaton to happen in the calling thread); can also - // represent a fetch failure if size == -1. - class FetchResult(val blockId: String, val size: Long, val deserialize: () => Iterator[Any]) { - def failed: Boolean = size == -1 - } - - // A queue to hold our results. - val results = new LinkedBlockingQueue[FetchResult] - // A request to fetch one or more blocks, complete with their sizes - class FetchRequest(val address: BlockManagerId, val blocks: Seq[(String, Long)]) { - val size = blocks.map(_._2).sum - } + def blockIdsToExecutorLocations(blockIds: Array[String], env: SparkEnv, blockManagerMaster: BlockManagerMaster = null): HashMap[String, List[String]] = { + // env == null and blockManagerMaster != null is used in tests + assert (env != null || blockManagerMaster != null) + val locationBlockIds: Seq[Seq[BlockManagerId]] = + if (env != null) { + env.blockManager.getLocationBlockIds(blockIds) + } else { + blockManagerMaster.getLocations(blockIds) + } - // Queue of fetch requests to issue; we'll pull requests off this gradually to make sure that - // the number of bytes in flight is limited to maxBytesInFlight - val fetchRequests = new Queue[FetchRequest] + // Convert from block master locations to executor locations (we need that for task scheduling) + val executorLocations = new HashMap[String, List[String]]() + for (i <- 0 until blockIds.length) { + val blockId = blockIds(i) + val blockLocations = locationBlockIds(i) - // Current bytes in flight from our requests - var bytesInFlight = 0L + val executors = new HashSet[String]() - def sendRequest(req: FetchRequest) { - logDebug("Sending request for %d blocks (%s) from %s".format( - req.blocks.size, Utils.memoryBytesToString(req.size), req.address.hostPort)) - val cmId = new ConnectionManagerId(req.address.host, req.address.port) - val blockMessageArray = new BlockMessageArray(req.blocks.map { - case (blockId, size) => BlockMessage.fromGetBlock(GetBlock(blockId)) - }) - bytesInFlight += req.size - val sizeMap = req.blocks.toMap // so we can look up the size of each blockID - val fetchStart = System.currentTimeMillis() - val future = connectionManager.sendMessageReliably(cmId, blockMessageArray.toBufferMessage) - future.onSuccess { - case Some(message) => { - val fetchDone = System.currentTimeMillis() - _remoteFetchTime += fetchDone - fetchStart - val bufferMessage = message.asInstanceOf[BufferMessage] - val blockMessageArray = BlockMessageArray.fromBufferMessage(bufferMessage) - for (blockMessage <- blockMessageArray) { - if (blockMessage.getType != BlockMessage.TYPE_GOT_BLOCK) { - throw new SparkException( - "Unexpected message " + blockMessage.getType + " received from " + cmId) - } - val blockId = blockMessage.getId - results.put(new FetchResult( - blockId, sizeMap(blockId), () => dataDeserialize(blockId, blockMessage.getData))) - _remoteBytesRead += req.size - logDebug("Got remote block " + blockId + " after " + Utils.getUsedTimeMs(startTime)) + if (env != null) { + for (bkLocation <- blockLocations) { + val executorHostPort = env.resolveExecutorIdToHostPort(bkLocation.executorId, bkLocation.host) + executors += executorHostPort + // logInfo("bkLocation = " + bkLocation + ", executorHostPort = " + executorHostPort) } - } - case None => { - logError("Could not get block(s) from " + cmId) - for ((blockId, size) <- req.blocks) { - results.put(new FetchResult(blockId, -1, null)) + } else { + // Typically while testing, etc - revert to simply using host. + for (bkLocation <- blockLocations) { + executors += bkLocation.host + // logInfo("bkLocation = " + bkLocation + ", executorHostPort = " + executorHostPort) } } - } - } - // Split local and remote blocks. Remote blocks are further split into FetchRequests of size - // at most maxBytesInFlight in order to limit the amount of data in flight. - val remoteRequests = new ArrayBuffer[FetchRequest] - for ((address, blockInfos) <- blocksByAddress) { - if (address == blockManagerId) { - localBlockIds ++= blockInfos.map(_._1) - } else { - remoteBlockIds ++= blockInfos.map(_._1) - // Make our requests at least maxBytesInFlight / 5 in length; the reason to keep them - // smaller than maxBytesInFlight is to allow multiple, parallel fetches from up to 5 - // nodes, rather than blocking on reading output from one node. - val minRequestSize = math.max(maxBytesInFlight / 5, 1L) - logInfo("maxBytesInFlight: " + maxBytesInFlight + ", minRequest: " + minRequestSize) - val iterator = blockInfos.iterator - var curRequestSize = 0L - var curBlocks = new ArrayBuffer[(String, Long)] - while (iterator.hasNext) { - val (blockId, size) = iterator.next() - curBlocks += ((blockId, size)) - curRequestSize += size - if (curRequestSize >= minRequestSize) { - // Add this FetchRequest - remoteRequests += new FetchRequest(address, curBlocks) - curRequestSize = 0 - curBlocks = new ArrayBuffer[(String, Long)] - } - } - // Add in the final request - if (!curBlocks.isEmpty) { - remoteRequests += new FetchRequest(address, curBlocks) - } + executorLocations.put(blockId, executors.toSeq.toList) } - } - // Add the remote requests into our queue in a random order - fetchRequests ++= Utils.randomize(remoteRequests) - // Send out initial requests for blocks, up to our maxBytesInFlight - while (!fetchRequests.isEmpty && - (bytesInFlight == 0 || bytesInFlight + fetchRequests.front.size <= maxBytesInFlight)) { - sendRequest(fetchRequests.dequeue()) + executorLocations } - val numGets = remoteBlockIds.size - fetchRequests.size - logInfo("Started " + numGets + " remote gets in " + Utils.getUsedTimeMs(startTime)) - - // Get the local blocks while remote blocks are being fetched. Note that it's okay to do - // these all at once because they will just memory-map some files, so they won't consume - // any memory that might exceed our maxBytesInFlight - startTime = System.currentTimeMillis - for (id <- localBlockIds) { - getLocal(id) match { - case Some(iter) => { - results.put(new FetchResult(id, 0, () => iter)) // Pass 0 as size since it's not in flight - logDebug("Got local block " + id) - } - case None => { - throw new BlockException(id, "Could not get block " + id + " from local machine") - } - } - } - logDebug("Got local blocks in " + Utils.getUsedTimeMs(startTime) + " ms") - - //an iterator that will read fetched blocks off the queue as they arrive. - @volatile private var resultsGotten = 0 - - def hasNext: Boolean = resultsGotten < totalBlocks - - def next(): (String, Option[Iterator[Any]]) = { - resultsGotten += 1 - val startFetchWait = System.currentTimeMillis() - val result = results.take() - val stopFetchWait = System.currentTimeMillis() - _fetchWaitTime += (stopFetchWait - startFetchWait) - if (! result.failed) bytesInFlight -= result.size - while (!fetchRequests.isEmpty && - (bytesInFlight == 0 || bytesInFlight + fetchRequests.front.size <= maxBytesInFlight)) { - sendRequest(fetchRequests.dequeue()) - } - (result.blockId, if (result.failed) None else Some(result.deserialize())) - } - - - //methods to profile the block fetching - def numLocalBlocks = localBlockIds.size - def numRemoteBlocks = remoteBlockIds.size - - def remoteFetchTime = _remoteFetchTime - def fetchWaitTime = _fetchWaitTime - - def remoteBytesRead = _remoteBytesRead - } + diff --git a/core/src/main/scala/spark/storage/BlockManagerId.scala b/core/src/main/scala/spark/storage/BlockManagerId.scala index f4a2181490..b36a6176c0 100644 --- a/core/src/main/scala/spark/storage/BlockManagerId.scala +++ b/core/src/main/scala/spark/storage/BlockManagerId.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.io.{Externalizable, IOException, ObjectInput, ObjectOutput} @@ -7,18 +24,19 @@ import spark.Utils /** * This class represent an unique identifier for a BlockManager. * The first 2 constructors of this class is made private to ensure that - * BlockManagerId objects can be created only using the factory method in - * [[spark.storage.BlockManager$]]. This allows de-duplication of ID objects. + * BlockManagerId objects can be created only using the apply method in + * the companion object. This allows de-duplication of ID objects. * Also, constructor parameters are private to ensure that parameters cannot * be modified from outside this class. */ private[spark] class BlockManagerId private ( private var executorId_ : String, private var host_ : String, - private var port_ : Int + private var port_ : Int, + private var nettyPort_ : Int ) extends Externalizable { - private def this() = this(null, null, 0) // For deserialization only + private def this() = this(null, null, 0, 0) // For deserialization only def executorId: String = executorId_ @@ -39,28 +57,32 @@ private[spark] class BlockManagerId private ( def port: Int = port_ + def nettyPort: Int = nettyPort_ + override def writeExternal(out: ObjectOutput) { out.writeUTF(executorId_) out.writeUTF(host_) out.writeInt(port_) + out.writeInt(nettyPort_) } override def readExternal(in: ObjectInput) { executorId_ = in.readUTF() host_ = in.readUTF() port_ = in.readInt() + nettyPort_ = in.readInt() } @throws(classOf[IOException]) private def readResolve(): Object = BlockManagerId.getCachedBlockManagerId(this) - override def toString = "BlockManagerId(%s, %s, %d)".format(executorId, host, port) + override def toString = "BlockManagerId(%s, %s, %d, %d)".format(executorId, host, port, nettyPort) - override def hashCode: Int = (executorId.hashCode * 41 + host.hashCode) * 41 + port + override def hashCode: Int = (executorId.hashCode * 41 + host.hashCode) * 41 + port + nettyPort override def equals(that: Any) = that match { case id: BlockManagerId => - executorId == id.executorId && port == id.port && host == id.host + executorId == id.executorId && port == id.port && host == id.host && nettyPort == id.nettyPort case _ => false } @@ -69,8 +91,17 @@ private[spark] class BlockManagerId private ( private[spark] object BlockManagerId { - def apply(execId: String, host: String, port: Int) = - getCachedBlockManagerId(new BlockManagerId(execId, host, port)) + /** + * Returns a [[spark.storage.BlockManagerId]] for the given configuraiton. + * + * @param execId ID of the executor. + * @param host Host name of the block manager. + * @param port Port of the block manager. + * @param nettyPort Optional port for the Netty-based shuffle sender. + * @return A new [[spark.storage.BlockManagerId]]. + */ + def apply(execId: String, host: String, port: Int, nettyPort: Int) = + getCachedBlockManagerId(new BlockManagerId(execId, host, port, nettyPort)) def apply(in: ObjectInput) = { val obj = new BlockManagerId() diff --git a/core/src/main/scala/spark/storage/BlockManagerMaster.scala b/core/src/main/scala/spark/storage/BlockManagerMaster.scala index 6fae62d373..3186f7c85b 100644 --- a/core/src/main/scala/spark/storage/BlockManagerMaster.scala +++ b/core/src/main/scala/spark/storage/BlockManagerMaster.scala @@ -1,19 +1,29 @@ -package spark.storage - -import java.io._ -import java.util.{HashMap => JHashMap} +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ -import scala.collection.JavaConverters._ -import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} -import scala.util.Random +package spark.storage -import akka.actor.{Actor, ActorRef, ActorSystem, Props} -import akka.dispatch.Await +import akka.actor.ActorRef +import akka.dispatch.{Await, Future} import akka.pattern.ask -import akka.util.{Duration, Timeout} -import akka.util.duration._ +import akka.util.Duration + +import spark.{Logging, SparkException} -import spark.{Logging, SparkException, Utils} private[spark] class BlockManagerMaster(var driverActor: ActorRef) extends Logging { @@ -88,6 +98,19 @@ private[spark] class BlockManagerMaster(var driverActor: ActorRef) extends Loggi } /** + * Remove all blocks belonging to the given RDD. + */ + def removeRdd(rddId: Int, blocking: Boolean) { + val future = askDriverWithReply[Future[Seq[Int]]](RemoveRdd(rddId)) + future onFailure { + case e: Throwable => logError("Failed to remove RDD " + rddId, e) + } + if (blocking) { + Await.result(future, timeout) + } + } + + /** * Return the memory status for each block manager, in the form of a map from * the block manager's id to two long values. The first value is the maximum * amount of memory allocated for the block manager, while the second is the @@ -98,7 +121,7 @@ private[spark] class BlockManagerMaster(var driverActor: ActorRef) extends Loggi } def getStorageStatus: Array[StorageStatus] = { - askDriverWithReply[ArrayBuffer[StorageStatus]](GetStorageStatus).toArray + askDriverWithReply[Array[StorageStatus]](GetStorageStatus) } /** Stop the driver actor, called only on the Spark driver node */ @@ -135,7 +158,7 @@ private[spark] class BlockManagerMaster(var driverActor: ActorRef) extends Loggi val future = driverActor.ask(message)(timeout) val result = Await.result(future, timeout) if (result == null) { - throw new Exception("BlockManagerMaster returned null") + throw new SparkException("BlockManagerMaster returned null") } return result.asInstanceOf[T] } catch { diff --git a/core/src/main/scala/spark/storage/BlockManagerMasterActor.scala b/core/src/main/scala/spark/storage/BlockManagerMasterActor.scala index 9b64f95df8..244000d952 100644 --- a/core/src/main/scala/spark/storage/BlockManagerMasterActor.scala +++ b/core/src/main/scala/spark/storage/BlockManagerMasterActor.scala @@ -1,16 +1,34 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.util.{HashMap => JHashMap} -import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} +import scala.collection.mutable import scala.collection.JavaConversions._ -import scala.util.Random import akka.actor.{Actor, ActorRef, Cancellable} -import akka.util.{Duration, Timeout} +import akka.dispatch.Future +import akka.pattern.ask +import akka.util.Duration import akka.util.duration._ -import spark.{Logging, Utils} +import spark.{Logging, Utils, SparkException} /** * BlockManagerMasterActor is an actor on the master node to track statuses of @@ -21,13 +39,16 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { // Mapping from block manager id to the block manager's information. private val blockManagerInfo = - new HashMap[BlockManagerId, BlockManagerMasterActor.BlockManagerInfo] + new mutable.HashMap[BlockManagerId, BlockManagerMasterActor.BlockManagerInfo] // Mapping from executor ID to block manager ID. - private val blockManagerIdByExecutor = new HashMap[String, BlockManagerId] + private val blockManagerIdByExecutor = new mutable.HashMap[String, BlockManagerId] // Mapping from block id to the set of block managers that have the block. - private val blockLocations = new JHashMap[String, Pair[Int, HashSet[BlockManagerId]]] + private val blockLocations = new JHashMap[String, mutable.HashSet[BlockManagerId]] + + val akkaTimeout = Duration.create( + System.getProperty("spark.akka.askTimeout", "10").toLong, "seconds") initLogging() @@ -35,7 +56,7 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { "" + (BlockManager.getHeartBeatFrequencyFromSystemProperties * 3)).toLong val checkTimeoutInterval = System.getProperty("spark.storage.blockManagerTimeoutIntervalMs", - "5000").toLong + "60000").toLong var timeoutCheckingTask: Cancellable = null @@ -50,28 +71,34 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { def receive = { case RegisterBlockManager(blockManagerId, maxMemSize, slaveActor) => register(blockManagerId, maxMemSize, slaveActor) + sender ! true case UpdateBlockInfo(blockManagerId, blockId, storageLevel, deserializedSize, size) => + // TODO: Ideally we want to handle all the message replies in receive instead of in the + // individual private methods. updateBlockInfo(blockManagerId, blockId, storageLevel, deserializedSize, size) case GetLocations(blockId) => - getLocations(blockId) + sender ! getLocations(blockId) case GetLocationsMultipleBlockIds(blockIds) => - getLocationsMultipleBlockIds(blockIds) + sender ! getLocationsMultipleBlockIds(blockIds) case GetPeers(blockManagerId, size) => - getPeersDeterministic(blockManagerId, size) - /*getPeers(blockManagerId, size)*/ + sender ! getPeers(blockManagerId, size) case GetMemoryStatus => - getMemoryStatus + sender ! memoryStatus case GetStorageStatus => - getStorageStatus + sender ! storageStatus + + case RemoveRdd(rddId) => + sender ! removeRdd(rddId) case RemoveBlock(blockId) => - removeBlock(blockId) + removeBlockFromWorkers(blockId) + sender ! true case RemoveExecutor(execId) => removeExecutor(execId) @@ -81,7 +108,7 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { logInfo("Stopping BlockManagerMaster") sender ! true if (timeoutCheckingTask != null) { - timeoutCheckingTask.cancel + timeoutCheckingTask.cancel() } context.stop(self) @@ -89,13 +116,36 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { expireDeadHosts() case HeartBeat(blockManagerId) => - heartBeat(blockManagerId) + sender ! heartBeat(blockManagerId) case other => - logInfo("Got unknown message: " + other) + logWarning("Got unknown message: " + other) + } + + private def removeRdd(rddId: Int): Future[Seq[Int]] = { + // First remove the metadata for the given RDD, and then asynchronously remove the blocks + // from the slaves. + + val prefix = "rdd_" + rddId + "_" + // Find all blocks for the given RDD, remove the block from both blockLocations and + // the blockManagerInfo that is tracking the blocks. + val blocks = blockLocations.keySet().filter(_.startsWith(prefix)) + blocks.foreach { blockId => + val bms: mutable.HashSet[BlockManagerId] = blockLocations.get(blockId) + bms.foreach(bm => blockManagerInfo.get(bm).foreach(_.removeBlock(blockId))) + blockLocations.remove(blockId) + } + + // Ask the slaves to remove the RDD, and put the result in a sequence of Futures. + // The dispatcher is used as an implicit argument into the Future sequence construction. + import context.dispatcher + val removeMsg = RemoveRdd(rddId) + Future.sequence(blockManagerInfo.values.map { bm => + bm.slaveActor.ask(removeMsg)(akkaTimeout).mapTo[Int] + }.toSeq) } - def removeBlockManager(blockManagerId: BlockManagerId) { + private def removeBlockManager(blockManagerId: BlockManagerId) { val info = blockManagerInfo(blockManagerId) // Remove the block manager from blockManagerIdByExecutor. @@ -106,7 +156,7 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { val iterator = info.blocks.keySet.iterator while (iterator.hasNext) { val blockId = iterator.next - val locations = blockLocations.get(blockId)._2 + val locations = blockLocations.get(blockId) locations -= blockManagerId if (locations.size == 0) { blockLocations.remove(locations) @@ -114,11 +164,11 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { } } - def expireDeadHosts() { + private def expireDeadHosts() { logTrace("Checking for hosts with no recent heart beats in BlockManagerMaster.") val now = System.currentTimeMillis() val minSeenTime = now - slaveTimeout - val toRemove = new HashSet[BlockManagerId] + val toRemove = new mutable.HashSet[BlockManagerId] for (info <- blockManagerInfo.values) { if (info.lastSeenMs < minSeenTime) { logWarning("Removing BlockManager " + info.blockManagerId + " with no recent heart beats: " + @@ -129,31 +179,26 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { toRemove.foreach(removeBlockManager) } - def removeExecutor(execId: String) { + private def removeExecutor(execId: String) { logInfo("Trying to remove executor " + execId + " from BlockManagerMaster.") blockManagerIdByExecutor.get(execId).foreach(removeBlockManager) - sender ! true } - def heartBeat(blockManagerId: BlockManagerId) { + private def heartBeat(blockManagerId: BlockManagerId): Boolean = { if (!blockManagerInfo.contains(blockManagerId)) { - if (blockManagerId.executorId == "<driver>" && !isLocal) { - sender ! true - } else { - sender ! false - } + blockManagerId.executorId == "<driver>" && !isLocal } else { blockManagerInfo(blockManagerId).updateLastSeenMs() - sender ! true + true } } // Remove a block from the slaves that have it. This can only be used to remove // blocks that the master knows about. - private def removeBlock(blockId: String) { - val block = blockLocations.get(blockId) - if (block != null) { - block._2.foreach { blockManagerId: BlockManagerId => + private def removeBlockFromWorkers(blockId: String) { + val locations = blockLocations.get(blockId) + if (locations != null) { + locations.foreach { blockManagerId: BlockManagerId => val blockManager = blockManagerInfo.get(blockManagerId) if (blockManager.isDefined) { // Remove the block from the slave's BlockManager. @@ -163,23 +208,20 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { } } } - sender ! true } // Return a map from the block manager id to max memory and remaining memory. - private def getMemoryStatus() { - val res = blockManagerInfo.map { case(blockManagerId, info) => + private def memoryStatus: Map[BlockManagerId, (Long, Long)] = { + blockManagerInfo.map { case(blockManagerId, info) => (blockManagerId, (info.maxMem, info.remainingMem)) }.toMap - sender ! res } - private def getStorageStatus() { - val res = blockManagerInfo.map { case(blockManagerId, info) => + private def storageStatus: Array[StorageStatus] = { + blockManagerInfo.map { case(blockManagerId, info) => import collection.JavaConverters._ StorageStatus(blockManagerId, info.maxMem, info.blocks.asScala.toMap) - } - sender ! res + }.toArray } private def register(id: BlockManagerId, maxMemSize: Long, slaveActor: ActorRef) { @@ -188,7 +230,8 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { } else if (!blockManagerInfo.contains(id)) { blockManagerIdByExecutor.get(id.executorId) match { case Some(manager) => - // A block manager of the same host name already exists + // A block manager of the same executor already exists. + // This should never happen. Let's just quit. logError("Got two different block manager registrations on " + id.executorId) System.exit(1) case None => @@ -197,7 +240,6 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { blockManagerInfo(id) = new BlockManagerMasterActor.BlockManagerInfo( id, System.currentTimeMillis(), maxMemSize, slaveActor) } - sender ! true } private def updateBlockInfo( @@ -226,12 +268,12 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { blockManagerInfo(blockManagerId).updateBlockInfo(blockId, storageLevel, memSize, diskSize) - var locations: HashSet[BlockManagerId] = null + var locations: mutable.HashSet[BlockManagerId] = null if (blockLocations.containsKey(blockId)) { - locations = blockLocations.get(blockId)._2 + locations = blockLocations.get(blockId) } else { - locations = new HashSet[BlockManagerId] - blockLocations.put(blockId, (storageLevel.replication, locations)) + locations = new mutable.HashSet[BlockManagerId] + blockLocations.put(blockId, locations) } if (storageLevel.isValid) { @@ -247,70 +289,24 @@ class BlockManagerMasterActor(val isLocal: Boolean) extends Actor with Logging { sender ! true } - private def getLocations(blockId: String) { - val startTimeMs = System.currentTimeMillis() - val tmp = " " + blockId + " " - if (blockLocations.containsKey(blockId)) { - var res: ArrayBuffer[BlockManagerId] = new ArrayBuffer[BlockManagerId] - res.appendAll(blockLocations.get(blockId)._2) - sender ! res.toSeq - } else { - var res: ArrayBuffer[BlockManagerId] = new ArrayBuffer[BlockManagerId] - sender ! res - } + private def getLocations(blockId: String): Seq[BlockManagerId] = { + if (blockLocations.containsKey(blockId)) blockLocations.get(blockId).toSeq else Seq.empty } - private def getLocationsMultipleBlockIds(blockIds: Array[String]) { - def getLocations(blockId: String): Seq[BlockManagerId] = { - val tmp = blockId - if (blockLocations.containsKey(blockId)) { - var res: ArrayBuffer[BlockManagerId] = new ArrayBuffer[BlockManagerId] - res.appendAll(blockLocations.get(blockId)._2) - return res.toSeq - } else { - var res: ArrayBuffer[BlockManagerId] = new ArrayBuffer[BlockManagerId] - return res.toSeq - } - } - - var res: ArrayBuffer[Seq[BlockManagerId]] = new ArrayBuffer[Seq[BlockManagerId]] - for (blockId <- blockIds) { - res.append(getLocations(blockId)) - } - sender ! res.toSeq + private def getLocationsMultipleBlockIds(blockIds: Array[String]): Seq[Seq[BlockManagerId]] = { + blockIds.map(blockId => getLocations(blockId)) } - private def getPeers(blockManagerId: BlockManagerId, size: Int) { - var peers: Array[BlockManagerId] = blockManagerInfo.keySet.toArray - var res: ArrayBuffer[BlockManagerId] = new ArrayBuffer[BlockManagerId] - res.appendAll(peers) - res -= blockManagerId - val rand = new Random(System.currentTimeMillis()) - while (res.length > size) { - res.remove(rand.nextInt(res.length)) - } - sender ! res.toSeq - } - - private def getPeersDeterministic(blockManagerId: BlockManagerId, size: Int) { - var peers: Array[BlockManagerId] = blockManagerInfo.keySet.toArray - var res: ArrayBuffer[BlockManagerId] = new ArrayBuffer[BlockManagerId] + private def getPeers(blockManagerId: BlockManagerId, size: Int): Seq[BlockManagerId] = { + val peers: Array[BlockManagerId] = blockManagerInfo.keySet.toArray val selfIndex = peers.indexOf(blockManagerId) if (selfIndex == -1) { - throw new Exception("Self index for " + blockManagerId + " not found") + throw new SparkException("Self index for " + blockManagerId + " not found") } // Note that this logic will select the same node multiple times if there aren't enough peers - var index = selfIndex - while (res.size < size) { - index += 1 - if (index == selfIndex) { - throw new Exception("More peer expected than available") - } - res += peers(index % peers.size) - } - sender ! res.toSeq + Array.tabulate[BlockManagerId](size) { i => peers((selfIndex + i + 1) % peers.length) }.toSeq } } @@ -384,6 +380,13 @@ object BlockManagerMasterActor { } } + def removeBlock(blockId: String) { + if (_blocks.containsKey(blockId)) { + _remainingMem += _blocks.get(blockId).memSize + _blocks.remove(blockId) + } + } + def remainingMem: Long = _remainingMem def lastSeenMs: Long = _lastSeenMs diff --git a/core/src/main/scala/spark/storage/BlockManagerMessages.scala b/core/src/main/scala/spark/storage/BlockManagerMessages.scala index cff48d9909..01de4ccb8f 100644 --- a/core/src/main/scala/spark/storage/BlockManagerMessages.scala +++ b/core/src/main/scala/spark/storage/BlockManagerMessages.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.io.{Externalizable, ObjectInput, ObjectOutput} @@ -16,6 +33,9 @@ sealed trait ToBlockManagerSlave private[spark] case class RemoveBlock(blockId: String) extends ToBlockManagerSlave +// Remove all blocks belonging to a specific RDD. +private[spark] case class RemoveRdd(rddId: Int) extends ToBlockManagerSlave + ////////////////////////////////////////////////////////////////////////////////// // Messages from slaves to the master. diff --git a/core/src/main/scala/spark/storage/BlockManagerSlaveActor.scala b/core/src/main/scala/spark/storage/BlockManagerSlaveActor.scala index f570cdc52d..45cffad810 100644 --- a/core/src/main/scala/spark/storage/BlockManagerSlaveActor.scala +++ b/core/src/main/scala/spark/storage/BlockManagerSlaveActor.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import akka.actor.Actor @@ -11,6 +28,12 @@ import spark.{Logging, SparkException, Utils} */ class BlockManagerSlaveActor(blockManager: BlockManager) extends Actor { override def receive = { - case RemoveBlock(blockId) => blockManager.removeBlock(blockId) + + case RemoveBlock(blockId) => + blockManager.removeBlock(blockId) + + case RemoveRdd(rddId) => + val numBlocksRemoved = blockManager.removeRdd(rddId) + sender ! numBlocksRemoved } } diff --git a/core/src/main/scala/spark/storage/BlockManagerSource.scala b/core/src/main/scala/spark/storage/BlockManagerSource.scala new file mode 100644 index 0000000000..4faa715c94 --- /dev/null +++ b/core/src/main/scala/spark/storage/BlockManagerSource.scala @@ -0,0 +1,48 @@ +package spark.storage + +import com.codahale.metrics.{Gauge,MetricRegistry} + +import spark.metrics.source.Source +import spark.storage._ + +private[spark] class BlockManagerSource(val blockManager: BlockManager) extends Source { + val metricRegistry = new MetricRegistry() + val sourceName = "BlockManager" + + metricRegistry.register(MetricRegistry.name("memory", "maxMem", "MBytes"), new Gauge[Long] { + override def getValue: Long = { + val storageStatusList = blockManager.master.getStorageStatus + val maxMem = storageStatusList.map(_.maxMem).reduce(_ + _) + maxMem / 1024 / 1024 + } + }) + + metricRegistry.register(MetricRegistry.name("memory", "remainingMem", "MBytes"), new Gauge[Long] { + override def getValue: Long = { + val storageStatusList = blockManager.master.getStorageStatus + val remainingMem = storageStatusList.map(_.memRemaining).reduce(_ + _) + remainingMem / 1024 / 1024 + } + }) + + metricRegistry.register(MetricRegistry.name("memory", "memUsed", "MBytes"), new Gauge[Long] { + override def getValue: Long = { + val storageStatusList = blockManager.master.getStorageStatus + val maxMem = storageStatusList.map(_.maxMem).reduce(_ + _) + val remainingMem = storageStatusList.map(_.memRemaining).reduce(_ + _) + (maxMem - remainingMem) / 1024 / 1024 + } + }) + + metricRegistry.register(MetricRegistry.name("disk", "diskSpaceUsed", "MBytes"), new Gauge[Long] { + override def getValue: Long = { + val storageStatusList = blockManager.master.getStorageStatus + val diskSpaceUsed = storageStatusList + .flatMap(_.blocks.values.map(_.diskSize)) + .reduceOption(_ + _) + .getOrElse(0L) + + diskSpaceUsed / 1024 / 1024 + } + }) +} diff --git a/core/src/main/scala/spark/storage/BlockManagerUI.scala b/core/src/main/scala/spark/storage/BlockManagerUI.scala deleted file mode 100644 index 07da572044..0000000000 --- a/core/src/main/scala/spark/storage/BlockManagerUI.scala +++ /dev/null @@ -1,76 +0,0 @@ -package spark.storage - -import akka.actor.{ActorRef, ActorSystem} -import akka.util.Duration -import akka.util.duration._ -import cc.spray.typeconversion.TwirlSupport._ -import cc.spray.Directives -import spark.{Logging, SparkContext} -import spark.util.AkkaUtils -import spark.Utils - - -/** - * Web UI server for the BlockManager inside each SparkContext. - */ -private[spark] -class BlockManagerUI(val actorSystem: ActorSystem, blockManagerMaster: ActorRef, sc: SparkContext) - extends Directives with Logging { - - val STATIC_RESOURCE_DIR = "spark/deploy/static" - - implicit val timeout = Duration.create(System.getProperty("spark.akka.askTimeout", "10").toLong, "seconds") - - /** Start a HTTP server to run the Web interface */ - def start() { - try { - val port = if (System.getProperty("spark.ui.port") != null) { - System.getProperty("spark.ui.port").toInt - } else { - // TODO: Unfortunately, it's not possible to pass port 0 to spray and figure out which - // random port it bound to, so we have to try to find a local one by creating a socket. - Utils.findFreePort() - } - AkkaUtils.startSprayServer(actorSystem, "0.0.0.0", port, handler, "BlockManagerHTTPServer") - logInfo("Started BlockManager web UI at http://%s:%d".format(Utils.localHostName(), port)) - } catch { - case e: Exception => - logError("Failed to create BlockManager WebUI", e) - System.exit(1) - } - } - - val handler = { - get { - path("") { - completeWith { - // Request the current storage status from the Master - val storageStatusList = sc.getExecutorStorageStatus - // Calculate macro-level statistics - val maxMem = storageStatusList.map(_.maxMem).reduce(_+_) - val remainingMem = storageStatusList.map(_.memRemaining).reduce(_+_) - val diskSpaceUsed = storageStatusList.flatMap(_.blocks.values.map(_.diskSize)) - .reduceOption(_+_).getOrElse(0L) - val rdds = StorageUtils.rddInfoFromStorageStatus(storageStatusList, sc) - spark.storage.html.index. - render(maxMem, remainingMem, diskSpaceUsed, rdds, storageStatusList) - } - } ~ - path("rdd") { - parameter("id") { id => - completeWith { - val prefix = "rdd_" + id.toString - val storageStatusList = sc.getExecutorStorageStatus - val filteredStorageStatusList = StorageUtils. - filterStorageStatusByPrefix(storageStatusList, prefix) - val rddInfo = StorageUtils.rddInfoFromStorageStatus(filteredStorageStatusList, sc).head - spark.storage.html.rdd.render(rddInfo, filteredStorageStatusList) - } - } - } ~ - pathPrefix("static") { - getFromResourceDirectory(STATIC_RESOURCE_DIR) - } - } - } -} diff --git a/core/src/main/scala/spark/storage/BlockManagerWorker.scala b/core/src/main/scala/spark/storage/BlockManagerWorker.scala index d2985559c1..39064bce92 100644 --- a/core/src/main/scala/spark/storage/BlockManagerWorker.scala +++ b/core/src/main/scala/spark/storage/BlockManagerWorker.scala @@ -1,14 +1,25 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.nio.ByteBuffer -import scala.actors._ -import scala.actors.Actor._ -import scala.actors.remote._ -import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} -import scala.util.Random - -import spark.{Logging, Utils, SparkEnv} +import spark.{Logging, Utils} import spark.network._ /** @@ -19,7 +30,7 @@ import spark.network._ */ private[spark] class BlockManagerWorker(val blockManager: BlockManager) extends Logging { initLogging() - + blockManager.connectionManager.onReceiveMessage(onBlockMessageReceive) def onBlockMessageReceive(msg: Message, id: ConnectionManagerId): Option[Message] = { @@ -51,7 +62,7 @@ private[spark] class BlockManagerWorker(val blockManager: BlockManager) extends logDebug("Received [" + pB + "]") putBlock(pB.id, pB.data, pB.level) return None - } + } case BlockMessage.TYPE_GET_BLOCK => { val gB = new GetBlock(blockMessage.getId) logDebug("Received [" + gB + "]") @@ -88,30 +99,26 @@ private[spark] class BlockManagerWorker(val blockManager: BlockManager) extends private[spark] object BlockManagerWorker extends Logging { private var blockManagerWorker: BlockManagerWorker = null - private val DATA_TRANSFER_TIME_OUT_MS: Long = 500 - private val REQUEST_RETRY_INTERVAL_MS: Long = 1000 - + initLogging() - + def startBlockManagerWorker(manager: BlockManager) { blockManagerWorker = new BlockManagerWorker(manager) } - + def syncPutBlock(msg: PutBlock, toConnManagerId: ConnectionManagerId): Boolean = { val blockManager = blockManagerWorker.blockManager - val connectionManager = blockManager.connectionManager - val serializer = blockManager.serializer + val connectionManager = blockManager.connectionManager val blockMessage = BlockMessage.fromPutBlock(msg) val blockMessageArray = new BlockMessageArray(blockMessage) val resultMessage = connectionManager.sendMessageReliablySync( toConnManagerId, blockMessageArray.toBufferMessage) return (resultMessage != None) } - + def syncGetBlock(msg: GetBlock, toConnManagerId: ConnectionManagerId): ByteBuffer = { val blockManager = blockManagerWorker.blockManager - val connectionManager = blockManager.connectionManager - val serializer = blockManager.serializer + val connectionManager = blockManager.connectionManager val blockMessage = BlockMessage.fromGetBlock(msg) val blockMessageArray = new BlockMessageArray(blockMessage) val responseMessage = connectionManager.sendMessageReliablySync( diff --git a/core/src/main/scala/spark/storage/BlockMessage.scala b/core/src/main/scala/spark/storage/BlockMessage.scala index 30d7500e01..ab72dbb62b 100644 --- a/core/src/main/scala/spark/storage/BlockMessage.scala +++ b/core/src/main/scala/spark/storage/BlockMessage.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/storage/BlockMessageArray.scala b/core/src/main/scala/spark/storage/BlockMessageArray.scala index ee0c5ff9a2..b0229d6124 100644 --- a/core/src/main/scala/spark/storage/BlockMessageArray.scala +++ b/core/src/main/scala/spark/storage/BlockMessageArray.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/storage/BlockObjectWriter.scala b/core/src/main/scala/spark/storage/BlockObjectWriter.scala new file mode 100644 index 0000000000..01ed6e8c1f --- /dev/null +++ b/core/src/main/scala/spark/storage/BlockObjectWriter.scala @@ -0,0 +1,67 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.storage + +import java.nio.ByteBuffer + + +/** + * An interface for writing JVM objects to some underlying storage. This interface allows + * appending data to an existing block, and can guarantee atomicity in the case of faults + * as it allows the caller to revert partial writes. + * + * This interface does not support concurrent writes. + */ +abstract class BlockObjectWriter(val blockId: String) { + + var closeEventHandler: () => Unit = _ + + def open(): BlockObjectWriter + + def close() { + closeEventHandler() + } + + def isOpen: Boolean + + def registerCloseEventHandler(handler: () => Unit) { + closeEventHandler = handler + } + + /** + * Flush the partial writes and commit them as a single atomic block. Return the + * number of bytes written for this commit. + */ + def commit(): Long + + /** + * Reverts writes that haven't been flushed yet. Callers should invoke this function + * when there are runtime exceptions. + */ + def revertPartialWrites() + + /** + * Writes an object. + */ + def write(value: Any) + + /** + * Size of the valid writes, in bytes. + */ + def size(): Long +} diff --git a/core/src/main/scala/spark/storage/BlockStore.scala b/core/src/main/scala/spark/storage/BlockStore.scala index 8188d3595e..c8db0022b0 100644 --- a/core/src/main/scala/spark/storage/BlockStore.scala +++ b/core/src/main/scala/spark/storage/BlockStore.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/storage/DiskStore.scala b/core/src/main/scala/spark/storage/DiskStore.scala index c9553d2e0f..3495d653bd 100644 --- a/core/src/main/scala/spark/storage/DiskStore.scala +++ b/core/src/main/scala/spark/storage/DiskStore.scala @@ -1,39 +1,135 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage +import java.io.{File, FileOutputStream, OutputStream, RandomAccessFile} import java.nio.ByteBuffer -import java.io.{File, FileOutputStream, RandomAccessFile} +import java.nio.channels.FileChannel import java.nio.channels.FileChannel.MapMode import java.util.{Random, Date} import java.text.SimpleDateFormat -import it.unimi.dsi.fastutil.io.FastBufferedOutputStream - import scala.collection.mutable.ArrayBuffer -import spark.executor.ExecutorExitCode +import it.unimi.dsi.fastutil.io.FastBufferedOutputStream import spark.Utils +import spark.executor.ExecutorExitCode +import spark.serializer.{Serializer, SerializationStream} +import spark.Logging +import spark.network.netty.ShuffleSender +import spark.network.netty.PathResolver + /** * Stores BlockManager blocks on disk. */ private class DiskStore(blockManager: BlockManager, rootDirs: String) - extends BlockStore(blockManager) { + extends BlockStore(blockManager) with Logging { + + class DiskBlockObjectWriter(blockId: String, serializer: Serializer, bufferSize: Int) + extends BlockObjectWriter(blockId) { - private val mapMode = MapMode.READ_ONLY - private var mapOpenMode = "r" + private val f: File = createFile(blockId /*, allowAppendExisting */) - val MAX_DIR_CREATION_ATTEMPTS: Int = 10 - val subDirsPerLocalDir = System.getProperty("spark.diskStore.subDirectories", "64").toInt + // The file channel, used for repositioning / truncating the file. + private var channel: FileChannel = null + private var bs: OutputStream = null + private var objOut: SerializationStream = null + private var lastValidPosition = 0L + private var initialized = false + override def open(): DiskBlockObjectWriter = { + val fos = new FileOutputStream(f, true) + channel = fos.getChannel() + bs = blockManager.wrapForCompression(blockId, new FastBufferedOutputStream(fos, bufferSize)) + objOut = serializer.newInstance().serializeStream(bs) + initialized = true + this + } + + override def close() { + if (initialized) { + objOut.close() + bs.close() + channel = null + bs = null + objOut = null + } + // Invoke the close callback handler. + super.close() + } + + override def isOpen: Boolean = objOut != null + + // Flush the partial writes, and set valid length to be the length of the entire file. + // Return the number of bytes written for this commit. + override def commit(): Long = { + if (initialized) { + // NOTE: Flush the serializer first and then the compressed/buffered output stream + objOut.flush() + bs.flush() + val prevPos = lastValidPosition + lastValidPosition = channel.position() + lastValidPosition - prevPos + } else { + // lastValidPosition is zero if stream is uninitialized + lastValidPosition + } + } + + override def revertPartialWrites() { + if (initialized) { + // Discard current writes. We do this by flushing the outstanding writes and + // truncate the file to the last valid position. + objOut.flush() + bs.flush() + channel.truncate(lastValidPosition) + } + } + + override def write(value: Any) { + if (!initialized) { + open() + } + objOut.writeObject(value) + } + + override def size(): Long = lastValidPosition + } + + private val MAX_DIR_CREATION_ATTEMPTS: Int = 10 + private val subDirsPerLocalDir = System.getProperty("spark.diskStore.subDirectories", "64").toInt + + private var shuffleSender : ShuffleSender = null // Create one local directory for each path mentioned in spark.local.dir; then, inside this // directory, create multiple subdirectories that we will hash files into, in order to avoid // having really large inodes at the top level. - val localDirs = createLocalDirs() - val subDirs = Array.fill(localDirs.length)(new Array[File](subDirsPerLocalDir)) + private val localDirs: Array[File] = createLocalDirs() + private val subDirs = Array.fill(localDirs.length)(new Array[File](subDirsPerLocalDir)) addShutdownHook() + def getBlockWriter(blockId: String, serializer: Serializer, bufferSize: Int) + : BlockObjectWriter = { + new DiskBlockObjectWriter(blockId, serializer, bufferSize) + } + override def getSize(blockId: String): Long = { getFile(blockId).length() } @@ -57,9 +153,9 @@ private class DiskStore(blockManager: BlockManager, rootDirs: String) private def getFileBytes(file: File): ByteBuffer = { val length = file.length() - val channel = new RandomAccessFile(file, mapOpenMode).getChannel() + val channel = new RandomAccessFile(file, "r").getChannel() val buffer = try { - channel.map(mapMode, 0, length) + channel.map(MapMode.READ_ONLY, 0, length) } finally { channel.close() } @@ -79,12 +175,14 @@ private class DiskStore(blockManager: BlockManager, rootDirs: String) val file = createFile(blockId) val fileOut = blockManager.wrapForCompression(blockId, new FastBufferedOutputStream(new FileOutputStream(file))) - val objOut = blockManager.serializer.newInstance().serializeStream(fileOut) + val objOut = blockManager.defaultSerializer.newInstance().serializeStream(fileOut) objOut.writeAll(values.iterator) objOut.close() val length = file.length() + + val timeTaken = System.currentTimeMillis - startTime logDebug("Block %s stored as %s file on disk in %d ms".format( - blockId, Utils.memoryBytesToString(length), (System.currentTimeMillis - startTime))) + blockId, Utils.memoryBytesToString(length), timeTaken)) if (returnValues) { // Return a byte buffer for the contents of the file @@ -105,6 +203,14 @@ private class DiskStore(blockManager: BlockManager, rootDirs: String) getBytes(blockId).map(bytes => blockManager.dataDeserialize(blockId, bytes)) } + /** + * A version of getValues that allows a custom serializer. This is used as part of the + * shuffle short-circuit code. + */ + def getValues(blockId: String, serializer: Serializer): Option[Iterator[Any]] = { + getBytes(blockId).map(bytes => blockManager.dataDeserialize(blockId, bytes, serializer)) + } + override def remove(blockId: String): Boolean = { val file = getFile(blockId) if (file.exists()) { @@ -118,10 +224,13 @@ private class DiskStore(blockManager: BlockManager, rootDirs: String) getFile(blockId).exists() } - private def createFile(blockId: String): File = { + private def createFile(blockId: String, allowAppendExisting: Boolean = false): File = { val file = getFile(blockId) - if (file.exists()) { - throw new Exception("File for block " + blockId + " already exists on disk: " + file) + if (!allowAppendExisting && file.exists()) { + // NOTE(shivaram): Delete the file if it exists. This might happen if a ShuffleMap task + // was rescheduled on the same machine as the old task. + logWarning("File for block " + blockId + " already exists on disk: " + file + ". Deleting") + file.delete() } file } @@ -156,8 +265,8 @@ private class DiskStore(blockManager: BlockManager, rootDirs: String) private def createLocalDirs(): Array[File] = { logDebug("Creating local directories at root dirs '" + rootDirs + "'") val dateFormat = new SimpleDateFormat("yyyyMMddHHmmss") - rootDirs.split(",").map(rootDir => { - var foundLocalDir: Boolean = false + rootDirs.split(",").map { rootDir => + var foundLocalDir = false var localDir: File = null var localDirId: String = null var tries = 0 @@ -168,12 +277,11 @@ private class DiskStore(blockManager: BlockManager, rootDirs: String) localDirId = "%s-%04x".format(dateFormat.format(new Date), rand.nextInt(65536)) localDir = new File(rootDir, "spark-local-" + localDirId) if (!localDir.exists) { - localDir.mkdirs() - foundLocalDir = true + foundLocalDir = localDir.mkdirs() } } catch { case e: Exception => - logWarning("Attempt " + tries + " to create local dir failed", e) + logWarning("Attempt " + tries + " to create local dir " + localDir + " failed", e) } } if (!foundLocalDir) { @@ -183,20 +291,40 @@ private class DiskStore(blockManager: BlockManager, rootDirs: String) } logInfo("Created local directory at " + localDir) localDir - }) + } } private def addShutdownHook() { - localDirs.foreach(localDir => Utils.registerShutdownDeleteDir(localDir) ) + localDirs.foreach(localDir => Utils.registerShutdownDeleteDir(localDir)) Runtime.getRuntime.addShutdownHook(new Thread("delete Spark local dirs") { override def run() { logDebug("Shutdown hook called") - try { - localDirs.foreach(localDir => if (! Utils.hasRootAsShutdownDeleteDir(localDir)) Utils.deleteRecursively(localDir)) - } catch { - case t: Throwable => logError("Exception while deleting local spark dirs", t) + localDirs.foreach { localDir => + try { + if (!Utils.hasRootAsShutdownDeleteDir(localDir)) Utils.deleteRecursively(localDir) + } catch { + case t: Throwable => + logError("Exception while deleting local spark dir: " + localDir, t) + } + } + if (shuffleSender != null) { + shuffleSender.stop } } }) } + + private[storage] def startShuffleBlockSender(port: Int): Int = { + val pResolver = new PathResolver { + override def getAbsolutePath(blockId: String): String = { + if (!blockId.startsWith("shuffle_")) { + return null + } + DiskStore.this.getFile(blockId).getAbsolutePath() + } + } + shuffleSender = new ShuffleSender(port, pResolver) + logInfo("Created ShuffleSender binding to port : "+ shuffleSender.port) + shuffleSender.port + } } diff --git a/core/src/main/scala/spark/storage/MemoryStore.scala b/core/src/main/scala/spark/storage/MemoryStore.scala index eba5ee507f..b5a86b85a7 100644 --- a/core/src/main/scala/spark/storage/MemoryStore.scala +++ b/core/src/main/scala/spark/storage/MemoryStore.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.util.LinkedHashMap diff --git a/core/src/main/scala/spark/storage/PutResult.scala b/core/src/main/scala/spark/storage/PutResult.scala index 76f236057b..3a0974fe15 100644 --- a/core/src/main/scala/spark/storage/PutResult.scala +++ b/core/src/main/scala/spark/storage/PutResult.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/storage/ShuffleBlockManager.scala b/core/src/main/scala/spark/storage/ShuffleBlockManager.scala new file mode 100644 index 0000000000..8a7a6f9ed3 --- /dev/null +++ b/core/src/main/scala/spark/storage/ShuffleBlockManager.scala @@ -0,0 +1,67 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.storage + +import spark.serializer.Serializer + + +private[spark] +class ShuffleWriterGroup(val id: Int, val writers: Array[BlockObjectWriter]) + + +private[spark] +trait ShuffleBlocks { + def acquireWriters(mapId: Int): ShuffleWriterGroup + def releaseWriters(group: ShuffleWriterGroup) +} + + +private[spark] +class ShuffleBlockManager(blockManager: BlockManager) { + + def forShuffle(shuffleId: Int, numBuckets: Int, serializer: Serializer): ShuffleBlocks = { + new ShuffleBlocks { + // Get a group of writers for a map task. + override def acquireWriters(mapId: Int): ShuffleWriterGroup = { + val bufferSize = System.getProperty("spark.shuffle.file.buffer.kb", "100").toInt * 1024 + val writers = Array.tabulate[BlockObjectWriter](numBuckets) { bucketId => + val blockId = ShuffleBlockManager.blockId(shuffleId, bucketId, mapId) + blockManager.getDiskBlockWriter(blockId, serializer, bufferSize) + } + new ShuffleWriterGroup(mapId, writers) + } + + override def releaseWriters(group: ShuffleWriterGroup) = { + // Nothing really to release here. + } + } + } +} + + +private[spark] +object ShuffleBlockManager { + + // Returns the block id for a given shuffle block. + def blockId(shuffleId: Int, bucketId: Int, groupId: Int): String = { + "shuffle_" + shuffleId + "_" + groupId + "_" + bucketId + } + + // Returns true if the block is a shuffle block. + def isShuffle(blockId: String): Boolean = blockId.startsWith("shuffle_") +} diff --git a/core/src/main/scala/spark/storage/StorageLevel.scala b/core/src/main/scala/spark/storage/StorageLevel.scala index cc0c354e7e..f52650988c 100644 --- a/core/src/main/scala/spark/storage/StorageLevel.scala +++ b/core/src/main/scala/spark/storage/StorageLevel.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.io.{Externalizable, IOException, ObjectInput, ObjectOutput} diff --git a/core/src/main/scala/spark/storage/StorageUtils.scala b/core/src/main/scala/spark/storage/StorageUtils.scala index dec47a9d41..2aeed4ea3c 100644 --- a/core/src/main/scala/spark/storage/StorageUtils.scala +++ b/core/src/main/scala/spark/storage/StorageUtils.scala @@ -1,12 +1,29 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import spark.{Utils, SparkContext} import BlockManagerMasterActor.BlockStatus private[spark] -case class StorageStatus(blockManagerId: BlockManagerId, maxMem: Long, +case class StorageStatus(blockManagerId: BlockManagerId, maxMem: Long, blocks: Map[String, BlockStatus]) { - + def memUsed(blockPrefix: String = "") = { blocks.filterKeys(_.startsWith(blockPrefix)).values.map(_.memSize). reduceOption(_+_).getOrElse(0l) @@ -22,53 +39,69 @@ case class StorageStatus(blockManagerId: BlockManagerId, maxMem: Long, } case class RDDInfo(id: Int, name: String, storageLevel: StorageLevel, - numCachedPartitions: Int, numPartitions: Int, memSize: Long, diskSize: Long) { + numCachedPartitions: Int, numPartitions: Int, memSize: Long, diskSize: Long) + extends Ordered[RDDInfo] { override def toString = { import Utils.memoryBytesToString "RDD \"%s\" (%d) Storage: %s; CachedPartitions: %d; TotalPartitions: %d; MemorySize: %s; DiskSize: %s".format(name, id, storageLevel.toString, numCachedPartitions, numPartitions, memoryBytesToString(memSize), memoryBytesToString(diskSize)) } + + override def compare(that: RDDInfo) = { + this.id - that.id + } } /* Helper methods for storage-related objects */ private[spark] object StorageUtils { - /* Given the current storage status of the BlockManager, returns information for each RDD */ - def rddInfoFromStorageStatus(storageStatusList: Array[StorageStatus], + /* Returns RDD-level information, compiled from a list of StorageStatus objects */ + def rddInfoFromStorageStatus(storageStatusList: Seq[StorageStatus], sc: SparkContext) : Array[RDDInfo] = { - rddInfoFromBlockStatusList(storageStatusList.flatMap(_.blocks).toMap, sc) + rddInfoFromBlockStatusList(storageStatusList.flatMap(_.blocks).toMap, sc) } - /* Given a list of BlockStatus objets, returns information for each RDD */ - def rddInfoFromBlockStatusList(infos: Map[String, BlockStatus], + /* Returns a map of blocks to their locations, compiled from a list of StorageStatus objects */ + def blockLocationsFromStorageStatus(storageStatusList: Seq[StorageStatus]) = { + val blockLocationPairs = storageStatusList + .flatMap(s => s.blocks.map(b => (b._1, s.blockManagerId.hostPort))) + blockLocationPairs.groupBy(_._1).map{case (k, v) => (k, v.unzip._2)}.toMap + } + + /* Given a list of BlockStatus objets, returns information for each RDD */ + def rddInfoFromBlockStatusList(infos: Map[String, BlockStatus], sc: SparkContext) : Array[RDDInfo] = { // Group by rddId, ignore the partition name - val groupedRddBlocks = infos.groupBy { case(k, v) => + val groupedRddBlocks = infos.filterKeys(_.startsWith("rdd_")).groupBy { case(k, v) => k.substring(0,k.lastIndexOf('_')) }.mapValues(_.values.toArray) // For each RDD, generate an RDDInfo object - groupedRddBlocks.map { case(rddKey, rddBlocks) => - + val rddInfos = groupedRddBlocks.map { case (rddKey, rddBlocks) => // Add up memory and disk sizes val memSize = rddBlocks.map(_.memSize).reduce(_ + _) val diskSize = rddBlocks.map(_.diskSize).reduce(_ + _) // Find the id of the RDD, e.g. rdd_1 => 1 val rddId = rddKey.split("_").last.toInt - // Get the friendly name for the rdd, if available. - val rdd = sc.persistentRdds(rddId) - val rddName = Option(rdd.name).getOrElse(rddKey) - val rddStorageLevel = rdd.getStorageLevel - RDDInfo(rddId, rddName, rddStorageLevel, rddBlocks.length, rdd.partitions.size, memSize, diskSize) - }.toArray + // Get the friendly name and storage level for the RDD, if available + sc.persistentRdds.get(rddId).map { r => + val rddName = Option(r.name).getOrElse(rddKey) + val rddStorageLevel = r.getStorageLevel + RDDInfo(rddId, rddName, rddStorageLevel, rddBlocks.length, r.partitions.size, memSize, diskSize) + } + }.flatten.toArray + + scala.util.Sorting.quickSort(rddInfos) + + rddInfos } - /* Removes all BlockStatus object that are not part of a block prefix */ - def filterStorageStatusByPrefix(storageStatusList: Array[StorageStatus], + /* Removes all BlockStatus object that are not part of a block prefix */ + def filterStorageStatusByPrefix(storageStatusList: Array[StorageStatus], prefix: String) : Array[StorageStatus] = { storageStatusList.map { status => diff --git a/core/src/main/scala/spark/storage/ThreadingTest.scala b/core/src/main/scala/spark/storage/ThreadingTest.scala index 5c406e68cb..b3ab1ff4b4 100644 --- a/core/src/main/scala/spark/storage/ThreadingTest.scala +++ b/core/src/main/scala/spark/storage/ThreadingTest.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import akka.actor._ diff --git a/core/src/main/scala/spark/ui/JettyUtils.scala b/core/src/main/scala/spark/ui/JettyUtils.scala new file mode 100644 index 0000000000..ca6088ad93 --- /dev/null +++ b/core/src/main/scala/spark/ui/JettyUtils.scala @@ -0,0 +1,132 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui + +import annotation.tailrec + +import javax.servlet.http.{HttpServletResponse, HttpServletRequest} + +import net.liftweb.json.{JValue, pretty, render} + +import org.eclipse.jetty.server.{Server, Request, Handler} +import org.eclipse.jetty.server.handler.{ResourceHandler, HandlerList, ContextHandler, AbstractHandler} +import org.eclipse.jetty.util.thread.QueuedThreadPool + +import scala.util.{Try, Success, Failure} +import scala.xml.Node + +import spark.Logging + +/** Utilities for launching a web server using Jetty's HTTP Server class */ +private[spark] object JettyUtils extends Logging { + // Base type for a function that returns something based on an HTTP request. Allows for + // implicit conversion from many types of functions to jetty Handlers. + type Responder[T] = HttpServletRequest => T + + // Conversions from various types of Responder's to jetty Handlers + implicit def jsonResponderToHandler(responder: Responder[JValue]): Handler = + createHandler(responder, "text/json", (in: JValue) => pretty(render(in))) + + implicit def htmlResponderToHandler(responder: Responder[Seq[Node]]): Handler = + createHandler(responder, "text/html", (in: Seq[Node]) => "<!DOCTYPE html>" + in.toString) + + implicit def textResponderToHandler(responder: Responder[String]): Handler = + createHandler(responder, "text/plain") + + private def createHandler[T <% AnyRef](responder: Responder[T], contentType: String, + extractFn: T => String = (in: Any) => in.toString): Handler = { + new AbstractHandler { + def handle(target: String, + baseRequest: Request, + request: HttpServletRequest, + response: HttpServletResponse) { + response.setContentType("%s;charset=utf-8".format(contentType)) + response.setStatus(HttpServletResponse.SC_OK) + baseRequest.setHandled(true) + val result = responder(request) + response.setHeader("Cache-Control", "no-cache, no-store, must-revalidate") + response.getWriter().println(extractFn(result)) + } + } + } + + /** Creates a handler that always redirects the user to a given path */ + def createRedirectHandler(newPath: String): Handler = { + new AbstractHandler { + def handle(target: String, + baseRequest: Request, + request: HttpServletRequest, + response: HttpServletResponse) { + response.setStatus(302) + response.setHeader("Location", baseRequest.getRootURL + newPath) + baseRequest.setHandled(true) + } + } + } + + /** Creates a handler for serving files from a static directory */ + def createStaticHandler(resourceBase: String): ResourceHandler = { + val staticHandler = new ResourceHandler + Option(getClass.getClassLoader.getResource(resourceBase)) match { + case Some(res) => + staticHandler.setResourceBase(res.toString) + case None => + throw new Exception("Could not find resource path for Web UI: " + resourceBase) + } + staticHandler + } + + /** + * Attempts to start a Jetty server at the supplied ip:port which uses the supplied handlers. + * + * If the desired port number is contented, continues incrementing ports until a free port is + * found. Returns the chosen port and the jetty Server object. + */ + def startJettyServer(ip: String, port: Int, handlers: Seq[(String, Handler)]): (Server, Int) = { + val handlersToRegister = handlers.map { case(path, handler) => + val contextHandler = new ContextHandler(path) + contextHandler.setHandler(handler) + contextHandler.asInstanceOf[org.eclipse.jetty.server.Handler] + } + + val handlerList = new HandlerList + handlerList.setHandlers(handlersToRegister.toArray) + + @tailrec + def connect(currentPort: Int): (Server, Int) = { + val server = new Server(currentPort) + val pool = new QueuedThreadPool + pool.setDaemon(true) + server.setThreadPool(pool) + server.setHandler(handlerList) + + Try { server.start() } match { + case s: Success[_] => + sys.addShutdownHook(server.stop()) // Be kind, un-bind + (server, server.getConnectors.head.getLocalPort) + case f: Failure[_] => + server.stop() + logInfo("Failed to create UI at port, %s. Trying again.".format(currentPort)) + logInfo("Error was: " + f.toString) + connect((currentPort + 1) % 65536) + } + } + + connect(port) + } +} diff --git a/core/src/main/scala/spark/ui/Page.scala b/core/src/main/scala/spark/ui/Page.scala new file mode 100644 index 0000000000..03034a4520 --- /dev/null +++ b/core/src/main/scala/spark/ui/Page.scala @@ -0,0 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui + +private[spark] object Page extends Enumeration { val Storage, Jobs, Environment, Executors = Value } diff --git a/core/src/main/scala/spark/ui/SparkUI.scala b/core/src/main/scala/spark/ui/SparkUI.scala new file mode 100644 index 0000000000..7599f82a94 --- /dev/null +++ b/core/src/main/scala/spark/ui/SparkUI.scala @@ -0,0 +1,83 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui + +import javax.servlet.http.HttpServletRequest + +import org.eclipse.jetty.server.{Handler, Server} + +import spark.{Logging, SparkContext, Utils} +import spark.ui.env.EnvironmentUI +import spark.ui.exec.ExecutorsUI +import spark.ui.storage.BlockManagerUI +import spark.ui.jobs.JobProgressUI +import spark.ui.JettyUtils._ + +/** Top level user interface for Spark */ +private[spark] class SparkUI(sc: SparkContext) extends Logging { + val host = Utils.localHostName() + val port = Option(System.getProperty("spark.ui.port")).getOrElse(SparkUI.DEFAULT_PORT).toInt + var boundPort: Option[Int] = None + var server: Option[Server] = None + + val handlers = Seq[(String, Handler)]( + ("/static", createStaticHandler(SparkUI.STATIC_RESOURCE_DIR)), + ("/", createRedirectHandler("/stages")) + ) + val storage = new BlockManagerUI(sc) + val jobs = new JobProgressUI(sc) + val env = new EnvironmentUI(sc) + val exec = new ExecutorsUI(sc) + val allHandlers = storage.getHandlers ++ jobs.getHandlers ++ env.getHandlers ++ + exec.getHandlers ++ handlers + + /** Bind the HTTP server which backs this web interface */ + def bind() { + try { + val (srv, usedPort) = JettyUtils.startJettyServer("0.0.0.0", port, allHandlers) + logInfo("Started Spark Web UI at http://%s:%d".format(host, usedPort)) + server = Some(srv) + boundPort = Some(usedPort) + } catch { + case e: Exception => + logError("Failed to create Spark JettyUtils", e) + System.exit(1) + } + } + + /** Initialize all components of the server */ + def start() { + // NOTE: This is decoupled from bind() because of the following dependency cycle: + // DAGScheduler() requires that the port of this server is known + // This server must register all handlers, including JobProgressUI, before binding + // JobProgressUI registers a listener with SparkContext, which requires sc to initialize + jobs.start() + exec.start() + } + + def stop() { + server.foreach(_.stop()) + } + + private[spark] def appUIAddress = "http://" + host + ":" + boundPort.getOrElse("-1") +} + +private[spark] object SparkUI { + val DEFAULT_PORT = "33000" + val STATIC_RESOURCE_DIR = "spark/ui/static" +} diff --git a/core/src/main/scala/spark/ui/UIUtils.scala b/core/src/main/scala/spark/ui/UIUtils.scala new file mode 100644 index 0000000000..e33c80282a --- /dev/null +++ b/core/src/main/scala/spark/ui/UIUtils.scala @@ -0,0 +1,140 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui + +import scala.xml.Node + +import spark.SparkContext + +/** Utility functions for generating XML pages with spark content. */ +private[spark] object UIUtils { + import Page._ + + /** Returns a spark page with correctly formatted headers */ + def headerSparkPage(content: => Seq[Node], sc: SparkContext, title: String, page: Page.Value) + : Seq[Node] = { + val storage = page match { + case Storage => <li class="active"><a href="/storage">Storage</a></li> + case _ => <li><a href="/storage">Storage</a></li> + } + val jobs = page match { + case Jobs => <li class="active"><a href="/stages">Jobs</a></li> + case _ => <li><a href="/stages">Jobs</a></li> + } + val environment = page match { + case Environment => <li class="active"><a href="/environment">Environment</a></li> + case _ => <li><a href="/environment">Environment</a></li> + } + val executors = page match { + case Executors => <li class="active"><a href="/executors">Executors</a></li> + case _ => <li><a href="/executors">Executors</a></li> + } + + <html> + <head> + <meta http-equiv="Content-type" content="text/html; charset=utf-8" /> + <link rel="stylesheet" href="/static/bootstrap.min.css" type="text/css" /> + <link rel="stylesheet" href="/static/webui.css" type="text/css" /> + <link rel="stylesheet" href="/static/bootstrap-responsive.min.css" type="text/css" /> + <script src="/static/sorttable.js"></script> + <title>{sc.appName} - {title}</title> + <style type="text/css"> + table.sortable thead {{ cursor: pointer; }} + </style> + </head> + <body> + <div class="container"> + + <div class="row"> + <div class="span12"> + <div class="navbar"> + <div class="navbar-inner"> + <div class="container"> + <div class="brand"><img src="/static/spark-logo-77x50px-hd.png" /></div> + <ul class="nav"> + {storage} + {jobs} + {environment} + {executors} + </ul> + <ul id="infolist"> + <li>Application: <strong>{sc.appName}</strong></li> + <li>Master: <strong>{sc.master}</strong></li> + <li>Executors: <strong>{sc.getExecutorStorageStatus.size}</strong></li> + </ul> + </div> + </div> + </div> + </div> + </div> + + <div class="row" style="padding-top: 5px;"> + <div class="span12"> + <h1 style="vertical-align: bottom; display: inline-block;"> + {title} + </h1> + </div> + </div> + <hr/> + {content} + </div> + </body> + </html> + } + + /** Returns a page with the spark css/js and a simple format. Used for scheduler UI. */ + def basicSparkPage(content: => Seq[Node], title: String): Seq[Node] = { + <html> + <head> + <meta http-equiv="Content-type" content="text/html; charset=utf-8" /> + <link rel="stylesheet" href="/static/bootstrap.min.css" type="text/css" /> + <link rel="stylesheet" href="/static/bootstrap-responsive.min.css" type="text/css" /> + <script src="/static/sorttable.js"></script> + <title>{title}</title> + <style type="text/css"> + table.sortable thead {{ cursor: pointer; }} + </style> + </head> + <body> + <div class="container"> + <div class="row"> + <div class="span2"> + <img src="/static/spark_logo.png" /> + </div> + <div class="span10"> + <h1 style="vertical-align: bottom; margin-top: 40px; display: inline-block;"> + {title} + </h1> + </div> + </div> + {content} + </div> + </body> + </html> + } + + /** Returns an HTML table constructed by generating a row for each object in a sequence. */ + def listingTable[T](headers: Seq[String], makeRow: T => Seq[Node], rows: Seq[T]): Seq[Node] = { + <table class="table table-bordered table-striped table-condensed sortable"> + <thead>{headers.map(h => <th>{h}</th>)}</thead> + <tbody> + {rows.map(r => makeRow(r))} + </tbody> + </table> + } +} diff --git a/core/src/main/scala/spark/ui/UIWorkloadGenerator.scala b/core/src/main/scala/spark/ui/UIWorkloadGenerator.scala new file mode 100644 index 0000000000..a80e2d7002 --- /dev/null +++ b/core/src/main/scala/spark/ui/UIWorkloadGenerator.scala @@ -0,0 +1,88 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui + +import scala.util.Random + +import spark.SparkContext +import spark.SparkContext._ + +/** + * Continuously generates jobs that expose various features of the WebUI (internal testing tool). + * + * Usage: ./run spark.ui.UIWorkloadGenerator [master] + */ +private[spark] object UIWorkloadGenerator { + val NUM_PARTITIONS = 100 + val INTER_JOB_WAIT_MS = 500 + + def main(args: Array[String]) { + val master = args(0) + val appName = "Spark UI Tester" + val sc = new SparkContext(master, appName) + + // NOTE: Right now there is no easy way for us to show spark.job.annotation for a given phase, + // but we pass it here anyways since it will be useful once we do. + def setName(s: String) = { + sc.addLocalProperties("spark.job.annotation", s) + } + val baseData = sc.makeRDD(1 to NUM_PARTITIONS * 10, NUM_PARTITIONS) + def nextFloat() = (new Random()).nextFloat() + + val jobs = Seq[(String, () => Long)]( + ("Count", baseData.count), + ("Cache and Count", baseData.map(x => x).cache.count), + ("Single Shuffle", baseData.map(x => (x % 10, x)).reduceByKey(_ + _).count), + ("Entirely failed phase", baseData.map(x => throw new Exception).count), + ("Partially failed phase", { + baseData.map{x => + val probFailure = (4.0 / NUM_PARTITIONS) + if (nextFloat() < probFailure) { + throw new Exception("This is a task failure") + } + 1 + }.count + }), + ("Partially failed phase (longer tasks)", { + baseData.map{x => + val probFailure = (4.0 / NUM_PARTITIONS) + if (nextFloat() < probFailure) { + Thread.sleep(100) + throw new Exception("This is a task failure") + } + 1 + }.count + }), + ("Job with delays", baseData.map(x => Thread.sleep(100)).count) + ) + + while (true) { + for ((desc, job) <- jobs) { + try { + setName(desc) + job() + println("Job funished: " + desc) + } catch { + case e: Exception => + println("Job Failed: " + desc) + } + Thread.sleep(INTER_JOB_WAIT_MS) + } + } + } +} diff --git a/core/src/main/scala/spark/ui/env/EnvironmentUI.scala b/core/src/main/scala/spark/ui/env/EnvironmentUI.scala new file mode 100644 index 0000000000..5ae7935ed4 --- /dev/null +++ b/core/src/main/scala/spark/ui/env/EnvironmentUI.scala @@ -0,0 +1,86 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui.env + +import javax.servlet.http.HttpServletRequest + +import org.eclipse.jetty.server.Handler + +import scala.collection.JavaConversions._ +import scala.util.Properties + +import spark.ui.JettyUtils._ +import spark.ui.UIUtils.headerSparkPage +import spark.ui.Page.Environment +import spark.SparkContext +import spark.ui.UIUtils + +import scala.xml.Node + +private[spark] class EnvironmentUI(sc: SparkContext) { + + def getHandlers = Seq[(String, Handler)]( + ("/environment", (request: HttpServletRequest) => envDetails(request)) + ) + + def envDetails(request: HttpServletRequest): Seq[Node] = { + val jvmInformation = Seq( + ("Java Version", "%s (%s)".format(Properties.javaVersion, Properties.javaVendor)), + ("Java Home", Properties.javaHome), + ("Scala Version", Properties.versionString), + ("Scala Home", Properties.scalaHome) + ) + def jvmRow(kv: (String, String)) = <tr><td>{kv._1}</td><td>{kv._2}</td></tr> + def jvmTable = UIUtils.listingTable(Seq("Name", "Value"), jvmRow, jvmInformation) + + val properties = System.getProperties.iterator.toSeq + val classPathProperty = properties + .filter{case (k, v) => k.contains("java.class.path")} + .headOption + .getOrElse("", "") + val sparkProperties = properties.filter(_._1.startsWith("spark")) + val otherProperties = properties.diff(sparkProperties :+ classPathProperty) + + val propertyHeaders = Seq("Name", "Value") + def propertyRow(kv: (String, String)) = <tr><td>{kv._1}</td><td>{kv._2}</td></tr> + val sparkPropertyTable = UIUtils.listingTable(propertyHeaders, propertyRow, sparkProperties) + val otherPropertyTable = UIUtils.listingTable(propertyHeaders, propertyRow, otherProperties) + + val classPathEntries = classPathProperty._2 + .split(System.getProperty("path.separator", ":")) + .filterNot(e => e.isEmpty) + .map(e => (e, "System Classpath")) + val addedJars = sc.addedJars.iterator.toSeq.map{case (path, time) => (path, "Added By User")} + val addedFiles = sc.addedFiles.iterator.toSeq.map{case (path, time) => (path, "Added By User")} + val classPath = addedJars ++ addedFiles ++ classPathEntries + + val classPathHeaders = Seq("Resource", "Source") + def classPathRow(data: (String, String)) = <tr><td>{data._1}</td><td>{data._2}</td></tr> + val classPathTable = UIUtils.listingTable(classPathHeaders, classPathRow, classPath) + + val content = + <span> + <h2>Runtime Information</h2> {jvmTable} + <h2>Spark Properties</h2> {sparkPropertyTable} + <h2>System Properties</h2> {otherPropertyTable} + <h2>Classpath Entries</h2> {classPathTable} + </span> + + headerSparkPage(content, sc, "Environment", Environment) + } +} diff --git a/core/src/main/scala/spark/ui/exec/ExecutorsUI.scala b/core/src/main/scala/spark/ui/exec/ExecutorsUI.scala new file mode 100644 index 0000000000..db1c902955 --- /dev/null +++ b/core/src/main/scala/spark/ui/exec/ExecutorsUI.scala @@ -0,0 +1,154 @@ +package spark.ui.exec + + +import javax.servlet.http.HttpServletRequest + +import org.eclipse.jetty.server.Handler + +import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} +import scala.util.Properties + +import spark.{ExceptionFailure, Logging, SparkContext, Success, Utils} +import spark.executor.TaskMetrics +import spark.scheduler.cluster.TaskInfo +import spark.scheduler._ +import spark.SparkContext +import spark.storage.{StorageStatus, StorageUtils} +import spark.ui.JettyUtils._ +import spark.ui.Page.Executors +import spark.ui.UIUtils.headerSparkPage +import spark.ui.UIUtils + +import scala.xml.{Node, XML} + +private[spark] class ExecutorsUI(val sc: SparkContext) { + + private var _listener: Option[ExecutorsListener] = None + def listener = _listener.get + + def start() { + _listener = Some(new ExecutorsListener) + sc.addSparkListener(listener) + } + + def getHandlers = Seq[(String, Handler)]( + ("/executors", (request: HttpServletRequest) => render(request)) + ) + + def render(request: HttpServletRequest): Seq[Node] = { + val storageStatusList = sc.getExecutorStorageStatus + + val maxMem = storageStatusList.map(_.maxMem).reduce(_+_) + val memUsed = storageStatusList.map(_.memUsed()).reduce(_+_) + val diskSpaceUsed = storageStatusList.flatMap(_.blocks.values.map(_.diskSize)) + .reduceOption(_+_).getOrElse(0L) + + val execHead = Seq("Executor ID", "Address", "RDD blocks", "Memory used", "Disk used", + "Active tasks", "Failed tasks", "Complete tasks", "Total tasks") + def execRow(kv: Seq[String]) = + <tr> + <td>{kv(0)}</td> + <td>{kv(1)}</td> + <td>{kv(2)}</td> + <td sorttable_customkey={kv(3)}> + {Utils.memoryBytesToString(kv(3).toLong)} / {Utils.memoryBytesToString(kv(4).toLong)} + </td> + <td sorttable_customkey={kv(5)}> + {Utils.memoryBytesToString(kv(5).toLong)} + </td> + <td>{kv(6)}</td> + <td>{kv(7)}</td> + <td>{kv(8)}</td> + <td>{kv(9)}</td> + </tr> + val execInfo = + for (b <- 0 until storageStatusList.size) + yield getExecInfo(b) + val execTable = UIUtils.listingTable(execHead, execRow, execInfo) + + val content = + <div class="row"> + <div class="span12"> + <ul class="unstyled"> + <li><strong>Memory:</strong> + {Utils.memoryBytesToString(memUsed)} Used + ({Utils.memoryBytesToString(maxMem)} Total) </li> + <li><strong>Disk:</strong> {Utils.memoryBytesToString(diskSpaceUsed)} Used </li> + </ul> + </div> + </div> + <div class = "row"> + <div class="span12"> + {execTable} + </div> + </div>; + + headerSparkPage(content, sc, "Executors", Executors) + } + + def getExecInfo(a: Int): Seq[String] = { + val execId = sc.getExecutorStorageStatus(a).blockManagerId.executorId + val hostPort = sc.getExecutorStorageStatus(a).blockManagerId.hostPort + val rddBlocks = sc.getExecutorStorageStatus(a).blocks.size.toString + val memUsed = sc.getExecutorStorageStatus(a).memUsed().toString + val maxMem = sc.getExecutorStorageStatus(a).maxMem.toString + val diskUsed = sc.getExecutorStorageStatus(a).diskUsed().toString + val activeTasks = listener.executorToTasksActive.getOrElse(a.toString, Seq[Long]()).size.toString + val failedTasks = listener.executorToTasksFailed.getOrElse(a.toString, 0).toString + val completedTasks = listener.executorToTasksComplete.getOrElse(a.toString, 0).toString + val totalTasks = listener.executorToTaskInfos(a.toString).size.toString + + Seq( + execId, + hostPort, + rddBlocks, + memUsed, + maxMem, + diskUsed, + activeTasks, + failedTasks, + completedTasks, + totalTasks + ) + } + + private[spark] class ExecutorsListener extends SparkListener with Logging { + val executorToTasksActive = HashMap[String, HashSet[Long]]() + val executorToTasksComplete = HashMap[String, Int]() + val executorToTasksFailed = HashMap[String, Int]() + val executorToTaskInfos = + HashMap[String, ArrayBuffer[(TaskInfo, Option[TaskMetrics], Option[ExceptionFailure])]]() + + override def onTaskStart(taskStart: SparkListenerTaskStart) { + val eid = taskStart.taskInfo.executorId + if (!executorToTasksActive.contains(eid)) + executorToTasksActive(eid) = HashSet[Long]() + executorToTasksActive(eid) += taskStart.taskInfo.taskId + val taskList = executorToTaskInfos.getOrElse( + eid, ArrayBuffer[(TaskInfo, Option[TaskMetrics], Option[ExceptionFailure])]()) + taskList += ((taskStart.taskInfo, None, None)) + executorToTaskInfos(eid) = taskList + } + + override def onTaskEnd(taskEnd: SparkListenerTaskEnd) { + val eid = taskEnd.taskInfo.executorId + if (!executorToTasksActive.contains(eid)) + executorToTasksActive(eid) = HashSet[Long]() + executorToTasksActive(eid) -= taskEnd.taskInfo.taskId + val (failureInfo, metrics): (Option[ExceptionFailure], Option[TaskMetrics]) = + taskEnd.reason match { + case e: ExceptionFailure => + executorToTasksFailed(eid) = executorToTasksFailed.getOrElse(eid, 0) + 1 + (Some(e), e.metrics) + case _ => + executorToTasksComplete(eid) = executorToTasksComplete.getOrElse(eid, 0) + 1 + (None, Some(taskEnd.taskMetrics)) + } + val taskList = executorToTaskInfos.getOrElse( + eid, ArrayBuffer[(TaskInfo, Option[TaskMetrics], Option[ExceptionFailure])]()) + taskList -= ((taskEnd.taskInfo, None, None)) + taskList += ((taskEnd.taskInfo, metrics, failureInfo)) + executorToTaskInfos(eid) = taskList + } + } +} diff --git a/core/src/main/scala/spark/ui/jobs/IndexPage.scala b/core/src/main/scala/spark/ui/jobs/IndexPage.scala new file mode 100644 index 0000000000..f31af3cda6 --- /dev/null +++ b/core/src/main/scala/spark/ui/jobs/IndexPage.scala @@ -0,0 +1,129 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui.jobs + +import java.util.Date + +import javax.servlet.http.HttpServletRequest + +import scala.Some +import scala.xml.{NodeSeq, Node} + +import spark.scheduler.Stage +import spark.ui.UIUtils._ +import spark.ui.Page._ +import spark.storage.StorageLevel + +/** Page showing list of all ongoing and recently finished stages */ +private[spark] class IndexPage(parent: JobProgressUI) { + def listener = parent.listener + val dateFmt = parent.dateFmt + + def render(request: HttpServletRequest): Seq[Node] = { + val activeStages = listener.activeStages.toSeq + val completedStages = listener.completedStages.reverse.toSeq + val failedStages = listener.failedStages.reverse.toSeq + + /** Special table which merges two header cells. */ + def stageTable[T](makeRow: T => Seq[Node], rows: Seq[T]): Seq[Node] = { + <table class="table table-bordered table-striped table-condensed sortable"> + <thead> + <th>Stage Id</th> + <th>Origin</th> + <th>Submitted</th> + <th>Duration</th> + <th colspan="2">Tasks: Complete/Total</th> + <th>Shuffle Activity</th> + <th>Stored RDD</th> + </thead> + <tbody> + {rows.map(r => makeRow(r))} + </tbody> + </table> + } + + val activeStageTable: NodeSeq = stageTable(stageRow, activeStages) + val completedStageTable = stageTable(stageRow, completedStages) + val failedStageTable: NodeSeq = stageTable(stageRow, failedStages) + + val content = <h2>Active Stages</h2> ++ activeStageTable ++ + <h2>Completed Stages</h2> ++ completedStageTable ++ + <h2>Failed Stages</h2> ++ failedStageTable + + headerSparkPage(content, parent.sc, "Spark Stages", Jobs) + } + + def getElapsedTime(submitted: Option[Long], completed: Long): String = { + submitted match { + case Some(t) => parent.formatDuration(completed - t) + case _ => "Unknown" + } + } + + def makeProgressBar(completed: Int, total: Int): Seq[Node] = { + val width=130 + val height=15 + val completeWidth = (completed.toDouble / total) * width + + <svg width={width.toString} height={height.toString}> + <rect width={width.toString} height={height.toString} + fill="white" stroke="rgb(51,51,51)" stroke-width="1" /> + <rect width={completeWidth.toString} height={height.toString} + fill="rgb(0,136,204)" stroke="black" stroke-width="1" /> + </svg> + } + + + def stageRow(s: Stage): Seq[Node] = { + val submissionTime = s.submissionTime match { + case Some(t) => dateFmt.format(new Date(t)) + case None => "Unknown" + } + val (read, write) = (listener.hasShuffleRead(s.id), listener.hasShuffleWrite(s.id)) + val shuffleInfo = (read, write) match { + case (true, true) => "Read/Write" + case (true, false) => "Read" + case (false, true) => "Write" + case _ => "" + } + val completedTasks = listener.stageToTasksComplete.getOrElse(s.id, 0) + val totalTasks = s.numPartitions + + <tr> + <td>{s.id}</td> + <td><a href={"/stages/stage?id=%s".format(s.id)}>{s.name}</a></td> + <td>{submissionTime}</td> + <td>{getElapsedTime(s.submissionTime, + s.completionTime.getOrElse(System.currentTimeMillis()))}</td> + <td class="progress-cell">{makeProgressBar(completedTasks, totalTasks)}</td> + <td style="border-left: 0; text-align: center;">{completedTasks} / {totalTasks} + {listener.stageToTasksFailed.getOrElse(s.id, 0) match { + case f if f > 0 => "(%s failed)".format(f) + case _ => + }} + </td> + <td>{shuffleInfo}</td> + <td>{if (s.rdd.getStorageLevel != StorageLevel.NONE) { + <a href={"/storage/rdd?id=%s".format(s.rdd.id)}> + {Option(s.rdd.name).getOrElse(s.rdd.id)} + </a> + }} + </td> + </tr> + } +} diff --git a/core/src/main/scala/spark/ui/jobs/JobProgressUI.scala b/core/src/main/scala/spark/ui/jobs/JobProgressUI.scala new file mode 100644 index 0000000000..6e332415db --- /dev/null +++ b/core/src/main/scala/spark/ui/jobs/JobProgressUI.scala @@ -0,0 +1,160 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui.jobs + +import akka.util.Duration + +import java.text.SimpleDateFormat + +import javax.servlet.http.HttpServletRequest + +import org.eclipse.jetty.server.Handler + +import scala.Seq +import scala.collection.mutable.{HashSet, ListBuffer, HashMap, ArrayBuffer} + +import spark.ui.JettyUtils._ +import spark.{ExceptionFailure, SparkContext, Success, Utils} +import spark.scheduler._ +import spark.scheduler.cluster.TaskInfo +import spark.executor.TaskMetrics +import collection.mutable + +/** Web UI showing progress status of all jobs in the given SparkContext. */ +private[spark] class JobProgressUI(val sc: SparkContext) { + private var _listener: Option[JobProgressListener] = None + def listener = _listener.get + val dateFmt = new SimpleDateFormat("yyyy/MM/dd HH:mm:ss") + + private val indexPage = new IndexPage(this) + private val stagePage = new StagePage(this) + + def start() { + _listener = Some(new JobProgressListener) + sc.addSparkListener(listener) + } + + def formatDuration(ms: Long) = Utils.msDurationToString(ms) + + def getHandlers = Seq[(String, Handler)]( + ("/stages/stage", (request: HttpServletRequest) => stagePage.render(request)), + ("/stages", (request: HttpServletRequest) => indexPage.render(request)) + ) +} + +private[spark] class JobProgressListener extends SparkListener { + // How many stages to remember + val RETAINED_STAGES = System.getProperty("spark.ui.retained_stages", "1000").toInt + + val activeStages = HashSet[Stage]() + val completedStages = ListBuffer[Stage]() + val failedStages = ListBuffer[Stage]() + + val stageToTasksActive = HashMap[Int, HashSet[Long]]() + val stageToTasksComplete = HashMap[Int, Int]() + val stageToTasksFailed = HashMap[Int, Int]() + val stageToTaskInfos = + HashMap[Int, ArrayBuffer[(TaskInfo, Option[TaskMetrics], Option[ExceptionFailure])]]() + + override def onJobStart(jobStart: SparkListenerJobStart) {} + + override def onStageCompleted(stageCompleted: StageCompleted) = { + val stage = stageCompleted.stageInfo.stage + activeStages -= stage + completedStages += stage + trimIfNecessary(completedStages) + } + + /** If stages is too large, remove and garbage collect old stages */ + def trimIfNecessary(stages: ListBuffer[Stage]) { + if (stages.size > RETAINED_STAGES) { + val toRemove = RETAINED_STAGES / 10 + stages.takeRight(toRemove).foreach( s => { + stageToTaskInfos.remove(s.id) + }) + stages.trimEnd(toRemove) + } + } + + override def onStageSubmitted(stageSubmitted: SparkListenerStageSubmitted) = + activeStages += stageSubmitted.stage + + override def onTaskStart(taskStart: SparkListenerTaskStart) { + val sid = taskStart.task.stageId + if (!stageToTasksActive.contains(sid)) + stageToTasksActive(sid) = HashSet[Long]() + stageToTasksActive(sid) += taskStart.taskInfo.taskId + val taskList = stageToTaskInfos.getOrElse( + sid, ArrayBuffer[(TaskInfo, Option[TaskMetrics], Option[ExceptionFailure])]()) + taskList += ((taskStart.taskInfo, None, None)) + stageToTaskInfos(sid) = taskList + } + + override def onTaskEnd(taskEnd: SparkListenerTaskEnd) { + val sid = taskEnd.task.stageId + if (!stageToTasksActive.contains(sid)) + stageToTasksActive(sid) = HashSet[Long]() + stageToTasksActive(sid) -= taskEnd.taskInfo.taskId + val (failureInfo, metrics): (Option[ExceptionFailure], Option[TaskMetrics]) = + taskEnd.reason match { + case e: ExceptionFailure => + stageToTasksFailed(sid) = stageToTasksFailed.getOrElse(sid, 0) + 1 + (Some(e), e.metrics) + case _ => + stageToTasksComplete(sid) = stageToTasksComplete.getOrElse(sid, 0) + 1 + (None, Some(taskEnd.taskMetrics)) + } + val taskList = stageToTaskInfos.getOrElse( + sid, ArrayBuffer[(TaskInfo, Option[TaskMetrics], Option[ExceptionFailure])]()) + taskList -= ((taskEnd.taskInfo, None, None)) + taskList += ((taskEnd.taskInfo, metrics, failureInfo)) + stageToTaskInfos(sid) = taskList + } + + override def onJobEnd(jobEnd: SparkListenerJobEnd) { + jobEnd match { + case end: SparkListenerJobEnd => + end.jobResult match { + case JobFailed(ex, Some(stage)) => + activeStages -= stage + failedStages += stage + trimIfNecessary(failedStages) + case _ => + } + case _ => + } + } + + /** Is this stage's input from a shuffle read. */ + def hasShuffleRead(stageID: Int): Boolean = { + // This is written in a slightly complicated way to avoid having to scan all tasks + for (s <- stageToTaskInfos.get(stageID).getOrElse(Seq())) { + if (s._2 != null) return s._2.flatMap(m => m.shuffleReadMetrics).isDefined + } + return false // No tasks have finished for this stage + } + + /** Is this stage's output to a shuffle write. */ + def hasShuffleWrite(stageID: Int): Boolean = { + // This is written in a slightly complicated way to avoid having to scan all tasks + for (s <- stageToTaskInfos.get(stageID).getOrElse(Seq())) { + if (s._2 != null) return s._2.flatMap(m => m.shuffleWriteMetrics).isDefined + } + return false // No tasks have finished for this stage + } +} diff --git a/core/src/main/scala/spark/ui/jobs/StagePage.scala b/core/src/main/scala/spark/ui/jobs/StagePage.scala new file mode 100644 index 0000000000..654f347723 --- /dev/null +++ b/core/src/main/scala/spark/ui/jobs/StagePage.scala @@ -0,0 +1,138 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui.jobs + +import java.util.Date + +import javax.servlet.http.HttpServletRequest + +import scala.xml.Node + +import spark.ui.UIUtils._ +import spark.ui.Page._ +import spark.util.Distribution +import spark.{ExceptionFailure, Utils} +import spark.scheduler.cluster.TaskInfo +import spark.executor.TaskMetrics + +/** Page showing statistics and task list for a given stage */ +private[spark] class StagePage(parent: JobProgressUI) { + def listener = parent.listener + val dateFmt = parent.dateFmt + + def render(request: HttpServletRequest): Seq[Node] = { + val stageId = request.getParameter("id").toInt + + if (!listener.stageToTaskInfos.contains(stageId)) { + val content = + <div> + <h2>Summary Metrics</h2> No tasks have started yet + <h2>Tasks</h2> No tasks have started yet + </div> + return headerSparkPage(content, parent.sc, "Stage Details: %s".format(stageId), Jobs) + } + + val tasks = listener.stageToTaskInfos(stageId) + + val shuffleRead = listener.hasShuffleRead(stageId) + val shuffleWrite = listener.hasShuffleWrite(stageId) + + val taskHeaders: Seq[String] = + Seq("Task ID", "Status", "Duration", "Locality Level", "Worker", "Launch Time") ++ + {if (shuffleRead) Seq("Shuffle Read") else Nil} ++ + {if (shuffleWrite) Seq("Shuffle Write") else Nil} ++ + Seq("Details") + + val taskTable = listingTable(taskHeaders, taskRow, tasks) + + // Excludes tasks which failed and have incomplete metrics + val validTasks = tasks.filter(t => t._1.status == "SUCCESS" && (Option(t._2).isDefined)) + + val summaryTable: Option[Seq[Node]] = + if (validTasks.size == 0) { + None + } + else { + val serviceTimes = validTasks.map{case (info, metrics, exception) => + metrics.get.executorRunTime.toDouble} + val serviceQuantiles = "Duration" +: Distribution(serviceTimes).get.getQuantiles().map( + ms => parent.formatDuration(ms.toLong)) + + def getQuantileCols(data: Seq[Double]) = + Distribution(data).get.getQuantiles().map(d => Utils.memoryBytesToString(d.toLong)) + + val shuffleReadSizes = validTasks.map { + case(info, metrics, exception) => + metrics.get.shuffleReadMetrics.map(_.remoteBytesRead).getOrElse(0L).toDouble + } + val shuffleReadQuantiles = "Shuffle Read (Remote)" +: getQuantileCols(shuffleReadSizes) + + val shuffleWriteSizes = validTasks.map { + case(info, metrics, exception) => + metrics.get.shuffleWriteMetrics.map(_.shuffleBytesWritten).getOrElse(0L).toDouble + } + val shuffleWriteQuantiles = "Shuffle Write" +: getQuantileCols(shuffleWriteSizes) + + val listings: Seq[Seq[String]] = Seq(serviceQuantiles, + if (shuffleRead) shuffleReadQuantiles else Nil, + if (shuffleWrite) shuffleWriteQuantiles else Nil) + + val quantileHeaders = Seq("Metric", "Min", "25%", "50%", "75%", "Max") + def quantileRow(data: Seq[String]): Seq[Node] = <tr> {data.map(d => <td>{d}</td>)} </tr> + Some(listingTable(quantileHeaders, quantileRow, listings)) + } + + val content = + <h2>Summary Metrics</h2> ++ summaryTable.getOrElse(Nil) ++ <h2>Tasks</h2> ++ taskTable; + + headerSparkPage(content, parent.sc, "Stage Details: %s".format(stageId), Jobs) + } + + + def taskRow(taskData: (TaskInfo, Option[TaskMetrics], Option[ExceptionFailure])): Seq[Node] = { + def fmtStackTrace(trace: Seq[StackTraceElement]): Seq[Node] = + trace.map(e => <span style="display:block;">{e.toString}</span>) + val (info, metrics, exception) = taskData + + val duration = if (info.status == "RUNNING") info.timeRunning(System.currentTimeMillis()) + else metrics.map(m => m.executorRunTime).getOrElse(1) + val formatDuration = if (info.status == "RUNNING") parent.formatDuration(duration) + else metrics.map(m => parent.formatDuration(m.executorRunTime)).getOrElse("") + + <tr> + <td>{info.taskId}</td> + <td>{info.status}</td> + <td sorttable_customkey={duration.toString}> + {formatDuration} + </td> + <td>{info.taskLocality}</td> + <td>{info.hostPort}</td> + <td>{dateFmt.format(new Date(info.launchTime))}</td> + {metrics.flatMap{m => m.shuffleReadMetrics}.map{s => + <td>{Utils.memoryBytesToString(s.remoteBytesRead)}</td>}.getOrElse("")} + {metrics.flatMap{m => m.shuffleWriteMetrics}.map{s => + <td>{Utils.memoryBytesToString(s.shuffleBytesWritten)}</td>}.getOrElse("")} + <td>{exception.map(e => + <span> + {e.className} ({e.description})<br/> + {fmtStackTrace(e.stackTrace)} + </span>).getOrElse("")} + </td> + </tr> + } +} diff --git a/core/src/main/scala/spark/ui/storage/BlockManagerUI.scala b/core/src/main/scala/spark/ui/storage/BlockManagerUI.scala new file mode 100644 index 0000000000..49ed069c75 --- /dev/null +++ b/core/src/main/scala/spark/ui/storage/BlockManagerUI.scala @@ -0,0 +1,41 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui.storage + +import akka.util.Duration + +import javax.servlet.http.HttpServletRequest + +import org.eclipse.jetty.server.Handler + +import spark.{Logging, SparkContext} +import spark.ui.JettyUtils._ + +/** Web UI showing storage status of all RDD's in the given SparkContext. */ +private[spark] class BlockManagerUI(val sc: SparkContext) extends Logging { + implicit val timeout = Duration.create( + System.getProperty("spark.akka.askTimeout", "10").toLong, "seconds") + + val indexPage = new IndexPage(this) + val rddPage = new RDDPage(this) + + def getHandlers = Seq[(String, Handler)]( + ("/storage/rdd", (request: HttpServletRequest) => rddPage.render(request)), + ("/storage", (request: HttpServletRequest) => indexPage.render(request)) + ) +} diff --git a/core/src/main/scala/spark/ui/storage/IndexPage.scala b/core/src/main/scala/spark/ui/storage/IndexPage.scala new file mode 100644 index 0000000000..f76192eba8 --- /dev/null +++ b/core/src/main/scala/spark/ui/storage/IndexPage.scala @@ -0,0 +1,65 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui.storage + +import javax.servlet.http.HttpServletRequest + +import scala.xml.Node + +import spark.storage.{RDDInfo, StorageUtils} +import spark.Utils +import spark.ui.UIUtils._ +import spark.ui.Page._ + +/** Page showing list of RDD's currently stored in the cluster */ +private[spark] class IndexPage(parent: BlockManagerUI) { + val sc = parent.sc + + def render(request: HttpServletRequest): Seq[Node] = { + val storageStatusList = sc.getExecutorStorageStatus + // Calculate macro-level statistics + + val rddHeaders = Seq( + "RDD Name", + "Storage Level", + "Cached Partitions", + "Fraction Partitions Cached", + "Size in Memory", + "Size on Disk") + val rdds = StorageUtils.rddInfoFromStorageStatus(storageStatusList, sc) + val content = listingTable(rddHeaders, rddRow, rdds) + + headerSparkPage(content, parent.sc, "Spark Storage ", Storage) + } + + def rddRow(rdd: RDDInfo): Seq[Node] = { + <tr> + <td> + <a href={"/storage/rdd?id=%s".format(rdd.id)}> + {rdd.name} + </a> + </td> + <td>{rdd.storageLevel.description} + </td> + <td>{rdd.numCachedPartitions}</td> + <td>{rdd.numCachedPartitions / rdd.numPartitions.toDouble}</td> + <td>{Utils.memoryBytesToString(rdd.memSize)}</td> + <td>{Utils.memoryBytesToString(rdd.diskSize)}</td> + </tr> + } +} diff --git a/core/src/main/scala/spark/ui/storage/RDDPage.scala b/core/src/main/scala/spark/ui/storage/RDDPage.scala new file mode 100644 index 0000000000..003be54ad8 --- /dev/null +++ b/core/src/main/scala/spark/ui/storage/RDDPage.scala @@ -0,0 +1,130 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui.storage + +import javax.servlet.http.HttpServletRequest + +import scala.xml.Node + +import spark.storage.{StorageStatus, StorageUtils} +import spark.ui.UIUtils._ +import spark.Utils +import spark.storage.BlockManagerMasterActor.BlockStatus +import spark.ui.Page._ + +/** Page showing storage details for a given RDD */ +private[spark] class RDDPage(parent: BlockManagerUI) { + val sc = parent.sc + + def render(request: HttpServletRequest): Seq[Node] = { + val id = request.getParameter("id") + val prefix = "rdd_" + id.toString + val storageStatusList = sc.getExecutorStorageStatus + val filteredStorageStatusList = StorageUtils. + filterStorageStatusByPrefix(storageStatusList, prefix) + val rddInfo = StorageUtils.rddInfoFromStorageStatus(filteredStorageStatusList, sc).head + + val workerHeaders = Seq("Host", "Memory Usage", "Disk Usage") + val workers = filteredStorageStatusList.map((prefix, _)) + val workerTable = listingTable(workerHeaders, workerRow, workers) + + val blockHeaders = Seq("Block Name", "Storage Level", "Size in Memory", "Size on Disk", + "Locations") + + val blockStatuses = filteredStorageStatusList.flatMap(_.blocks).toArray.sortWith(_._1 < _._1) + val blockLocations = StorageUtils.blockLocationsFromStorageStatus(filteredStorageStatusList) + val blocks = blockStatuses.map { + case(id, status) => (id, status, blockLocations.get(id).getOrElse(Seq("UNKNOWN"))) + } + val blockTable = listingTable(blockHeaders, blockRow, blocks) + + val content = + <div class="row"> + <div class="span12"> + <ul class="unstyled"> + <li> + <strong>Storage Level:</strong> + {rddInfo.storageLevel.description} + </li> + <li> + <strong>Cached Partitions:</strong> + {rddInfo.numCachedPartitions} + </li> + <li> + <strong>Total Partitions:</strong> + {rddInfo.numPartitions} + </li> + <li> + <strong>Memory Size:</strong> + {Utils.memoryBytesToString(rddInfo.memSize)} + </li> + <li> + <strong>Disk Size:</strong> + {Utils.memoryBytesToString(rddInfo.diskSize)} + </li> + </ul> + </div> + </div> + <hr/> + <div class="row"> + <div class="span12"> + {workerTable} + </div> + </div> + <hr/> + <div class="row"> + <div class="span12"> + <h3> RDD Summary </h3> + {blockTable} + </div> + </div>; + + headerSparkPage(content, parent.sc, "RDD Info: " + rddInfo.name, Jobs) + } + + def blockRow(row: (String, BlockStatus, Seq[String])): Seq[Node] = { + val (id, block, locations) = row + <tr> + <td>{id}</td> + <td> + {block.storageLevel.description} + </td> + <td sorttable_customkey={block.memSize.toString}> + {Utils.memoryBytesToString(block.memSize)} + </td> + <td sorttable_customkey={block.diskSize.toString}> + {Utils.memoryBytesToString(block.diskSize)} + </td> + <td> + {locations.map(l => <span>{l}<br/></span>)} + </td> + </tr> + } + + def workerRow(worker: (String, StorageStatus)): Seq[Node] = { + val (prefix, status) = worker + <tr> + <td>{status.blockManagerId.host + ":" + status.blockManagerId.port}</td> + <td> + {Utils.memoryBytesToString(status.memUsed(prefix))} + ({Utils.memoryBytesToString(status.memRemaining)} Total Available) + </td> + <td>{Utils.memoryBytesToString(status.diskUsed(prefix))}</td> + </tr> + } +} diff --git a/core/src/main/scala/spark/util/AkkaUtils.scala b/core/src/main/scala/spark/util/AkkaUtils.scala index 9fb7e001ba..9233277bdb 100644 --- a/core/src/main/scala/spark/util/AkkaUtils.scala +++ b/core/src/main/scala/spark/util/AkkaUtils.scala @@ -1,18 +1,27 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util -import akka.actor.{ActorRef, Props, ActorSystemImpl, ActorSystem} +import akka.actor.{ActorSystem, ExtendedActorSystem} import com.typesafe.config.ConfigFactory import akka.util.duration._ -import akka.pattern.ask import akka.remote.RemoteActorRefProvider -import cc.spray.Route -import cc.spray.io.IoWorker -import cc.spray.{SprayCanRootService, HttpService} -import cc.spray.can.server.HttpServer -import cc.spray.io.pipelines.MessageHandlerDispatch.SingletonHandler -import akka.dispatch.Await -import spark.{Utils, SparkException} -import java.util.concurrent.TimeoutException + /** * Various utility classes for working with Akka. @@ -29,7 +38,7 @@ private[spark] object AkkaUtils { def createActorSystem(name: String, host: String, port: Int): (ActorSystem, Int) = { val akkaThreads = System.getProperty("spark.akka.threads", "4").toInt val akkaBatchSize = System.getProperty("spark.akka.batchSize", "15").toInt - val akkaTimeout = System.getProperty("spark.akka.timeout", "20").toInt + val akkaTimeout = System.getProperty("spark.akka.timeout", "60").toInt val akkaFrameSize = System.getProperty("spark.akka.frameSize", "10").toInt val lifecycleEvents = if (System.getProperty("spark.akka.logLifecycleEvents", "false").toBoolean) "on" else "off" // 10 seconds is the default akka timeout, but in a cluster, we need higher by default. @@ -52,40 +61,12 @@ private[spark] object AkkaUtils { """.format(host, port, akkaTimeout, akkaFrameSize, akkaThreads, akkaBatchSize, lifecycleEvents, akkaWriteTimeout)) - val actorSystem = ActorSystem(name, akkaConf, getClass.getClassLoader) + val actorSystem = ActorSystem(name, akkaConf) // Figure out the port number we bound to, in case port was passed as 0. This is a bit of a // hack because Akka doesn't let you figure out the port through the public API yet. - val provider = actorSystem.asInstanceOf[ActorSystemImpl].provider + val provider = actorSystem.asInstanceOf[ExtendedActorSystem].provider val boundPort = provider.asInstanceOf[RemoteActorRefProvider].transport.address.port.get return (actorSystem, boundPort) } - - /** - * Creates a Spray HTTP server bound to a given IP and port with a given Spray Route object to - * handle requests. Returns the bound port or throws a SparkException on failure. - * TODO: Not changing ip to host here - is it required ? - */ - def startSprayServer(actorSystem: ActorSystem, ip: String, port: Int, route: Route, - name: String = "HttpServer"): ActorRef = { - val ioWorker = new IoWorker(actorSystem).start() - val httpService = actorSystem.actorOf(Props(new HttpService(route))) - val rootService = actorSystem.actorOf(Props(new SprayCanRootService(httpService))) - val server = actorSystem.actorOf( - Props(new HttpServer(ioWorker, SingletonHandler(rootService))), name = name) - actorSystem.registerOnTermination { ioWorker.stop() } - val timeout = 3.seconds - val future = server.ask(HttpServer.Bind(ip, port))(timeout) - try { - Await.result(future, timeout) match { - case bound: HttpServer.Bound => - return server - case other: Any => - throw new SparkException("Failed to bind web UI to port " + port + ": " + other) - } - } catch { - case e: TimeoutException => - throw new SparkException("Failed to bind web UI to port " + port) - } - } } diff --git a/core/src/main/scala/spark/util/BoundedPriorityQueue.scala b/core/src/main/scala/spark/util/BoundedPriorityQueue.scala new file mode 100644 index 0000000000..0575497f5d --- /dev/null +++ b/core/src/main/scala/spark/util/BoundedPriorityQueue.scala @@ -0,0 +1,62 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.util + +import java.io.Serializable +import java.util.{PriorityQueue => JPriorityQueue} +import scala.collection.generic.Growable +import scala.collection.JavaConverters._ + +/** + * Bounded priority queue. This class wraps the original PriorityQueue + * class and modifies it such that only the top K elements are retained. + * The top K elements are defined by an implicit Ordering[A]. + */ +class BoundedPriorityQueue[A](maxSize: Int)(implicit ord: Ordering[A]) + extends Iterable[A] with Growable[A] with Serializable { + + private val underlying = new JPriorityQueue[A](maxSize, ord) + + override def iterator: Iterator[A] = underlying.iterator.asScala + + override def ++=(xs: TraversableOnce[A]): this.type = { + xs.foreach { this += _ } + this + } + + override def +=(elem: A): this.type = { + if (size < maxSize) underlying.offer(elem) + else maybeReplaceLowest(elem) + this + } + + override def +=(elem1: A, elem2: A, elems: A*): this.type = { + this += elem1 += elem2 ++= elems + } + + override def clear() { underlying.clear() } + + private def maybeReplaceLowest(a: A): Boolean = { + val head = underlying.peek() + if (head != null && ord.gt(a, head)) { + underlying.poll() + underlying.offer(a) + } else false + } +} + diff --git a/core/src/main/scala/spark/util/ByteBufferInputStream.scala b/core/src/main/scala/spark/util/ByteBufferInputStream.scala index d7e67497fe..47a28e2f76 100644 --- a/core/src/main/scala/spark/util/ByteBufferInputStream.scala +++ b/core/src/main/scala/spark/util/ByteBufferInputStream.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import java.io.InputStream diff --git a/core/src/main/scala/spark/util/CompletionIterator.scala b/core/src/main/scala/spark/util/CompletionIterator.scala index 8139183780..210450892b 100644 --- a/core/src/main/scala/spark/util/CompletionIterator.scala +++ b/core/src/main/scala/spark/util/CompletionIterator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util /** @@ -22,4 +39,4 @@ object CompletionIterator { def completion() = completionFunction } } -}
\ No newline at end of file +} diff --git a/core/src/main/scala/spark/util/Distribution.scala b/core/src/main/scala/spark/util/Distribution.scala index 24738b4307..5d4d7a6c50 100644 --- a/core/src/main/scala/spark/util/Distribution.scala +++ b/core/src/main/scala/spark/util/Distribution.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import java.io.PrintStream @@ -62,4 +79,4 @@ object Distribution { quantiles.foreach{q => out.print(q + "\t")} out.println } -}
\ No newline at end of file +} diff --git a/core/src/main/scala/spark/util/IdGenerator.scala b/core/src/main/scala/spark/util/IdGenerator.scala index b6e309fe1a..3422280559 100644 --- a/core/src/main/scala/spark/util/IdGenerator.scala +++ b/core/src/main/scala/spark/util/IdGenerator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import java.util.concurrent.atomic.AtomicInteger diff --git a/core/src/main/scala/spark/util/IntParam.scala b/core/src/main/scala/spark/util/IntParam.scala index 0427646747..daf0d58fa2 100644 --- a/core/src/main/scala/spark/util/IntParam.scala +++ b/core/src/main/scala/spark/util/IntParam.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util /** diff --git a/core/src/main/scala/spark/util/MemoryParam.scala b/core/src/main/scala/spark/util/MemoryParam.scala index 3726738842..298562323a 100644 --- a/core/src/main/scala/spark/util/MemoryParam.scala +++ b/core/src/main/scala/spark/util/MemoryParam.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import spark.Utils diff --git a/core/src/main/scala/spark/util/MetadataCleaner.scala b/core/src/main/scala/spark/util/MetadataCleaner.scala index dafa906712..92909e0959 100644 --- a/core/src/main/scala/spark/util/MetadataCleaner.scala +++ b/core/src/main/scala/spark/util/MetadataCleaner.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import java.util.concurrent.{TimeUnit, ScheduledFuture, Executors} diff --git a/core/src/main/scala/spark/util/NextIterator.scala b/core/src/main/scala/spark/util/NextIterator.scala index 48b5018ddd..22163ece8d 100644 --- a/core/src/main/scala/spark/util/NextIterator.scala +++ b/core/src/main/scala/spark/util/NextIterator.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util /** Provides a basic/boilerplate Iterator implementation. */ @@ -68,4 +85,4 @@ private[spark] abstract class NextIterator[U] extends Iterator[U] { gotNext = false nextValue } -}
\ No newline at end of file +} diff --git a/core/src/main/scala/spark/util/RateLimitedOutputStream.scala b/core/src/main/scala/spark/util/RateLimitedOutputStream.scala index e3f00ea8c7..00f782bbe7 100644 --- a/core/src/main/scala/spark/util/RateLimitedOutputStream.scala +++ b/core/src/main/scala/spark/util/RateLimitedOutputStream.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import scala.annotation.tailrec diff --git a/core/src/main/scala/spark/util/SerializableBuffer.scala b/core/src/main/scala/spark/util/SerializableBuffer.scala index 09d588fe1c..7e6842628a 100644 --- a/core/src/main/scala/spark/util/SerializableBuffer.scala +++ b/core/src/main/scala/spark/util/SerializableBuffer.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import java.nio.ByteBuffer diff --git a/core/src/main/scala/spark/util/StatCounter.scala b/core/src/main/scala/spark/util/StatCounter.scala index 5f80180339..76358d4151 100644 --- a/core/src/main/scala/spark/util/StatCounter.scala +++ b/core/src/main/scala/spark/util/StatCounter.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util /** @@ -37,17 +54,23 @@ class StatCounter(values: TraversableOnce[Double]) extends Serializable { if (other == this) { merge(other.copy()) // Avoid overwriting fields in a weird order } else { - val delta = other.mu - mu - if (other.n * 10 < n) { - mu = mu + (delta * other.n) / (n + other.n) - } else if (n * 10 < other.n) { - mu = other.mu - (delta * n) / (n + other.n) - } else { - mu = (mu * n + other.mu * other.n) / (n + other.n) + if (n == 0) { + mu = other.mu + m2 = other.m2 + n = other.n + } else if (other.n != 0) { + val delta = other.mu - mu + if (other.n * 10 < n) { + mu = mu + (delta * other.n) / (n + other.n) + } else if (n * 10 < other.n) { + mu = other.mu - (delta * n) / (n + other.n) + } else { + mu = (mu * n + other.mu * other.n) / (n + other.n) + } + m2 += other.m2 + (delta * delta * n * other.n) / (n + other.n) + n += other.n } - m2 += other.m2 + (delta * delta * n * other.n) / (n + other.n) - n += other.n - this + this } } diff --git a/core/src/main/scala/spark/util/TimeStampedHashMap.scala b/core/src/main/scala/spark/util/TimeStampedHashMap.scala index 92dfaa6e6f..07772a0afb 100644 --- a/core/src/main/scala/spark/util/TimeStampedHashMap.scala +++ b/core/src/main/scala/spark/util/TimeStampedHashMap.scala @@ -1,8 +1,26 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import java.util.concurrent.ConcurrentHashMap import scala.collection.JavaConversions import scala.collection.mutable.Map +import scala.collection.immutable import spark.scheduler.MapStatus /** @@ -82,6 +100,8 @@ class TimeStampedHashMap[A, B] extends Map[A, B]() with spark.Logging { } } + def toMap: immutable.Map[A, B] = iterator.toMap + /** * Removes old key-value pairs that have timestamp earlier than `threshTime` */ diff --git a/core/src/main/scala/spark/util/TimeStampedHashSet.scala b/core/src/main/scala/spark/util/TimeStampedHashSet.scala index 5f1cc93752..41e3fd8cba 100644 --- a/core/src/main/scala/spark/util/TimeStampedHashSet.scala +++ b/core/src/main/scala/spark/util/TimeStampedHashSet.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import scala.collection.mutable.Set diff --git a/core/src/main/scala/spark/util/Vector.scala b/core/src/main/scala/spark/util/Vector.scala index 835822edb2..ed49386f18 100644 --- a/core/src/main/scala/spark/util/Vector.scala +++ b/core/src/main/scala/spark/util/Vector.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util class Vector(val elements: Array[Double]) extends Serializable { diff --git a/core/src/main/twirl/spark/common/layout.scala.html b/core/src/main/twirl/spark/common/layout.scala.html deleted file mode 100644 index b9192060aa..0000000000 --- a/core/src/main/twirl/spark/common/layout.scala.html +++ /dev/null @@ -1,35 +0,0 @@ -@(title: String)(content: Html) - -<!DOCTYPE html> -<html> - - <head> - <meta http-equiv="Content-type" content="text/html; charset=utf-8"> - <link rel="stylesheet" href="static/bootstrap.min.css" type="text/css"> - <link rel="stylesheet" href="static/bootstrap-responsive.min.css" type="text/css"> - <script src="static/sorttable.js"></script> - <title>@title</title> - <style type="text/css"> - table.sortable thead { cursor: pointer; } - </style> - </head> - - <body> - <div class="container"> - - <!-- HEADER --> - <div class="row"> - <div class="span12"> - <img src="static/spark_logo.png"> - <h1 style="vertical-align: bottom; margin-bottom: 10px; margin-left: 30px; display: inline-block;"> @title </h1> - </div> - </div> - - <hr/> - - @content - - </div> - </body> - -</html>
\ No newline at end of file diff --git a/core/src/main/twirl/spark/deploy/master/app_details.scala.html b/core/src/main/twirl/spark/deploy/master/app_details.scala.html deleted file mode 100644 index 301a7e2124..0000000000 --- a/core/src/main/twirl/spark/deploy/master/app_details.scala.html +++ /dev/null @@ -1,40 +0,0 @@ -@(app: spark.deploy.master.ApplicationInfo) - -@spark.common.html.layout(title = "Application Details") { - - <!-- Application Details --> - <div class="row"> - <div class="span12"> - <ul class="unstyled"> - <li><strong>ID:</strong> @app.id</li> - <li><strong>Description:</strong> @app.desc.name</li> - <li><strong>User:</strong> @app.desc.user</li> - <li><strong>Cores:</strong> - @app.desc.cores - (@app.coresGranted Granted - @if(app.desc.cores == Integer.MAX_VALUE) { - - } else { - , @app.coresLeft - } - ) - </li> - <li><strong>Memory per Slave:</strong> @app.desc.memoryPerSlave</li> - <li><strong>Submit Date:</strong> @app.submitDate</li> - <li><strong>State:</strong> @app.state</li> - </ul> - </div> - </div> - - <hr/> - - <!-- Executors --> - <div class="row"> - <div class="span12"> - <h3> Executor Summary </h3> - <br/> - @executors_table(app.executors.values.toList) - </div> - </div> - -} diff --git a/core/src/main/twirl/spark/deploy/master/app_row.scala.html b/core/src/main/twirl/spark/deploy/master/app_row.scala.html deleted file mode 100644 index feb306f35c..0000000000 --- a/core/src/main/twirl/spark/deploy/master/app_row.scala.html +++ /dev/null @@ -1,20 +0,0 @@ -@(app: spark.deploy.master.ApplicationInfo) - -@import spark.Utils -@import spark.deploy.WebUI.formatDate -@import spark.deploy.WebUI.formatDuration - -<tr> - <td> - <a href="app?appId=@(app.id)">@app.id</a> - </td> - <td>@app.desc.name</td> - <td> - @app.coresGranted - </td> - <td>@Utils.memoryMegabytesToString(app.desc.memoryPerSlave)</td> - <td>@formatDate(app.submitDate)</td> - <td>@app.desc.user</td> - <td>@app.state.toString()</td> - <td>@formatDuration(app.duration)</td> -</tr> diff --git a/core/src/main/twirl/spark/deploy/master/app_table.scala.html b/core/src/main/twirl/spark/deploy/master/app_table.scala.html deleted file mode 100644 index f789cee0f1..0000000000 --- a/core/src/main/twirl/spark/deploy/master/app_table.scala.html +++ /dev/null @@ -1,21 +0,0 @@ -@(apps: Array[spark.deploy.master.ApplicationInfo]) - -<table class="table table-bordered table-striped table-condensed sortable"> - <thead> - <tr> - <th>ID</th> - <th>Description</th> - <th>Cores</th> - <th>Memory per Node</th> - <th>Submit Time</th> - <th>User</th> - <th>State</th> - <th>Duration</th> - </tr> - </thead> - <tbody> - @for(j <- apps) { - @app_row(j) - } - </tbody> -</table> diff --git a/core/src/main/twirl/spark/deploy/master/executor_row.scala.html b/core/src/main/twirl/spark/deploy/master/executor_row.scala.html deleted file mode 100644 index d2d80fad48..0000000000 --- a/core/src/main/twirl/spark/deploy/master/executor_row.scala.html +++ /dev/null @@ -1,15 +0,0 @@ -@(executor: spark.deploy.master.ExecutorInfo) - -<tr> - <td>@executor.id</td> - <td> - <a href="@executor.worker.webUiAddress">@executor.worker.id</href> - </td> - <td>@executor.cores</td> - <td>@executor.memory</td> - <td>@executor.state</td> - <td> - <a href="@(executor.worker.webUiAddress)/log?appId=@(executor.application.id)&executorId=@(executor.id)&logType=stdout">stdout</a> - <a href="@(executor.worker.webUiAddress)/log?appId=@(executor.application.id)&executorId=@(executor.id)&logType=stderr">stderr</a> - </td> -</tr> diff --git a/core/src/main/twirl/spark/deploy/master/executors_table.scala.html b/core/src/main/twirl/spark/deploy/master/executors_table.scala.html deleted file mode 100644 index cafc42c80e..0000000000 --- a/core/src/main/twirl/spark/deploy/master/executors_table.scala.html +++ /dev/null @@ -1,19 +0,0 @@ -@(executors: List[spark.deploy.master.ExecutorInfo]) - -<table class="table table-bordered table-striped table-condensed"> - <thead> - <tr> - <th>ExecutorID</th> - <th>Worker</th> - <th>Cores</th> - <th>Memory</th> - <th>State</th> - <th>Logs</th> - </tr> - </thead> - <tbody> - @for(e <- executors) { - @executor_row(e) - } - </tbody> -</table>
\ No newline at end of file diff --git a/core/src/main/twirl/spark/deploy/master/index.scala.html b/core/src/main/twirl/spark/deploy/master/index.scala.html deleted file mode 100644 index b9b9f08810..0000000000 --- a/core/src/main/twirl/spark/deploy/master/index.scala.html +++ /dev/null @@ -1,53 +0,0 @@ -@(state: spark.deploy.MasterState) -@import spark.deploy.master._ -@import spark.Utils - -@spark.common.html.layout(title = "Spark Master on " + state.host + ":" + state.port) { - - <!-- Cluster Details --> - <div class="row"> - <div class="span12"> - <ul class="unstyled"> - <li><strong>URL:</strong> @(state.uri)</li> - <li><strong>Workers:</strong> @state.workers.size </li> - <li><strong>Cores:</strong> @{state.workers.map(_.cores).sum} Total, - @{state.workers.map(_.coresUsed).sum} Used</li> - <li><strong>Memory:</strong> @{Utils.memoryMegabytesToString(state.workers.map(_.memory).sum)} Total, - @{Utils.memoryMegabytesToString(state.workers.map(_.memoryUsed).sum)} Used</li> - <li><strong>Applications:</strong> @state.activeApps.size Running, @state.completedApps.size Completed </li> - </ul> - </div> - </div> - - <!-- Worker Summary --> - <div class="row"> - <div class="span12"> - <h3> Workers </h3> - <br/> - @worker_table(state.workers.sortBy(_.id)) - </div> - </div> - - <hr/> - - <!-- App Summary (Running) --> - <div class="row"> - <div class="span12"> - <h3> Running Applications </h3> - <br/> - @app_table(state.activeApps.sortBy(_.startTime).reverse) - </div> - </div> - - <hr/> - - <!-- App Summary (Completed) --> - <div class="row"> - <div class="span12"> - <h3> Completed Applications </h3> - <br/> - @app_table(state.completedApps.sortBy(_.endTime).reverse) - </div> - </div> - -} diff --git a/core/src/main/twirl/spark/deploy/master/worker_row.scala.html b/core/src/main/twirl/spark/deploy/master/worker_row.scala.html deleted file mode 100644 index be69e9bf02..0000000000 --- a/core/src/main/twirl/spark/deploy/master/worker_row.scala.html +++ /dev/null @@ -1,14 +0,0 @@ -@(worker: spark.deploy.master.WorkerInfo) - -@import spark.Utils - -<tr> - <td> - <a href="@worker.webUiAddress">@worker.id</href> - </td> - <td>@{worker.host}:@{worker.port}</td> - <td>@worker.state</td> - <td>@worker.cores (@worker.coresUsed Used)</td> - <td>@{Utils.memoryMegabytesToString(worker.memory)} - (@{Utils.memoryMegabytesToString(worker.memoryUsed)} Used)</td> -</tr> diff --git a/core/src/main/twirl/spark/deploy/master/worker_table.scala.html b/core/src/main/twirl/spark/deploy/master/worker_table.scala.html deleted file mode 100644 index b249411a62..0000000000 --- a/core/src/main/twirl/spark/deploy/master/worker_table.scala.html +++ /dev/null @@ -1,18 +0,0 @@ -@(workers: Array[spark.deploy.master.WorkerInfo]) - -<table class="table table-bordered table-striped table-condensed sortable"> - <thead> - <tr> - <th>ID</th> - <th>Address</th> - <th>State</th> - <th>Cores</th> - <th>Memory</th> - </tr> - </thead> - <tbody> - @for(w <- workers) { - @worker_row(w) - } - </tbody> -</table> diff --git a/core/src/main/twirl/spark/deploy/worker/executor_row.scala.html b/core/src/main/twirl/spark/deploy/worker/executor_row.scala.html deleted file mode 100644 index dad0a89080..0000000000 --- a/core/src/main/twirl/spark/deploy/worker/executor_row.scala.html +++ /dev/null @@ -1,20 +0,0 @@ -@(executor: spark.deploy.worker.ExecutorRunner) - -@import spark.Utils - -<tr> - <td>@executor.execId</td> - <td>@executor.cores</td> - <td>@Utils.memoryMegabytesToString(executor.memory)</td> - <td> - <ul class="unstyled"> - <li><strong>ID:</strong> @executor.appId</li> - <li><strong>Name:</strong> @executor.appDesc.name</li> - <li><strong>User:</strong> @executor.appDesc.user</li> - </ul> - </td> - <td> - <a href="log?appId=@(executor.appId)&executorId=@(executor.execId)&logType=stdout">stdout</a> - <a href="log?appId=@(executor.appId)&executorId=@(executor.execId)&logType=stderr">stderr</a> - </td> -</tr> diff --git a/core/src/main/twirl/spark/deploy/worker/executors_table.scala.html b/core/src/main/twirl/spark/deploy/worker/executors_table.scala.html deleted file mode 100644 index 327a2399c7..0000000000 --- a/core/src/main/twirl/spark/deploy/worker/executors_table.scala.html +++ /dev/null @@ -1,18 +0,0 @@ -@(executors: List[spark.deploy.worker.ExecutorRunner]) - -<table class="table table-bordered table-striped table-condensed sortable"> - <thead> - <tr> - <th>ExecutorID</th> - <th>Cores</th> - <th>Memory</th> - <th>Job Details</th> - <th>Logs</th> - </tr> - </thead> - <tbody> - @for(e <- executors) { - @executor_row(e) - } - </tbody> -</table>
\ No newline at end of file diff --git a/core/src/main/twirl/spark/deploy/worker/index.scala.html b/core/src/main/twirl/spark/deploy/worker/index.scala.html deleted file mode 100644 index 0e66af9284..0000000000 --- a/core/src/main/twirl/spark/deploy/worker/index.scala.html +++ /dev/null @@ -1,44 +0,0 @@ -@(worker: spark.deploy.WorkerState) -@import spark.Utils - -@spark.common.html.layout(title = "Spark Worker on " + worker.host + ":" + worker.port) { - - <!-- Worker Details --> - <div class="row"> - <div class="span12"> - <ul class="unstyled"> - <li><strong>ID:</strong> @worker.workerId</li> - <li><strong> - Master URL:</strong> @worker.masterUrl - </li> - <li><strong>Cores:</strong> @worker.cores (@worker.coresUsed Used)</li> - <li><strong>Memory:</strong> @{Utils.memoryMegabytesToString(worker.memory)} - (@{Utils.memoryMegabytesToString(worker.memoryUsed)} Used)</li> - </ul> - <p><a href="@worker.masterWebUiUrl">Back to Master</a></p> - </div> - </div> - - <hr/> - - <!-- Running Executors --> - <div class="row"> - <div class="span12"> - <h3> Running Executors </h3> - <br/> - @executors_table(worker.executors) - </div> - </div> - - <hr/> - - <!-- Finished Executors --> - <div class="row"> - <div class="span12"> - <h3> Finished Executors </h3> - <br/> - @executors_table(worker.finishedExecutors) - </div> - </div> - -} diff --git a/core/src/main/twirl/spark/storage/index.scala.html b/core/src/main/twirl/spark/storage/index.scala.html deleted file mode 100644 index 2b337f6133..0000000000 --- a/core/src/main/twirl/spark/storage/index.scala.html +++ /dev/null @@ -1,40 +0,0 @@ -@(maxMem: Long, remainingMem: Long, diskSpaceUsed: Long, rdds: Array[spark.storage.RDDInfo], storageStatusList: Array[spark.storage.StorageStatus]) -@import spark.Utils - -@spark.common.html.layout(title = "Storage Dashboard") { - - <!-- High-Level Information --> - <div class="row"> - <div class="span12"> - <ul class="unstyled"> - <li><strong>Memory:</strong> - @{Utils.memoryBytesToString(maxMem - remainingMem)} Used - (@{Utils.memoryBytesToString(remainingMem)} Available) </li> - <li><strong>Disk:</strong> @{Utils.memoryBytesToString(diskSpaceUsed)} Used </li> - </ul> - </div> - </div> - - <hr/> - - <!-- RDD Summary --> - <div class="row"> - <div class="span12"> - <h3> RDD Summary </h3> - <br/> - @rdd_table(rdds) - </div> - </div> - - <hr/> - - <!-- Worker Summary --> - <div class="row"> - <div class="span12"> - <h3> Worker Summary </h3> - <br/> - @worker_table(storageStatusList) - </div> - </div> - -}
\ No newline at end of file diff --git a/core/src/main/twirl/spark/storage/rdd.scala.html b/core/src/main/twirl/spark/storage/rdd.scala.html deleted file mode 100644 index d85addeb17..0000000000 --- a/core/src/main/twirl/spark/storage/rdd.scala.html +++ /dev/null @@ -1,81 +0,0 @@ -@(rddInfo: spark.storage.RDDInfo, storageStatusList: Array[spark.storage.StorageStatus]) -@import spark.Utils - -@spark.common.html.layout(title = "RDD Info ") { - - <!-- High-Level Information --> - <div class="row"> - <div class="span12"> - <ul class="unstyled"> - <li> - <strong>Storage Level:</strong> - @(rddInfo.storageLevel.description) - <li> - <strong>Cached Partitions:</strong> - @(rddInfo.numCachedPartitions) - </li> - <li> - <strong>Total Partitions:</strong> - @(rddInfo.numPartitions) - </li> - <li> - <strong>Memory Size:</strong> - @{Utils.memoryBytesToString(rddInfo.memSize)} - </li> - <li> - <strong>Disk Size:</strong> - @{Utils.memoryBytesToString(rddInfo.diskSize)} - </li> - </ul> - </div> - </div> - - <hr/> - - <!-- RDD Summary --> - <div class="row"> - <div class="span12"> - <h3> RDD Summary </h3> - <br/> - - - <!-- Block Table Summary --> - <table class="table table-bordered table-striped table-condensed sortable"> - <thead> - <tr> - <th>Block Name</th> - <th>Storage Level</th> - <th>Size in Memory</th> - <th>Size on Disk</th> - </tr> - </thead> - <tbody> - @storageStatusList.flatMap(_.blocks).toArray.sortWith(_._1 < _._1).map { case (k,v) => - <tr> - <td>@k</td> - <td> - @(v.storageLevel.description) - </td> - <td>@{Utils.memoryBytesToString(v.memSize)}</td> - <td>@{Utils.memoryBytesToString(v.diskSize)}</td> - </tr> - } - </tbody> - </table> - - - </div> - </div> - - <hr/> - - <!-- Worker Table --> - <div class="row"> - <div class="span12"> - <h3> Worker Summary </h3> - <br/> - @worker_table(storageStatusList, "rdd_" + rddInfo.id ) - </div> - </div> - -}
\ No newline at end of file diff --git a/core/src/main/twirl/spark/storage/rdd_table.scala.html b/core/src/main/twirl/spark/storage/rdd_table.scala.html deleted file mode 100644 index a51e64aed0..0000000000 --- a/core/src/main/twirl/spark/storage/rdd_table.scala.html +++ /dev/null @@ -1,32 +0,0 @@ -@(rdds: Array[spark.storage.RDDInfo]) -@import spark.Utils - -<table class="table table-bordered table-striped table-condensed sortable"> - <thead> - <tr> - <th>RDD Name</th> - <th>Storage Level</th> - <th>Cached Partitions</th> - <th>Fraction Partitions Cached</th> - <th>Size in Memory</th> - <th>Size on Disk</th> - </tr> - </thead> - <tbody> - @for(rdd <- rdds) { - <tr> - <td> - <a href="rdd?id=@(rdd.id)"> - @rdd.name - </a> - </td> - <td>@(rdd.storageLevel.description) - </td> - <td>@rdd.numCachedPartitions</td> - <td>@(rdd.numCachedPartitions / rdd.numPartitions.toDouble)</td> - <td>@{Utils.memoryBytesToString(rdd.memSize)}</td> - <td>@{Utils.memoryBytesToString(rdd.diskSize)}</td> - </tr> - } - </tbody> -</table>
\ No newline at end of file diff --git a/core/src/main/twirl/spark/storage/worker_table.scala.html b/core/src/main/twirl/spark/storage/worker_table.scala.html deleted file mode 100644 index cd72a688c1..0000000000 --- a/core/src/main/twirl/spark/storage/worker_table.scala.html +++ /dev/null @@ -1,24 +0,0 @@ -@(workersStatusList: Array[spark.storage.StorageStatus], prefix: String = "") -@import spark.Utils - -<table class="table table-bordered table-striped table-condensed sortable"> - <thead> - <tr> - <th>Host</th> - <th>Memory Usage</th> - <th>Disk Usage</th> - </tr> - </thead> - <tbody> - @for(status <- workersStatusList) { - <tr> - <td>@(status.blockManagerId.host + ":" + status.blockManagerId.port)</td> - <td> - @(Utils.memoryBytesToString(status.memUsed(prefix))) - (@(Utils.memoryBytesToString(status.memRemaining)) Total Available) - </td> - <td>@(Utils.memoryBytesToString(status.diskUsed(prefix)))</td> - </tr> - } - </tbody> -</table>
\ No newline at end of file diff --git a/core/src/test/resources/fairscheduler.xml b/core/src/test/resources/fairscheduler.xml new file mode 100644 index 0000000000..996ffb1864 --- /dev/null +++ b/core/src/test/resources/fairscheduler.xml @@ -0,0 +1,32 @@ +<?xml version="1.0"?> +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one or more + ~ contributor license agreements. See the NOTICE file distributed with + ~ this work for additional information regarding copyright ownership. + ~ The ASF licenses this file to You under the Apache License, Version 2.0 + ~ (the "License"); you may not use this file except in compliance with + ~ the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, software + ~ distributed under the License is distributed on an "AS IS" BASIS, + ~ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + ~ See the License for the specific language governing permissions and + ~ limitations under the License. + --> + +<allocations> +<pool name="1"> + <minShare>2</minShare> + <weight>1</weight> + <schedulingMode>FIFO</schedulingMode> +</pool> +<pool name="2"> + <minShare>3</minShare> + <weight>1</weight> + <schedulingMode>FIFO</schedulingMode> +</pool> +<pool name="3"> +</pool> +</allocations> diff --git a/core/src/test/resources/log4j.properties b/core/src/test/resources/log4j.properties index 6ec89c0184..f6fef03689 100644 --- a/core/src/test/resources/log4j.properties +++ b/core/src/test/resources/log4j.properties @@ -1,4 +1,21 @@ -# Set everything to be logged to the file core/target/unit-tests.log +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Set everything to be logged to the file core/target/unit-tests.log log4j.rootCategory=INFO, file log4j.appender.file=org.apache.log4j.FileAppender log4j.appender.file.append=false @@ -8,3 +25,4 @@ log4j.appender.file.layout.ConversionPattern=%d{yy/MM/dd HH:mm:ss.SSS} %p %c{1}: # Ignore messages below warning level from Jetty, because it's a bit verbose log4j.logger.org.eclipse.jetty=WARN +org.eclipse.jetty.LEVEL=WARN diff --git a/core/src/test/resources/test_metrics_config.properties b/core/src/test/resources/test_metrics_config.properties new file mode 100644 index 0000000000..2b31ddf2eb --- /dev/null +++ b/core/src/test/resources/test_metrics_config.properties @@ -0,0 +1,6 @@ +*.sink.console.period = 10 +*.sink.console.unit = seconds +*.source.jvm.class = spark.metrics.source.JvmSource +master.sink.console.period = 20 +master.sink.console.unit = minutes + diff --git a/core/src/test/resources/test_metrics_system.properties b/core/src/test/resources/test_metrics_system.properties new file mode 100644 index 0000000000..d5479f0298 --- /dev/null +++ b/core/src/test/resources/test_metrics_system.properties @@ -0,0 +1,7 @@ +*.sink.console.period = 10 +*.sink.console.unit = seconds +test.sink.console.class = spark.metrics.sink.ConsoleSink +test.sink.dummy.class = spark.metrics.sink.DummySink +test.source.dummy.class = spark.metrics.source.DummySource +test.sink.console.period = 20 +test.sink.console.unit = minutes diff --git a/core/src/test/scala/spark/AccumulatorSuite.scala b/core/src/test/scala/spark/AccumulatorSuite.scala index ac8ae7d308..0af175f316 100644 --- a/core/src/test/scala/spark/AccumulatorSuite.scala +++ b/core/src/test/scala/spark/AccumulatorSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite diff --git a/core/src/test/scala/spark/BroadcastSuite.scala b/core/src/test/scala/spark/BroadcastSuite.scala index 362a31fb0d..785721ece8 100644 --- a/core/src/test/scala/spark/BroadcastSuite.scala +++ b/core/src/test/scala/spark/BroadcastSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite diff --git a/core/src/test/scala/spark/CheckpointSuite.scala b/core/src/test/scala/spark/CheckpointSuite.scala index ca385972fb..a84c89e3c9 100644 --- a/core/src/test/scala/spark/CheckpointSuite.scala +++ b/core/src/test/scala/spark/CheckpointSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite @@ -27,6 +44,16 @@ class CheckpointSuite extends FunSuite with LocalSparkContext with Logging { } } + test("basic checkpointing") { + val parCollection = sc.makeRDD(1 to 4) + val flatMappedRDD = parCollection.flatMap(x => 1 to x) + flatMappedRDD.checkpoint() + assert(flatMappedRDD.dependencies.head.rdd == parCollection) + val result = flatMappedRDD.collect() + assert(flatMappedRDD.dependencies.head.rdd != parCollection) + assert(flatMappedRDD.collect() === result) + } + test("RDDs with one-to-one dependencies") { testCheckpointing(_.map(x => x.toString)) testCheckpointing(_.flatMap(x => 1 to x)) diff --git a/core/src/test/scala/spark/ClosureCleanerSuite.scala b/core/src/test/scala/spark/ClosureCleanerSuite.scala index b2d0dd4627..7d2831e19c 100644 --- a/core/src/test/scala/spark/ClosureCleanerSuite.scala +++ b/core/src/test/scala/spark/ClosureCleanerSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io.NotSerializableException diff --git a/core/src/test/scala/spark/DistributedSuite.scala b/core/src/test/scala/spark/DistributedSuite.scala index c9b4707def..e11efe459c 100644 --- a/core/src/test/scala/spark/DistributedSuite.scala +++ b/core/src/test/scala/spark/DistributedSuite.scala @@ -1,13 +1,33 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import network.ConnectionManagerId import org.scalatest.FunSuite import org.scalatest.BeforeAndAfter +import org.scalatest.concurrent.Timeouts._ import org.scalatest.matchers.ShouldMatchers import org.scalatest.prop.Checkers +import org.scalatest.time.{Span, Millis} import org.scalacheck.Arbitrary._ import org.scalacheck.Gen import org.scalacheck.Prop._ +import org.eclipse.jetty.server.{Server, Request, Handler} import com.google.common.io.Files @@ -15,8 +35,15 @@ import scala.collection.mutable.ArrayBuffer import SparkContext._ import storage.{GetBlock, BlockManagerWorker, StorageLevel} +import ui.JettyUtils + + +class NotSerializableClass +class NotSerializableExn(val notSer: NotSerializableClass) extends Throwable() {} -class DistributedSuite extends FunSuite with ShouldMatchers with BeforeAndAfter with LocalSparkContext { + +class DistributedSuite extends FunSuite with ShouldMatchers with BeforeAndAfter + with LocalSparkContext { val clusterUrl = "local-cluster[2,1,512]" @@ -25,6 +52,24 @@ class DistributedSuite extends FunSuite with ShouldMatchers with BeforeAndAfter System.clearProperty("spark.storage.memoryFraction") } + test("task throws not serializable exception") { + // Ensures that executors do not crash when an exn is not serializable. If executors crash, + // this test will hang. Correct behavior is that executors don't crash but fail tasks + // and the scheduler throws a SparkException. + + // numSlaves must be less than numPartitions + val numSlaves = 3 + val numPartitions = 10 + + sc = new SparkContext("local-cluster[%s,1,512]".format(numSlaves), "test") + val data = sc.parallelize(1 to 100, numPartitions). + map(x => throw new NotSerializableExn(new NotSerializableClass)) + intercept[SparkException] { + data.count() + } + resetSparkContext() + } + test("local-cluster format") { sc = new SparkContext("local-cluster[2,1,512]", "test") assert(sc.parallelize(1 to 2, 2).count() == 2) @@ -196,7 +241,6 @@ class DistributedSuite extends FunSuite with ShouldMatchers with BeforeAndAfter sc = new SparkContext(clusterUrl, "test") val data = sc.parallelize(Seq(true, true), 2) assert(data.count === 2) // force executors to start - val masterId = SparkEnv.get.blockManager.blockManagerId assert(data.map(markNodeIfIdentity).collect.size === 2) assert(data.map(failOnMarkedIdentity).collect.size === 2) } @@ -252,12 +296,48 @@ class DistributedSuite extends FunSuite with ShouldMatchers with BeforeAndAfter assert(data2.count === 2) } } + + test("unpersist RDDs") { + DistributedSuite.amMaster = true + sc = new SparkContext("local-cluster[3,1,512]", "test") + val data = sc.parallelize(Seq(true, false, false, false), 4) + data.persist(StorageLevel.MEMORY_ONLY_2) + data.count + assert(sc.persistentRdds.isEmpty === false) + data.unpersist() + assert(sc.persistentRdds.isEmpty === true) + + failAfter(Span(3000, Millis)) { + try { + while (! sc.getRDDStorageInfo.isEmpty) { + Thread.sleep(200) + } + } catch { + case _ => { Thread.sleep(10) } + // Do nothing. We might see exceptions because block manager + // is racing this thread to remove entries from the driver. + } + } + } + + test("job should fail if TaskResult exceeds Akka frame size") { + // We must use local-cluster mode since results are returned differently + // when running under LocalScheduler: + sc = new SparkContext("local-cluster[1,1,512]", "test") + val akkaFrameSize = + sc.env.actorSystem.settings.config.getBytes("akka.remote.netty.message-frame-size").toInt + val rdd = sc.parallelize(Seq(1)).map{x => new Array[Byte](akkaFrameSize)} + val exception = intercept[SparkException] { + rdd.reduce((x, y) => x) + } + exception.getMessage should endWith("result exceeded Akka frame size") + } } object DistributedSuite { // Indicates whether this JVM is marked for failure. var mark = false - + // Set by test to remember if we are in the driver program so we can assert // that we are not. var amMaster = false @@ -274,9 +354,9 @@ object DistributedSuite { // Act like an identity function, but if mark was set to true previously, fail, // crashing the entire JVM. def failOnMarkedIdentity(item: Boolean): Boolean = { - if (mark) { + if (mark) { System.exit(42) - } + } item - } + } } diff --git a/core/src/test/scala/spark/DriverSuite.scala b/core/src/test/scala/spark/DriverSuite.scala index 5e84b3a66a..ed16b9d8ef 100644 --- a/core/src/test/scala/spark/DriverSuite.scala +++ b/core/src/test/scala/spark/DriverSuite.scala @@ -1,7 +1,27 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io.File +import org.apache.log4j.Logger +import org.apache.log4j.Level + import org.scalatest.FunSuite import org.scalatest.concurrent.Timeouts import org.scalatest.prop.TableDrivenPropertyChecks._ @@ -27,6 +47,7 @@ class DriverSuite extends FunSuite with Timeouts { */ object DriverWithoutCleanup { def main(args: Array[String]) { + Logger.getRootLogger().setLevel(Level.WARN) val sc = new SparkContext(args(0), "DriverWithoutCleanup") sc.parallelize(1 to 100, 4).count() } diff --git a/core/src/test/scala/spark/FailureSuite.scala b/core/src/test/scala/spark/FailureSuite.scala index 8c1445a465..5b133cdd6e 100644 --- a/core/src/test/scala/spark/FailureSuite.scala +++ b/core/src/test/scala/spark/FailureSuite.scala @@ -1,9 +1,23 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite -import org.scalatest.prop.Checkers - -import scala.collection.mutable.ArrayBuffer import SparkContext._ @@ -23,7 +37,7 @@ object FailureSuiteState { } class FailureSuite extends FunSuite with LocalSparkContext { - + // Run a 3-task map job in which task 1 deterministically fails once, and check // whether the job completes successfully and we ran 4 tasks in total. test("failure in a single-stage job") { @@ -49,7 +63,7 @@ class FailureSuite extends FunSuite with LocalSparkContext { test("failure in a two-stage job") { sc = new SparkContext("local[1,1]", "test") val results = sc.makeRDD(1 to 3).map(x => (x, x)).groupByKey(3).map { - case (k, v) => + case (k, v) => FailureSuiteState.synchronized { FailureSuiteState.tasksRun += 1 if (k == 1 && FailureSuiteState.tasksFailed == 0) { @@ -70,15 +84,43 @@ class FailureSuite extends FunSuite with LocalSparkContext { sc = new SparkContext("local[1,1]", "test") val results = sc.makeRDD(1 to 3).map(x => new NonSerializable) - val thrown = intercept[spark.SparkException] { + val thrown = intercept[SparkException] { results.collect() } - assert(thrown.getClass === classOf[spark.SparkException]) + assert(thrown.getClass === classOf[SparkException]) assert(thrown.getMessage.contains("NotSerializableException")) FailureSuiteState.clear() } + test("failure because task closure is not serializable") { + sc = new SparkContext("local[1,1]", "test") + val a = new NonSerializable + + // Non-serializable closure in the final result stage + val thrown = intercept[SparkException] { + sc.parallelize(1 to 10, 2).map(x => a).count() + } + assert(thrown.getClass === classOf[SparkException]) + assert(thrown.getMessage.contains("NotSerializableException")) + + // Non-serializable closure in an earlier stage + val thrown1 = intercept[SparkException] { + sc.parallelize(1 to 10, 2).map(x => (x, a)).partitionBy(new HashPartitioner(3)).count() + } + assert(thrown1.getClass === classOf[SparkException]) + assert(thrown1.getMessage.contains("NotSerializableException")) + + // Non-serializable closure in foreach function + val thrown2 = intercept[SparkException] { + sc.parallelize(1 to 10, 2).foreach(x => println(a)) + } + assert(thrown2.getClass === classOf[SparkException]) + assert(thrown2.getMessage.contains("NotSerializableException")) + + FailureSuiteState.clear() + } + // TODO: Need to add tests with shuffle fetch failures. } diff --git a/core/src/test/scala/spark/FileServerSuite.scala b/core/src/test/scala/spark/FileServerSuite.scala index f1a35bced3..242ae971f8 100644 --- a/core/src/test/scala/spark/FileServerSuite.scala +++ b/core/src/test/scala/spark/FileServerSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import com.google.common.io.Files @@ -35,26 +52,26 @@ class FileServerSuite extends FunSuite with LocalSparkContext { val testData = Array((1,1), (1,1), (2,1), (3,5), (2,2), (3,0)) val result = sc.parallelize(testData).reduceByKey { val path = SparkFiles.get("FileServerSuite.txt") - val in = new BufferedReader(new FileReader(path)) + val in = new BufferedReader(new FileReader(path)) val fileVal = in.readLine().toInt in.close() _ * fileVal + _ * fileVal - }.collect + }.collect() assert(result.toSet === Set((1,200), (2,300), (3,500))) } test("Distributing files locally using URL as input") { // addFile("file:///....") sc = new SparkContext("local[4]", "test") - sc.addFile((new File(tmpFile.toString)).toURL.toString) + sc.addFile(new File(tmpFile.toString).toURI.toString) val testData = Array((1,1), (1,1), (2,1), (3,5), (2,2), (3,0)) val result = sc.parallelize(testData).reduceByKey { val path = SparkFiles.get("FileServerSuite.txt") - val in = new BufferedReader(new FileReader(path)) + val in = new BufferedReader(new FileReader(path)) val fileVal = in.readLine().toInt in.close() _ * fileVal + _ * fileVal - }.collect + }.collect() assert(result.toSet === Set((1,200), (2,300), (3,500))) } @@ -80,12 +97,11 @@ class FileServerSuite extends FunSuite with LocalSparkContext { val testData = Array((1,1), (1,1), (2,1), (3,5), (2,2), (3,0)) val result = sc.parallelize(testData).reduceByKey { val path = SparkFiles.get("FileServerSuite.txt") - val in = new BufferedReader(new FileReader(path)) + val in = new BufferedReader(new FileReader(path)) val fileVal = in.readLine().toInt in.close() _ * fileVal + _ * fileVal - }.collect - println(result) + }.collect() assert(result.toSet === Set((1,200), (2,300), (3,500))) } diff --git a/core/src/test/scala/spark/FileSuite.scala b/core/src/test/scala/spark/FileSuite.scala index 91b48c7456..1e2c257c4b 100644 --- a/core/src/test/scala/spark/FileSuite.scala +++ b/core/src/test/scala/spark/FileSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.io.{FileWriter, PrintWriter, File} @@ -7,6 +24,8 @@ import scala.io.Source import com.google.common.io.Files import org.scalatest.FunSuite import org.apache.hadoop.io._ +import org.apache.hadoop.io.compress.{DefaultCodec, CompressionCodec, GzipCodec} + import SparkContext._ @@ -26,6 +45,28 @@ class FileSuite extends FunSuite with LocalSparkContext { assert(sc.textFile(outputDir).collect().toList === List("1", "2", "3", "4")) } + test("text files (compressed)") { + sc = new SparkContext("local", "test") + val tempDir = Files.createTempDir() + val normalDir = new File(tempDir, "output_normal").getAbsolutePath + val compressedOutputDir = new File(tempDir, "output_compressed").getAbsolutePath + val codec = new DefaultCodec() + + val data = sc.parallelize("a" * 10000, 1) + data.saveAsTextFile(normalDir) + data.saveAsTextFile(compressedOutputDir, classOf[DefaultCodec]) + + val normalFile = new File(normalDir, "part-00000") + val normalContent = sc.textFile(normalDir).collect + assert(normalContent === Array.fill(10000)("a")) + + val compressedFile = new File(compressedOutputDir, "part-00000" + codec.getDefaultExtension) + val compressedContent = sc.textFile(compressedOutputDir).collect + assert(compressedContent === Array.fill(10000)("a")) + + assert(compressedFile.length < normalFile.length) + } + test("SequenceFiles") { sc = new SparkContext("local", "test") val tempDir = Files.createTempDir() @@ -37,6 +78,28 @@ class FileSuite extends FunSuite with LocalSparkContext { assert(output.map(_.toString).collect().toList === List("(1,a)", "(2,aa)", "(3,aaa)")) } + test("SequenceFile (compressed)") { + sc = new SparkContext("local", "test") + val tempDir = Files.createTempDir() + val normalDir = new File(tempDir, "output_normal").getAbsolutePath + val compressedOutputDir = new File(tempDir, "output_compressed").getAbsolutePath + val codec = new DefaultCodec() + + val data = sc.parallelize(Seq.fill(100)("abc"), 1).map(x => (x, x)) + data.saveAsSequenceFile(normalDir) + data.saveAsSequenceFile(compressedOutputDir, Some(classOf[DefaultCodec])) + + val normalFile = new File(normalDir, "part-00000") + val normalContent = sc.sequenceFile[String, String](normalDir).collect + assert(normalContent === Array.fill(100)("abc", "abc")) + + val compressedFile = new File(compressedOutputDir, "part-00000" + codec.getDefaultExtension) + val compressedContent = sc.sequenceFile[String, String](compressedOutputDir).collect + assert(compressedContent === Array.fill(100)("abc", "abc")) + + assert(compressedFile.length < normalFile.length) + } + test("SequenceFile with writable key") { sc = new SparkContext("local", "test") val tempDir = Files.createTempDir() diff --git a/core/src/test/scala/spark/JavaAPISuite.java b/core/src/test/scala/spark/JavaAPISuite.java index d3dcd3bbeb..5e2bf2d231 100644 --- a/core/src/test/scala/spark/JavaAPISuite.java +++ b/core/src/test/scala/spark/JavaAPISuite.java @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark; import java.io.File; @@ -8,6 +25,7 @@ import java.util.*; import scala.Tuple2; import com.google.common.base.Charsets; +import org.apache.hadoop.io.compress.DefaultCodec; import com.google.common.io.Files; import org.apache.hadoop.io.IntWritable; import org.apache.hadoop.io.Text; @@ -308,7 +326,9 @@ public class JavaAPISuite implements Serializable { Assert.assertEquals(20/6.0, rdd.mean(), 0.01); Assert.assertEquals(20/6.0, rdd.mean(), 0.01); Assert.assertEquals(6.22222, rdd.variance(), 0.01); + Assert.assertEquals(7.46667, rdd.sampleVariance(), 0.01); Assert.assertEquals(2.49444, rdd.stdev(), 0.01); + Assert.assertEquals(2.73252, rdd.sampleStdev(), 0.01); Double first = rdd.first(); List<Double> take = rdd.take(5); @@ -474,6 +494,19 @@ public class JavaAPISuite implements Serializable { } @Test + public void textFilesCompressed() throws IOException { + File tempDir = Files.createTempDir(); + String outputDir = new File(tempDir, "output").getAbsolutePath(); + JavaRDD<Integer> rdd = sc.parallelize(Arrays.asList(1, 2, 3, 4)); + rdd.saveAsTextFile(outputDir, DefaultCodec.class); + + // Try reading it in as a text file RDD + List<String> expected = Arrays.asList("1", "2", "3", "4"); + JavaRDD<String> readRDD = sc.textFile(outputDir); + Assert.assertEquals(expected, readRDD.collect()); + } + + @Test public void sequenceFile() { File tempDir = Files.createTempDir(); String outputDir = new File(tempDir, "output").getAbsolutePath(); @@ -620,6 +653,37 @@ public class JavaAPISuite implements Serializable { } @Test + public void hadoopFileCompressed() { + File tempDir = Files.createTempDir(); + String outputDir = new File(tempDir, "output_compressed").getAbsolutePath(); + List<Tuple2<Integer, String>> pairs = Arrays.asList( + new Tuple2<Integer, String>(1, "a"), + new Tuple2<Integer, String>(2, "aa"), + new Tuple2<Integer, String>(3, "aaa") + ); + JavaPairRDD<Integer, String> rdd = sc.parallelizePairs(pairs); + + rdd.map(new PairFunction<Tuple2<Integer, String>, IntWritable, Text>() { + @Override + public Tuple2<IntWritable, Text> call(Tuple2<Integer, String> pair) { + return new Tuple2<IntWritable, Text>(new IntWritable(pair._1()), new Text(pair._2())); + } + }).saveAsHadoopFile(outputDir, IntWritable.class, Text.class, SequenceFileOutputFormat.class, + DefaultCodec.class); + + JavaPairRDD<IntWritable, Text> output = sc.hadoopFile(outputDir, + SequenceFileInputFormat.class, IntWritable.class, Text.class); + + Assert.assertEquals(pairs.toString(), output.map(new Function<Tuple2<IntWritable, Text>, + String>() { + @Override + public String call(Tuple2<IntWritable, Text> x) { + return x.toString(); + } + }).collect().toString()); + } + + @Test public void zip() { JavaRDD<Integer> rdd = sc.parallelize(Arrays.asList(1, 2, 3, 4, 5)); JavaDoubleRDD doubles = rdd.map(new DoubleFunction<Integer>() { @@ -633,6 +697,32 @@ public class JavaAPISuite implements Serializable { } @Test + public void zipPartitions() { + JavaRDD<Integer> rdd1 = sc.parallelize(Arrays.asList(1, 2, 3, 4, 5, 6), 2); + JavaRDD<String> rdd2 = sc.parallelize(Arrays.asList("1", "2", "3", "4"), 2); + FlatMapFunction2<Iterator<Integer>, Iterator<String>, Integer> sizesFn = + new FlatMapFunction2<Iterator<Integer>, Iterator<String>, Integer>() { + @Override + public Iterable<Integer> call(Iterator<Integer> i, Iterator<String> s) { + int sizeI = 0; + int sizeS = 0; + while (i.hasNext()) { + sizeI += 1; + i.next(); + } + while (s.hasNext()) { + sizeS += 1; + s.next(); + } + return Arrays.asList(sizeI, sizeS); + } + }; + + JavaRDD<Integer> sizes = rdd1.zipPartitions(sizesFn, rdd2); + Assert.assertEquals("[3, 2, 3, 2]", sizes.collect().toString()); + } + + @Test public void accumulators() { JavaRDD<Integer> rdd = sc.parallelize(Arrays.asList(1, 2, 3, 4, 5)); diff --git a/core/src/test/scala/spark/KryoSerializerSuite.scala b/core/src/test/scala/spark/KryoSerializerSuite.scala index 327e2ff848..30d2d5282b 100644 --- a/core/src/test/scala/spark/KryoSerializerSuite.scala +++ b/core/src/test/scala/spark/KryoSerializerSuite.scala @@ -1,13 +1,27 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import scala.collection.mutable -import scala.collection.immutable import org.scalatest.FunSuite import com.esotericsoftware.kryo._ -import SparkContext._ - class KryoSerializerSuite extends FunSuite { test("basic types") { val ser = (new KryoSerializer).newInstance() @@ -36,6 +50,7 @@ class KryoSerializerSuite extends FunSuite { check(Array(true, false, true)) check(Array('a', 'b', 'c')) check(Array[Int]()) + check(Array(Array("1", "2"), Array("1", "2", "3", "4"))) } test("pairs") { @@ -86,7 +101,7 @@ class KryoSerializerSuite extends FunSuite { } test("custom registrator") { - import spark.test._ + import KryoTest._ System.setProperty("spark.kryo.registrator", classOf[MyRegistrator].getName) val ser = (new KryoSerializer).newInstance() @@ -106,14 +121,14 @@ class KryoSerializerSuite extends FunSuite { val hashMap = new java.util.HashMap[String, String] hashMap.put("foo", "bar") check(hashMap) - + System.clearProperty("spark.kryo.registrator") } } -package test { +object KryoTest { case class CaseClass(i: Int, s: String) {} - + class ClassWithNoArgConstructor { var x: Int = 0 override def equals(other: Any) = other match { @@ -137,4 +152,4 @@ package test { k.register(classOf[java.util.HashMap[_, _]]) } } -} +}
\ No newline at end of file diff --git a/core/src/test/scala/spark/LocalSparkContext.scala b/core/src/test/scala/spark/LocalSparkContext.scala index 76d5258b02..ddc212d290 100644 --- a/core/src/test/scala/spark/LocalSparkContext.scala +++ b/core/src/test/scala/spark/LocalSparkContext.scala @@ -1,13 +1,39 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.Suite import org.scalatest.BeforeAndAfterEach +import org.scalatest.BeforeAndAfterAll + +import org.jboss.netty.logging.InternalLoggerFactory +import org.jboss.netty.logging.Slf4JLoggerFactory /** Manages a local `sc` {@link SparkContext} variable, correctly stopping it after each test. */ -trait LocalSparkContext extends BeforeAndAfterEach { self: Suite => +trait LocalSparkContext extends BeforeAndAfterEach with BeforeAndAfterAll { self: Suite => @transient var sc: SparkContext = _ + override def beforeAll() { + InternalLoggerFactory.setDefaultFactory(new Slf4JLoggerFactory()); + super.beforeAll() + } + override def afterEach() { resetSparkContext() super.afterEach() diff --git a/core/src/test/scala/spark/MapOutputTrackerSuite.scala b/core/src/test/scala/spark/MapOutputTrackerSuite.scala index e95818db61..ce6cec0451 100644 --- a/core/src/test/scala/spark/MapOutputTrackerSuite.scala +++ b/core/src/test/scala/spark/MapOutputTrackerSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite @@ -8,7 +25,7 @@ import spark.storage.BlockManagerId import spark.util.AkkaUtils class MapOutputTrackerSuite extends FunSuite with LocalSparkContext { - + test("compressSize") { assert(MapOutputTracker.compressSize(0L) === 0) assert(MapOutputTracker.compressSize(1L) === 1) @@ -45,13 +62,13 @@ class MapOutputTrackerSuite extends FunSuite with LocalSparkContext { val compressedSize10000 = MapOutputTracker.compressSize(10000L) val size1000 = MapOutputTracker.decompressSize(compressedSize1000) val size10000 = MapOutputTracker.decompressSize(compressedSize10000) - tracker.registerMapOutput(10, 0, new MapStatus(BlockManagerId("a", "hostA", 1000), + tracker.registerMapOutput(10, 0, new MapStatus(BlockManagerId("a", "hostA", 1000, 0), Array(compressedSize1000, compressedSize10000))) - tracker.registerMapOutput(10, 1, new MapStatus(BlockManagerId("b", "hostB", 1000), + tracker.registerMapOutput(10, 1, new MapStatus(BlockManagerId("b", "hostB", 1000, 0), Array(compressedSize10000, compressedSize1000))) val statuses = tracker.getServerStatuses(10, 0) - assert(statuses.toSeq === Seq((BlockManagerId("a", "hostA", 1000), size1000), - (BlockManagerId("b", "hostB", 1000), size10000))) + assert(statuses.toSeq === Seq((BlockManagerId("a", "hostA", 1000, 0), size1000), + (BlockManagerId("b", "hostB", 1000, 0), size10000))) tracker.stop() } @@ -64,14 +81,14 @@ class MapOutputTrackerSuite extends FunSuite with LocalSparkContext { val compressedSize10000 = MapOutputTracker.compressSize(10000L) val size1000 = MapOutputTracker.decompressSize(compressedSize1000) val size10000 = MapOutputTracker.decompressSize(compressedSize10000) - tracker.registerMapOutput(10, 0, new MapStatus(BlockManagerId("a", "hostA", 1000), + tracker.registerMapOutput(10, 0, new MapStatus(BlockManagerId("a", "hostA", 1000, 0), Array(compressedSize1000, compressedSize1000, compressedSize1000))) - tracker.registerMapOutput(10, 1, new MapStatus(BlockManagerId("b", "hostB", 1000), + tracker.registerMapOutput(10, 1, new MapStatus(BlockManagerId("b", "hostB", 1000, 0), Array(compressedSize10000, compressedSize1000, compressedSize1000))) // As if we had two simulatenous fetch failures - tracker.unregisterMapOutput(10, 0, BlockManagerId("a", "hostA", 1000)) - tracker.unregisterMapOutput(10, 0, BlockManagerId("a", "hostA", 1000)) + tracker.unregisterMapOutput(10, 0, BlockManagerId("a", "hostA", 1000, 0)) + tracker.unregisterMapOutput(10, 0, BlockManagerId("a", "hostA", 1000, 0)) // The remaining reduce task might try to grab the output despite the shuffle failure; // this should cause it to fail, and the scheduler will ignore the failure due to the @@ -80,19 +97,20 @@ class MapOutputTrackerSuite extends FunSuite with LocalSparkContext { } test("remote fetch") { - val (actorSystem, boundPort) = AkkaUtils.createActorSystem("spark", "localhost", 0) + val hostname = "localhost" + val (actorSystem, boundPort) = AkkaUtils.createActorSystem("spark", hostname, 0) System.setProperty("spark.driver.port", boundPort.toString) // Will be cleared by LocalSparkContext - System.setProperty("spark.hostPort", "localhost:" + boundPort) + System.setProperty("spark.hostPort", hostname + ":" + boundPort) val masterTracker = new MapOutputTracker() masterTracker.trackerActor = actorSystem.actorOf( Props(new MapOutputTrackerActor(masterTracker)), "MapOutputTracker") - - val (slaveSystem, _) = AkkaUtils.createActorSystem("spark-slave", "localhost", 0) + + val (slaveSystem, _) = AkkaUtils.createActorSystem("spark-slave", hostname, 0) val slaveTracker = new MapOutputTracker() slaveTracker.trackerActor = slaveSystem.actorFor( "akka://spark@localhost:" + boundPort + "/user/MapOutputTracker") - + masterTracker.registerShuffle(10, 1) masterTracker.incrementGeneration() slaveTracker.updateGeneration(masterTracker.getGeneration) @@ -101,13 +119,13 @@ class MapOutputTrackerSuite extends FunSuite with LocalSparkContext { val compressedSize1000 = MapOutputTracker.compressSize(1000L) val size1000 = MapOutputTracker.decompressSize(compressedSize1000) masterTracker.registerMapOutput(10, 0, new MapStatus( - BlockManagerId("a", "hostA", 1000), Array(compressedSize1000))) + BlockManagerId("a", "hostA", 1000, 0), Array(compressedSize1000))) masterTracker.incrementGeneration() slaveTracker.updateGeneration(masterTracker.getGeneration) assert(slaveTracker.getServerStatuses(10, 0).toSeq === - Seq((BlockManagerId("a", "hostA", 1000), size1000))) + Seq((BlockManagerId("a", "hostA", 1000, 0), size1000))) - masterTracker.unregisterMapOutput(10, 0, BlockManagerId("a", "hostA", 1000)) + masterTracker.unregisterMapOutput(10, 0, BlockManagerId("a", "hostA", 1000, 0)) masterTracker.incrementGeneration() slaveTracker.updateGeneration(masterTracker.getGeneration) intercept[FetchFailedException] { slaveTracker.getServerStatuses(10, 0) } diff --git a/core/src/test/scala/spark/PairRDDFunctionsSuite.scala b/core/src/test/scala/spark/PairRDDFunctionsSuite.scala new file mode 100644 index 0000000000..b102eaf4e6 --- /dev/null +++ b/core/src/test/scala/spark/PairRDDFunctionsSuite.scala @@ -0,0 +1,304 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark + +import scala.collection.mutable.ArrayBuffer +import scala.collection.mutable.HashSet + +import org.scalatest.FunSuite +import org.scalatest.prop.Checkers +import org.scalacheck.Arbitrary._ +import org.scalacheck.Gen +import org.scalacheck.Prop._ + +import com.google.common.io.Files + +import spark.rdd.ShuffledRDD +import spark.SparkContext._ + +class PairRDDFunctionsSuite extends FunSuite with SharedSparkContext { + test("groupByKey") { + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (2, 1))) + val groups = pairs.groupByKey().collect() + assert(groups.size === 2) + val valuesFor1 = groups.find(_._1 == 1).get._2 + assert(valuesFor1.toList.sorted === List(1, 2, 3)) + val valuesFor2 = groups.find(_._1 == 2).get._2 + assert(valuesFor2.toList.sorted === List(1)) + } + + test("groupByKey with duplicates") { + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) + val groups = pairs.groupByKey().collect() + assert(groups.size === 2) + val valuesFor1 = groups.find(_._1 == 1).get._2 + assert(valuesFor1.toList.sorted === List(1, 1, 2, 3)) + val valuesFor2 = groups.find(_._1 == 2).get._2 + assert(valuesFor2.toList.sorted === List(1)) + } + + test("groupByKey with negative key hash codes") { + val pairs = sc.parallelize(Array((-1, 1), (-1, 2), (-1, 3), (2, 1))) + val groups = pairs.groupByKey().collect() + assert(groups.size === 2) + val valuesForMinus1 = groups.find(_._1 == -1).get._2 + assert(valuesForMinus1.toList.sorted === List(1, 2, 3)) + val valuesFor2 = groups.find(_._1 == 2).get._2 + assert(valuesFor2.toList.sorted === List(1)) + } + + test("groupByKey with many output partitions") { + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (2, 1))) + val groups = pairs.groupByKey(10).collect() + assert(groups.size === 2) + val valuesFor1 = groups.find(_._1 == 1).get._2 + assert(valuesFor1.toList.sorted === List(1, 2, 3)) + val valuesFor2 = groups.find(_._1 == 2).get._2 + assert(valuesFor2.toList.sorted === List(1)) + } + + test("reduceByKey") { + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) + val sums = pairs.reduceByKey(_+_).collect() + assert(sums.toSet === Set((1, 7), (2, 1))) + } + + test("reduceByKey with collectAsMap") { + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) + val sums = pairs.reduceByKey(_+_).collectAsMap() + assert(sums.size === 2) + assert(sums(1) === 7) + assert(sums(2) === 1) + } + + test("reduceByKey with many output partitons") { + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) + val sums = pairs.reduceByKey(_+_, 10).collect() + assert(sums.toSet === Set((1, 7), (2, 1))) + } + + test("reduceByKey with partitioner") { + val p = new Partitioner() { + def numPartitions = 2 + def getPartition(key: Any) = key.asInstanceOf[Int] + } + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 1), (0, 1))).partitionBy(p) + val sums = pairs.reduceByKey(_+_) + assert(sums.collect().toSet === Set((1, 4), (0, 1))) + assert(sums.partitioner === Some(p)) + // count the dependencies to make sure there is only 1 ShuffledRDD + val deps = new HashSet[RDD[_]]() + def visit(r: RDD[_]) { + for (dep <- r.dependencies) { + deps += dep.rdd + visit(dep.rdd) + } + } + visit(sums) + assert(deps.size === 2) // ShuffledRDD, ParallelCollection + } + + test("join") { + val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) + val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) + val joined = rdd1.join(rdd2).collect() + assert(joined.size === 4) + assert(joined.toSet === Set( + (1, (1, 'x')), + (1, (2, 'x')), + (2, (1, 'y')), + (2, (1, 'z')) + )) + } + + test("join all-to-all") { + val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (1, 3))) + val rdd2 = sc.parallelize(Array((1, 'x'), (1, 'y'))) + val joined = rdd1.join(rdd2).collect() + assert(joined.size === 6) + assert(joined.toSet === Set( + (1, (1, 'x')), + (1, (1, 'y')), + (1, (2, 'x')), + (1, (2, 'y')), + (1, (3, 'x')), + (1, (3, 'y')) + )) + } + + test("leftOuterJoin") { + val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) + val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) + val joined = rdd1.leftOuterJoin(rdd2).collect() + assert(joined.size === 5) + assert(joined.toSet === Set( + (1, (1, Some('x'))), + (1, (2, Some('x'))), + (2, (1, Some('y'))), + (2, (1, Some('z'))), + (3, (1, None)) + )) + } + + test("rightOuterJoin") { + val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) + val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) + val joined = rdd1.rightOuterJoin(rdd2).collect() + assert(joined.size === 5) + assert(joined.toSet === Set( + (1, (Some(1), 'x')), + (1, (Some(2), 'x')), + (2, (Some(1), 'y')), + (2, (Some(1), 'z')), + (4, (None, 'w')) + )) + } + + test("join with no matches") { + val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) + val rdd2 = sc.parallelize(Array((4, 'x'), (5, 'y'), (5, 'z'), (6, 'w'))) + val joined = rdd1.join(rdd2).collect() + assert(joined.size === 0) + } + + test("join with many output partitions") { + val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) + val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) + val joined = rdd1.join(rdd2, 10).collect() + assert(joined.size === 4) + assert(joined.toSet === Set( + (1, (1, 'x')), + (1, (2, 'x')), + (2, (1, 'y')), + (2, (1, 'z')) + )) + } + + test("groupWith") { + val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) + val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) + val joined = rdd1.groupWith(rdd2).collect() + assert(joined.size === 4) + assert(joined.toSet === Set( + (1, (ArrayBuffer(1, 2), ArrayBuffer('x'))), + (2, (ArrayBuffer(1), ArrayBuffer('y', 'z'))), + (3, (ArrayBuffer(1), ArrayBuffer())), + (4, (ArrayBuffer(), ArrayBuffer('w'))) + )) + } + + test("zero-partition RDD") { + val emptyDir = Files.createTempDir() + val file = sc.textFile(emptyDir.getAbsolutePath) + assert(file.partitions.size == 0) + assert(file.collect().toList === Nil) + // Test that a shuffle on the file works, because this used to be a bug + assert(file.map(line => (line, 1)).reduceByKey(_ + _).collect().toList === Nil) + } + + test("keys and values") { + val rdd = sc.parallelize(Array((1, "a"), (2, "b"))) + assert(rdd.keys.collect().toList === List(1, 2)) + assert(rdd.values.collect().toList === List("a", "b")) + } + + test("default partitioner uses partition size") { + // specify 2000 partitions + val a = sc.makeRDD(Array(1, 2, 3, 4), 2000) + // do a map, which loses the partitioner + val b = a.map(a => (a, (a * 2).toString)) + // then a group by, and see we didn't revert to 2 partitions + val c = b.groupByKey() + assert(c.partitions.size === 2000) + } + + test("default partitioner uses largest partitioner") { + val a = sc.makeRDD(Array((1, "a"), (2, "b")), 2) + val b = sc.makeRDD(Array((1, "a"), (2, "b")), 2000) + val c = a.join(b) + assert(c.partitions.size === 2000) + } + + test("subtract") { + val a = sc.parallelize(Array(1, 2, 3), 2) + val b = sc.parallelize(Array(2, 3, 4), 4) + val c = a.subtract(b) + assert(c.collect().toSet === Set(1)) + assert(c.partitions.size === a.partitions.size) + } + + test("subtract with narrow dependency") { + // use a deterministic partitioner + val p = new Partitioner() { + def numPartitions = 5 + def getPartition(key: Any) = key.asInstanceOf[Int] + } + // partitionBy so we have a narrow dependency + val a = sc.parallelize(Array((1, "a"), (2, "b"), (3, "c"))).partitionBy(p) + // more partitions/no partitioner so a shuffle dependency + val b = sc.parallelize(Array((2, "b"), (3, "cc"), (4, "d")), 4) + val c = a.subtract(b) + assert(c.collect().toSet === Set((1, "a"), (3, "c"))) + // Ideally we could keep the original partitioner... + assert(c.partitioner === None) + } + + test("subtractByKey") { + val a = sc.parallelize(Array((1, "a"), (1, "a"), (2, "b"), (3, "c")), 2) + val b = sc.parallelize(Array((2, 20), (3, 30), (4, 40)), 4) + val c = a.subtractByKey(b) + assert(c.collect().toSet === Set((1, "a"), (1, "a"))) + assert(c.partitions.size === a.partitions.size) + } + + test("subtractByKey with narrow dependency") { + // use a deterministic partitioner + val p = new Partitioner() { + def numPartitions = 5 + def getPartition(key: Any) = key.asInstanceOf[Int] + } + // partitionBy so we have a narrow dependency + val a = sc.parallelize(Array((1, "a"), (1, "a"), (2, "b"), (3, "c"))).partitionBy(p) + // more partitions/no partitioner so a shuffle dependency + val b = sc.parallelize(Array((2, "b"), (3, "cc"), (4, "d")), 4) + val c = a.subtractByKey(b) + assert(c.collect().toSet === Set((1, "a"), (1, "a"))) + assert(c.partitioner.get === p) + } + + test("foldByKey") { + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) + val sums = pairs.foldByKey(0)(_+_).collect() + assert(sums.toSet === Set((1, 7), (2, 1))) + } + + test("foldByKey with mutable result type") { + val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) + val bufs = pairs.mapValues(v => ArrayBuffer(v)).cache() + // Fold the values using in-place mutation + val sums = bufs.foldByKey(new ArrayBuffer[Int])(_ ++= _).collect() + assert(sums.toSet === Set((1, ArrayBuffer(1, 2, 3, 1)), (2, ArrayBuffer(1)))) + // Check that the mutable objects in the original RDD were not changed + assert(bufs.collect().toSet === Set( + (1, ArrayBuffer(1)), + (1, ArrayBuffer(2)), + (1, ArrayBuffer(3)), + (1, ArrayBuffer(1)), + (2, ArrayBuffer(1)))) + } +} diff --git a/core/src/test/scala/spark/PartitioningSuite.scala b/core/src/test/scala/spark/PartitioningSuite.scala index 60db759c25..b1e0b2b4d0 100644 --- a/core/src/test/scala/spark/PartitioningSuite.scala +++ b/core/src/test/scala/spark/PartitioningSuite.scala @@ -1,13 +1,30 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite - import scala.collection.mutable.ArrayBuffer - import SparkContext._ +import spark.util.StatCounter +import scala.math.abs + +class PartitioningSuite extends FunSuite with SharedSparkContext { -class PartitioningSuite extends FunSuite with LocalSparkContext { - test("HashPartitioner equality") { val p2 = new HashPartitioner(2) val p4 = new HashPartitioner(4) @@ -21,8 +38,6 @@ class PartitioningSuite extends FunSuite with LocalSparkContext { } test("RangePartitioner equality") { - sc = new SparkContext("local", "test") - // Make an RDD where all the elements are the same so that the partition range bounds // are deterministically all the same. val rdd = sc.parallelize(Seq(1, 1, 1, 1)).map(x => (x, x)) @@ -50,7 +65,6 @@ class PartitioningSuite extends FunSuite with LocalSparkContext { } test("HashPartitioner not equal to RangePartitioner") { - sc = new SparkContext("local", "test") val rdd = sc.parallelize(1 to 10).map(x => (x, x)) val rangeP2 = new RangePartitioner(2, rdd) val hashP2 = new HashPartitioner(2) @@ -61,8 +75,6 @@ class PartitioningSuite extends FunSuite with LocalSparkContext { } test("partitioner preservation") { - sc = new SparkContext("local", "test") - val rdd = sc.parallelize(1 to 10, 4).map(x => (x, x)) val grouped2 = rdd.groupByKey(2) @@ -101,7 +113,6 @@ class PartitioningSuite extends FunSuite with LocalSparkContext { } test("partitioning Java arrays should fail") { - sc = new SparkContext("local", "test") val arrs: RDD[Array[Int]] = sc.parallelize(Array(1, 2, 3, 4), 2).map(x => Array(x)) val arrPairs: RDD[(Array[Int], Int)] = sc.parallelize(Array(1, 2, 3, 4), 2).map(x => (Array(x), x)) @@ -120,4 +131,20 @@ class PartitioningSuite extends FunSuite with LocalSparkContext { assert(intercept[SparkException]{ arrPairs.reduceByKeyLocally(_ + _) }.getMessage.contains("array")) assert(intercept[SparkException]{ arrPairs.reduceByKey(_ + _) }.getMessage.contains("array")) } + + test("zero-length partitions should be correctly handled") { + // Create RDD with some consecutive empty partitions (including the "first" one) + val rdd: RDD[Double] = sc + .parallelize(Array(-1.0, -1.0, -1.0, -1.0, 2.0, 4.0, -1.0, -1.0), 8) + .filter(_ >= 0.0) + + // Run the partitions, including the consecutive empty ones, through StatCounter + val stats: StatCounter = rdd.stats(); + assert(abs(6.0 - stats.sum) < 0.01); + assert(abs(6.0/2 - rdd.mean) < 0.01); + assert(abs(1.0 - rdd.variance) < 0.01); + assert(abs(1.0 - rdd.stdev) < 0.01); + + // Add other tests here for classes that should be able to handle empty partitions correctly + } } diff --git a/core/src/test/scala/spark/PipedRDDSuite.scala b/core/src/test/scala/spark/PipedRDDSuite.scala index a6344edf8f..35c04710a3 100644 --- a/core/src/test/scala/spark/PipedRDDSuite.scala +++ b/core/src/test/scala/spark/PipedRDDSuite.scala @@ -1,12 +1,28 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite import SparkContext._ -class PipedRDDSuite extends FunSuite with LocalSparkContext { - +class PipedRDDSuite extends FunSuite with SharedSparkContext { + test("basic pipe") { - sc = new SparkContext("local", "test") val nums = sc.makeRDD(Array(1, 2, 3, 4), 2) val piped = nums.pipe(Seq("cat")) @@ -19,8 +35,45 @@ class PipedRDDSuite extends FunSuite with LocalSparkContext { assert(c(3) === "4") } + test("advanced pipe") { + val nums = sc.makeRDD(Array(1, 2, 3, 4), 2) + val bl = sc.broadcast(List("0")) + + val piped = nums.pipe(Seq("cat"), + Map[String, String](), + (f: String => Unit) => {bl.value.map(f(_));f("\u0001")}, + (i:Int, f: String=> Unit) => f(i + "_")) + + val c = piped.collect() + + assert(c.size === 8) + assert(c(0) === "0") + assert(c(1) === "\u0001") + assert(c(2) === "1_") + assert(c(3) === "2_") + assert(c(4) === "0") + assert(c(5) === "\u0001") + assert(c(6) === "3_") + assert(c(7) === "4_") + + val nums1 = sc.makeRDD(Array("a\t1", "b\t2", "a\t3", "b\t4"), 2) + val d = nums1.groupBy(str=>str.split("\t")(0)). + pipe(Seq("cat"), + Map[String, String](), + (f: String => Unit) => {bl.value.map(f(_));f("\u0001")}, + (i:Tuple2[String, Seq[String]], f: String=> Unit) => {for (e <- i._2){ f(e + "_")}}).collect() + assert(d.size === 8) + assert(d(0) === "0") + assert(d(1) === "\u0001") + assert(d(2) === "b\t2_") + assert(d(3) === "b\t4_") + assert(d(4) === "0") + assert(d(5) === "\u0001") + assert(d(6) === "a\t1_") + assert(d(7) === "a\t3_") + } + test("pipe with env variable") { - sc = new SparkContext("local", "test") val nums = sc.makeRDD(Array(1, 2, 3, 4), 2) val piped = nums.pipe(Seq("printenv", "MY_TEST_ENV"), Map("MY_TEST_ENV" -> "LALALA")) val c = piped.collect() @@ -30,9 +83,8 @@ class PipedRDDSuite extends FunSuite with LocalSparkContext { } test("pipe with non-zero exit status") { - sc = new SparkContext("local", "test") val nums = sc.makeRDD(Array(1, 2, 3, 4), 2) - val piped = nums.pipe("cat nonexistent_file") + val piped = nums.pipe(Seq("cat nonexistent_file", "2>", "/dev/null")) intercept[SparkException] { piped.collect() } diff --git a/core/src/test/scala/spark/RDDSuite.scala b/core/src/test/scala/spark/RDDSuite.scala index 7fbdd44340..cbddf4e523 100644 --- a/core/src/test/scala/spark/RDDSuite.scala +++ b/core/src/test/scala/spark/RDDSuite.scala @@ -1,14 +1,32 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import scala.collection.mutable.HashMap import org.scalatest.FunSuite +import org.scalatest.concurrent.Timeouts._ +import org.scalatest.time.{Span, Millis} import spark.SparkContext._ -import spark.rdd.{CoalescedRDD, CoGroupedRDD, PartitionPruningRDD, ShuffledRDD} +import spark.rdd.{CoalescedRDD, CoGroupedRDD, EmptyRDD, PartitionPruningRDD, ShuffledRDD} -class RDDSuite extends FunSuite with LocalSparkContext { +class RDDSuite extends FunSuite with SharedSparkContext { test("basic operations") { - sc = new SparkContext("local", "test") val nums = sc.makeRDD(Array(1, 2, 3, 4), 2) assert(nums.collect().toList === List(1, 2, 3, 4)) val dups = sc.makeRDD(Array(1, 1, 2, 2, 3, 3, 4, 4), 2) @@ -44,7 +62,6 @@ class RDDSuite extends FunSuite with LocalSparkContext { } test("SparkContext.union") { - sc = new SparkContext("local", "test") val nums = sc.makeRDD(Array(1, 2, 3, 4), 2) assert(sc.union(nums).collect().toList === List(1, 2, 3, 4)) assert(sc.union(nums, nums).collect().toList === List(1, 2, 3, 4, 1, 2, 3, 4)) @@ -53,7 +70,6 @@ class RDDSuite extends FunSuite with LocalSparkContext { } test("aggregate") { - sc = new SparkContext("local", "test") val pairs = sc.makeRDD(Array(("a", 1), ("b", 2), ("a", 2), ("c", 5), ("a", 3))) type StringMap = HashMap[String, Int] val emptyMap = new StringMap { @@ -73,27 +89,7 @@ class RDDSuite extends FunSuite with LocalSparkContext { assert(result.toSet === Set(("a", 6), ("b", 2), ("c", 5))) } - test("basic checkpointing") { - import java.io.File - val checkpointDir = File.createTempFile("temp", "") - checkpointDir.delete() - - sc = new SparkContext("local", "test") - sc.setCheckpointDir(checkpointDir.toString) - val parCollection = sc.makeRDD(1 to 4) - val flatMappedRDD = parCollection.flatMap(x => 1 to x) - flatMappedRDD.checkpoint() - assert(flatMappedRDD.dependencies.head.rdd == parCollection) - val result = flatMappedRDD.collect() - Thread.sleep(1000) - assert(flatMappedRDD.dependencies.head.rdd != parCollection) - assert(flatMappedRDD.collect() === result) - - checkpointDir.deleteOnExit() - } - test("basic caching") { - sc = new SparkContext("local", "test") val rdd = sc.makeRDD(Array(1, 2, 3, 4), 2).cache() assert(rdd.collect().toList === List(1, 2, 3, 4)) assert(rdd.collect().toList === List(1, 2, 3, 4)) @@ -101,7 +97,6 @@ class RDDSuite extends FunSuite with LocalSparkContext { } test("caching with failures") { - sc = new SparkContext("local", "test") val onlySplit = new Partition { override def index: Int = 0 } var shouldFail = true val rdd = new RDD[Int](sc, Nil) { @@ -123,38 +118,26 @@ class RDDSuite extends FunSuite with LocalSparkContext { assert(rdd.collect().toList === List(1, 2, 3, 4)) } - test("cogrouped RDDs") { - sc = new SparkContext("local", "test") - val rdd1 = sc.makeRDD(Array((1, "one"), (1, "another one"), (2, "two"), (3, "three")), 2) - val rdd2 = sc.makeRDD(Array((1, "one1"), (1, "another one1"), (2, "two1")), 2) - - // Use cogroup function - val cogrouped = rdd1.cogroup(rdd2).collectAsMap() - assert(cogrouped(1) === (Seq("one", "another one"), Seq("one1", "another one1"))) - assert(cogrouped(2) === (Seq("two"), Seq("two1"))) - assert(cogrouped(3) === (Seq("three"), Seq())) - - // Construct CoGroupedRDD directly, with map side combine enabled - val cogrouped1 = new CoGroupedRDD[Int]( - Seq(rdd1.asInstanceOf[RDD[(Int, Any)]], rdd2.asInstanceOf[RDD[(Int, Any)]]), - new HashPartitioner(3), - true).collectAsMap() - assert(cogrouped1(1).toSeq === Seq(Seq("one", "another one"), Seq("one1", "another one1"))) - assert(cogrouped1(2).toSeq === Seq(Seq("two"), Seq("two1"))) - assert(cogrouped1(3).toSeq === Seq(Seq("three"), Seq())) - - // Construct CoGroupedRDD directly, with map side combine disabled - val cogrouped2 = new CoGroupedRDD[Int]( - Seq(rdd1.asInstanceOf[RDD[(Int, Any)]], rdd2.asInstanceOf[RDD[(Int, Any)]]), - new HashPartitioner(3), - false).collectAsMap() - assert(cogrouped2(1).toSeq === Seq(Seq("one", "another one"), Seq("one1", "another one1"))) - assert(cogrouped2(2).toSeq === Seq(Seq("two"), Seq("two1"))) - assert(cogrouped2(3).toSeq === Seq(Seq("three"), Seq())) + test("empty RDD") { + val empty = new EmptyRDD[Int](sc) + assert(empty.count === 0) + assert(empty.collect().size === 0) + + val thrown = intercept[UnsupportedOperationException]{ + empty.reduce(_+_) + } + assert(thrown.getMessage.contains("empty")) + + val emptyKv = new EmptyRDD[(Int, Int)](sc) + val rdd = sc.parallelize(1 to 2, 2).map(x => (x, x)) + assert(rdd.join(emptyKv).collect().size === 0) + assert(rdd.rightOuterJoin(emptyKv).collect().size === 0) + assert(rdd.leftOuterJoin(emptyKv).collect().size === 2) + assert(rdd.cogroup(emptyKv).collect().size === 2) + assert(rdd.union(emptyKv).collect().size === 2) } - test("coalesced RDDs") { - sc = new SparkContext("local", "test") + test("cogrouped RDDs") { val data = sc.parallelize(1 to 10, 10) val coalesced1 = data.coalesce(2) @@ -192,7 +175,6 @@ class RDDSuite extends FunSuite with LocalSparkContext { } test("zipped RDDs") { - sc = new SparkContext("local", "test") val nums = sc.makeRDD(Array(1, 2, 3, 4), 2) val zipped = nums.zip(nums.map(_ + 1.0)) assert(zipped.glom().map(_.toList).collect().toList === @@ -204,7 +186,6 @@ class RDDSuite extends FunSuite with LocalSparkContext { } test("partition pruning") { - sc = new SparkContext("local", "test") val data = sc.parallelize(1 to 10, 10) // Note that split number starts from 0, so > 8 means only 10th partition left. val prunedRdd = new PartitionPruningRDD(data, splitNum => splitNum > 8) @@ -216,7 +197,6 @@ class RDDSuite extends FunSuite with LocalSparkContext { test("mapWith") { import java.util.Random - sc = new SparkContext("local", "test") val ones = sc.makeRDD(Array(1, 1, 1, 1, 1, 1), 2) val randoms = ones.mapWith( (index: Int) => new Random(index + 42)) @@ -235,7 +215,6 @@ class RDDSuite extends FunSuite with LocalSparkContext { test("flatMapWith") { import java.util.Random - sc = new SparkContext("local", "test") val ones = sc.makeRDD(Array(1, 1, 1, 1, 1, 1), 2) val randoms = ones.flatMapWith( (index: Int) => new Random(index + 42)) @@ -257,7 +236,6 @@ class RDDSuite extends FunSuite with LocalSparkContext { test("filterWith") { import java.util.Random - sc = new SparkContext("local", "test") val ints = sc.makeRDD(Array(1, 2, 3, 4, 5, 6), 2) val sample = ints.filterWith( (index: Int) => new Random(index + 42)) @@ -273,4 +251,78 @@ class RDDSuite extends FunSuite with LocalSparkContext { assert(sample.size === checkSample.size) for (i <- 0 until sample.size) assert(sample(i) === checkSample(i)) } + + test("top with predefined ordering") { + val nums = Array.range(1, 100000) + val ints = sc.makeRDD(scala.util.Random.shuffle(nums), 2) + val topK = ints.top(5) + assert(topK.size === 5) + assert(topK === nums.reverse.take(5)) + } + + test("top with custom ordering") { + val words = Vector("a", "b", "c", "d") + implicit val ord = implicitly[Ordering[String]].reverse + val rdd = sc.makeRDD(words, 2) + val topK = rdd.top(2) + assert(topK.size === 2) + assert(topK.sorted === Array("b", "a")) + } + + test("takeOrdered with predefined ordering") { + val nums = Array(1, 2, 3, 4, 5, 6, 7, 8, 9, 10) + val rdd = sc.makeRDD(nums, 2) + val sortedLowerK = rdd.takeOrdered(5) + assert(sortedLowerK.size === 5) + assert(sortedLowerK === Array(1, 2, 3, 4, 5)) + } + + test("takeOrdered with custom ordering") { + val nums = Array(1, 2, 3, 4, 5, 6, 7, 8, 9, 10) + implicit val ord = implicitly[Ordering[Int]].reverse + val rdd = sc.makeRDD(nums, 2) + val sortedTopK = rdd.takeOrdered(5) + assert(sortedTopK.size === 5) + assert(sortedTopK === Array(10, 9, 8, 7, 6)) + assert(sortedTopK === nums.sorted(ord).take(5)) + } + + test("takeSample") { + val data = sc.parallelize(1 to 100, 2) + for (seed <- 1 to 5) { + val sample = data.takeSample(withReplacement=false, 20, seed) + assert(sample.size === 20) // Got exactly 20 elements + assert(sample.toSet.size === 20) // Elements are distinct + assert(sample.forall(x => 1 <= x && x <= 100), "elements not in [1, 100]") + } + for (seed <- 1 to 5) { + val sample = data.takeSample(withReplacement=false, 200, seed) + assert(sample.size === 100) // Got only 100 elements + assert(sample.toSet.size === 100) // Elements are distinct + assert(sample.forall(x => 1 <= x && x <= 100), "elements not in [1, 100]") + } + for (seed <- 1 to 5) { + val sample = data.takeSample(withReplacement=true, 20, seed) + assert(sample.size === 20) // Got exactly 20 elements + assert(sample.forall(x => 1 <= x && x <= 100), "elements not in [1, 100]") + } + for (seed <- 1 to 5) { + val sample = data.takeSample(withReplacement=true, 100, seed) + assert(sample.size === 100) // Got exactly 100 elements + // Chance of getting all distinct elements is astronomically low, so test we got < 100 + assert(sample.toSet.size < 100, "sampling with replacement returned all distinct elements") + } + for (seed <- 1 to 5) { + val sample = data.takeSample(withReplacement=true, 200, seed) + assert(sample.size === 200) // Got exactly 200 elements + // Chance of getting all distinct elements is still quite low, so test we got < 100 + assert(sample.toSet.size < 100, "sampling with replacement returned all distinct elements") + } + } + + test("runJob on an invalid partition") { + intercept[IllegalArgumentException] { + sc.runJob(sc.parallelize(1 to 10, 2), {iter: Iterator[Int] => iter.size}, Seq(0, 1, 2), false) + } + } } diff --git a/core/src/test/scala/spark/SharedSparkContext.scala b/core/src/test/scala/spark/SharedSparkContext.scala new file mode 100644 index 0000000000..70c24515be --- /dev/null +++ b/core/src/test/scala/spark/SharedSparkContext.scala @@ -0,0 +1,42 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark + +import org.scalatest.Suite +import org.scalatest.BeforeAndAfterAll + +/** Shares a local `SparkContext` between all tests in a suite and closes it at the end */ +trait SharedSparkContext extends BeforeAndAfterAll { self: Suite => + + @transient private var _sc: SparkContext = _ + + def sc: SparkContext = _sc + + override def beforeAll() { + _sc = new SparkContext("local", "test") + super.beforeAll() + } + + override def afterAll() { + if (_sc != null) { + LocalSparkContext.stop(_sc) + _sc = null + } + super.afterAll() + } +} diff --git a/core/src/test/scala/spark/ShuffleNettySuite.scala b/core/src/test/scala/spark/ShuffleNettySuite.scala new file mode 100644 index 0000000000..6bad6c1d13 --- /dev/null +++ b/core/src/test/scala/spark/ShuffleNettySuite.scala @@ -0,0 +1,34 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark + +import org.scalatest.BeforeAndAfterAll + + +class ShuffleNettySuite extends ShuffleSuite with BeforeAndAfterAll { + + // This test suite should run all tests in ShuffleSuite with Netty shuffle mode. + + override def beforeAll(configMap: Map[String, Any]) { + System.setProperty("spark.shuffle.use.netty", "true") + } + + override def afterAll(configMap: Map[String, Any]) { + System.setProperty("spark.shuffle.use.netty", "false") + } +} diff --git a/core/src/test/scala/spark/ShuffleSuite.scala b/core/src/test/scala/spark/ShuffleSuite.scala index 2b2a90defa..752e4b85e6 100644 --- a/core/src/test/scala/spark/ShuffleSuite.scala +++ b/core/src/test/scala/spark/ShuffleSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import scala.collection.mutable.ArrayBuffer @@ -16,54 +33,9 @@ import spark.rdd.ShuffledRDD import spark.SparkContext._ class ShuffleSuite extends FunSuite with ShouldMatchers with LocalSparkContext { - - test("groupByKey") { - sc = new SparkContext("local", "test") - val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (2, 1))) - val groups = pairs.groupByKey().collect() - assert(groups.size === 2) - val valuesFor1 = groups.find(_._1 == 1).get._2 - assert(valuesFor1.toList.sorted === List(1, 2, 3)) - val valuesFor2 = groups.find(_._1 == 2).get._2 - assert(valuesFor2.toList.sorted === List(1)) - } - - test("groupByKey with duplicates") { - sc = new SparkContext("local", "test") - val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) - val groups = pairs.groupByKey().collect() - assert(groups.size === 2) - val valuesFor1 = groups.find(_._1 == 1).get._2 - assert(valuesFor1.toList.sorted === List(1, 1, 2, 3)) - val valuesFor2 = groups.find(_._1 == 2).get._2 - assert(valuesFor2.toList.sorted === List(1)) - } - - test("groupByKey with negative key hash codes") { - sc = new SparkContext("local", "test") - val pairs = sc.parallelize(Array((-1, 1), (-1, 2), (-1, 3), (2, 1))) - val groups = pairs.groupByKey().collect() - assert(groups.size === 2) - val valuesForMinus1 = groups.find(_._1 == -1).get._2 - assert(valuesForMinus1.toList.sorted === List(1, 2, 3)) - val valuesFor2 = groups.find(_._1 == 2).get._2 - assert(valuesFor2.toList.sorted === List(1)) - } - - test("groupByKey with many output partitions") { - sc = new SparkContext("local", "test") - val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (2, 1))) - val groups = pairs.groupByKey(10).collect() - assert(groups.size === 2) - val valuesFor1 = groups.find(_._1 == 1).get._2 - assert(valuesFor1.toList.sorted === List(1, 2, 3)) - val valuesFor2 = groups.find(_._1 == 2).get._2 - assert(valuesFor2.toList.sorted === List(1)) - } - - test("groupByKey with compression") { + test("groupByKey without compression") { try { - System.setProperty("spark.blockManager.compress", "true") + System.setProperty("spark.shuffle.compress", "false") sc = new SparkContext("local", "test") val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (2, 1)), 4) val groups = pairs.groupByKey(4).collect() @@ -73,243 +45,104 @@ class ShuffleSuite extends FunSuite with ShouldMatchers with LocalSparkContext { val valuesFor2 = groups.find(_._1 == 2).get._2 assert(valuesFor2.toList.sorted === List(1)) } finally { - System.setProperty("spark.blockManager.compress", "false") + System.setProperty("spark.shuffle.compress", "true") } } - test("reduceByKey") { - sc = new SparkContext("local", "test") - val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) - val sums = pairs.reduceByKey(_+_).collect() - assert(sums.toSet === Set((1, 7), (2, 1))) - } - - test("reduceByKey with collectAsMap") { - sc = new SparkContext("local", "test") - val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) - val sums = pairs.reduceByKey(_+_).collectAsMap() - assert(sums.size === 2) - assert(sums(1) === 7) - assert(sums(2) === 1) - } + test("shuffle non-zero block size") { + sc = new SparkContext("local-cluster[2,1,512]", "test") + val NUM_BLOCKS = 3 - test("reduceByKey with many output partitons") { - sc = new SparkContext("local", "test") - val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1))) - val sums = pairs.reduceByKey(_+_, 10).collect() - assert(sums.toSet === Set((1, 7), (2, 1))) - } - - test("reduceByKey with partitioner") { - sc = new SparkContext("local", "test") - val p = new Partitioner() { - def numPartitions = 2 - def getPartition(key: Any) = key.asInstanceOf[Int] + val a = sc.parallelize(1 to 10, 2) + val b = a.map { x => + (x, new ShuffleSuite.NonJavaSerializableClass(x * 2)) } - val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 1), (0, 1))).partitionBy(p) - val sums = pairs.reduceByKey(_+_) - assert(sums.collect().toSet === Set((1, 4), (0, 1))) - assert(sums.partitioner === Some(p)) - // count the dependencies to make sure there is only 1 ShuffledRDD - val deps = new HashSet[RDD[_]]() - def visit(r: RDD[_]) { - for (dep <- r.dependencies) { - deps += dep.rdd - visit(dep.rdd) - } + // If the Kryo serializer is not used correctly, the shuffle would fail because the + // default Java serializer cannot handle the non serializable class. + val c = new ShuffledRDD(b, new HashPartitioner(NUM_BLOCKS), + classOf[spark.KryoSerializer].getName) + val shuffleId = c.dependencies.head.asInstanceOf[ShuffleDependency[Int, Int]].shuffleId + + assert(c.count === 10) + + // All blocks must have non-zero size + (0 until NUM_BLOCKS).foreach { id => + val statuses = SparkEnv.get.mapOutputTracker.getServerStatuses(shuffleId, id) + assert(statuses.forall(s => s._2 > 0)) } - visit(sums) - assert(deps.size === 2) // ShuffledRDD, ParallelCollection - } - - test("join") { - sc = new SparkContext("local", "test") - val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) - val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) - val joined = rdd1.join(rdd2).collect() - assert(joined.size === 4) - assert(joined.toSet === Set( - (1, (1, 'x')), - (1, (2, 'x')), - (2, (1, 'y')), - (2, (1, 'z')) - )) } - test("join all-to-all") { - sc = new SparkContext("local", "test") - val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (1, 3))) - val rdd2 = sc.parallelize(Array((1, 'x'), (1, 'y'))) - val joined = rdd1.join(rdd2).collect() - assert(joined.size === 6) - assert(joined.toSet === Set( - (1, (1, 'x')), - (1, (1, 'y')), - (1, (2, 'x')), - (1, (2, 'y')), - (1, (3, 'x')), - (1, (3, 'y')) - )) + test("shuffle serializer") { + // Use a local cluster with 2 processes to make sure there are both local and remote blocks + sc = new SparkContext("local-cluster[2,1,512]", "test") + val a = sc.parallelize(1 to 10, 2) + val b = a.map { x => + (x, new ShuffleSuite.NonJavaSerializableClass(x * 2)) + } + // If the Kryo serializer is not used correctly, the shuffle would fail because the + // default Java serializer cannot handle the non serializable class. + val c = new ShuffledRDD(b, new HashPartitioner(3), classOf[spark.KryoSerializer].getName) + assert(c.count === 10) } - test("leftOuterJoin") { - sc = new SparkContext("local", "test") - val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) - val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) - val joined = rdd1.leftOuterJoin(rdd2).collect() - assert(joined.size === 5) - assert(joined.toSet === Set( - (1, (1, Some('x'))), - (1, (2, Some('x'))), - (2, (1, Some('y'))), - (2, (1, Some('z'))), - (3, (1, None)) - )) - } + test("zero sized blocks") { + // Use a local cluster with 2 processes to make sure there are both local and remote blocks + sc = new SparkContext("local-cluster[2,1,512]", "test") - test("rightOuterJoin") { - sc = new SparkContext("local", "test") - val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) - val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) - val joined = rdd1.rightOuterJoin(rdd2).collect() - assert(joined.size === 5) - assert(joined.toSet === Set( - (1, (Some(1), 'x')), - (1, (Some(2), 'x')), - (2, (Some(1), 'y')), - (2, (Some(1), 'z')), - (4, (None, 'w')) - )) - } + // 10 partitions from 4 keys + val NUM_BLOCKS = 10 + val a = sc.parallelize(1 to 4, NUM_BLOCKS) + val b = a.map(x => (x, x*2)) - test("join with no matches") { - sc = new SparkContext("local", "test") - val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) - val rdd2 = sc.parallelize(Array((4, 'x'), (5, 'y'), (5, 'z'), (6, 'w'))) - val joined = rdd1.join(rdd2).collect() - assert(joined.size === 0) - } + // NOTE: The default Java serializer doesn't create zero-sized blocks. + // So, use Kryo + val c = new ShuffledRDD(b, new HashPartitioner(10), classOf[spark.KryoSerializer].getName) - test("join with many output partitions") { - sc = new SparkContext("local", "test") - val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) - val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) - val joined = rdd1.join(rdd2, 10).collect() - assert(joined.size === 4) - assert(joined.toSet === Set( - (1, (1, 'x')), - (1, (2, 'x')), - (2, (1, 'y')), - (2, (1, 'z')) - )) - } + val shuffleId = c.dependencies.head.asInstanceOf[ShuffleDependency[Int, Int]].shuffleId + assert(c.count === 4) - test("groupWith") { - sc = new SparkContext("local", "test") - val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1))) - val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w'))) - val joined = rdd1.groupWith(rdd2).collect() - assert(joined.size === 4) - assert(joined.toSet === Set( - (1, (ArrayBuffer(1, 2), ArrayBuffer('x'))), - (2, (ArrayBuffer(1), ArrayBuffer('y', 'z'))), - (3, (ArrayBuffer(1), ArrayBuffer())), - (4, (ArrayBuffer(), ArrayBuffer('w'))) - )) - } + val blockSizes = (0 until NUM_BLOCKS).flatMap { id => + val statuses = SparkEnv.get.mapOutputTracker.getServerStatuses(shuffleId, id) + statuses.map(x => x._2) + } + val nonEmptyBlocks = blockSizes.filter(x => x > 0) - test("zero-partition RDD") { - sc = new SparkContext("local", "test") - val emptyDir = Files.createTempDir() - val file = sc.textFile(emptyDir.getAbsolutePath) - assert(file.partitions.size == 0) - assert(file.collect().toList === Nil) - // Test that a shuffle on the file works, because this used to be a bug - assert(file.map(line => (line, 1)).reduceByKey(_ + _).collect().toList === Nil) + // We should have at most 4 non-zero sized partitions + assert(nonEmptyBlocks.size <= 4) } - test("keys and values") { - sc = new SparkContext("local", "test") - val rdd = sc.parallelize(Array((1, "a"), (2, "b"))) - assert(rdd.keys.collect().toList === List(1, 2)) - assert(rdd.values.collect().toList === List("a", "b")) - } + test("zero sized blocks without kryo") { + // Use a local cluster with 2 processes to make sure there are both local and remote blocks + sc = new SparkContext("local-cluster[2,1,512]", "test") - test("default partitioner uses partition size") { - sc = new SparkContext("local", "test") - // specify 2000 partitions - val a = sc.makeRDD(Array(1, 2, 3, 4), 2000) - // do a map, which loses the partitioner - val b = a.map(a => (a, (a * 2).toString)) - // then a group by, and see we didn't revert to 2 partitions - val c = b.groupByKey() - assert(c.partitions.size === 2000) - } + // 10 partitions from 4 keys + val NUM_BLOCKS = 10 + val a = sc.parallelize(1 to 4, NUM_BLOCKS) + val b = a.map(x => (x, x*2)) - test("default partitioner uses largest partitioner") { - sc = new SparkContext("local", "test") - val a = sc.makeRDD(Array((1, "a"), (2, "b")), 2) - val b = sc.makeRDD(Array((1, "a"), (2, "b")), 2000) - val c = a.join(b) - assert(c.partitions.size === 2000) - } + // NOTE: The default Java serializer should create zero-sized blocks + val c = new ShuffledRDD(b, new HashPartitioner(10)) - test("subtract") { - sc = new SparkContext("local", "test") - val a = sc.parallelize(Array(1, 2, 3), 2) - val b = sc.parallelize(Array(2, 3, 4), 4) - val c = a.subtract(b) - assert(c.collect().toSet === Set(1)) - assert(c.partitions.size === a.partitions.size) - } + val shuffleId = c.dependencies.head.asInstanceOf[ShuffleDependency[Int, Int]].shuffleId + assert(c.count === 4) - test("subtract with narrow dependency") { - sc = new SparkContext("local", "test") - // use a deterministic partitioner - val p = new Partitioner() { - def numPartitions = 5 - def getPartition(key: Any) = key.asInstanceOf[Int] + val blockSizes = (0 until NUM_BLOCKS).flatMap { id => + val statuses = SparkEnv.get.mapOutputTracker.getServerStatuses(shuffleId, id) + statuses.map(x => x._2) } - // partitionBy so we have a narrow dependency - val a = sc.parallelize(Array((1, "a"), (2, "b"), (3, "c"))).partitionBy(p) - // more partitions/no partitioner so a shuffle dependency - val b = sc.parallelize(Array((2, "b"), (3, "cc"), (4, "d")), 4) - val c = a.subtract(b) - assert(c.collect().toSet === Set((1, "a"), (3, "c"))) - // Ideally we could keep the original partitioner... - assert(c.partitioner === None) - } - - test("subtractByKey") { - sc = new SparkContext("local", "test") - val a = sc.parallelize(Array((1, "a"), (1, "a"), (2, "b"), (3, "c")), 2) - val b = sc.parallelize(Array((2, 20), (3, 30), (4, 40)), 4) - val c = a.subtractByKey(b) - assert(c.collect().toSet === Set((1, "a"), (1, "a"))) - assert(c.partitions.size === a.partitions.size) - } + val nonEmptyBlocks = blockSizes.filter(x => x > 0) - test("subtractByKey with narrow dependency") { - sc = new SparkContext("local", "test") - // use a deterministic partitioner - val p = new Partitioner() { - def numPartitions = 5 - def getPartition(key: Any) = key.asInstanceOf[Int] - } - // partitionBy so we have a narrow dependency - val a = sc.parallelize(Array((1, "a"), (1, "a"), (2, "b"), (3, "c"))).partitionBy(p) - // more partitions/no partitioner so a shuffle dependency - val b = sc.parallelize(Array((2, "b"), (3, "cc"), (4, "d")), 4) - val c = a.subtractByKey(b) - assert(c.collect().toSet === Set((1, "a"), (1, "a"))) - assert(c.partitioner.get === p) + // We should have at most 4 non-zero sized partitions + assert(nonEmptyBlocks.size <= 4) } - } object ShuffleSuite { + def mergeCombineException(x: Int, y: Int): Int = { throw new SparkException("Exception for map-side combine.") x + y } + + class NonJavaSerializableClass(val value: Int) } diff --git a/core/src/test/scala/spark/SizeEstimatorSuite.scala b/core/src/test/scala/spark/SizeEstimatorSuite.scala index e235ef2f67..1ef812dfbd 100644 --- a/core/src/test/scala/spark/SizeEstimatorSuite.scala +++ b/core/src/test/scala/spark/SizeEstimatorSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite @@ -35,7 +52,7 @@ class SizeEstimatorSuite var oldOops: String = _ override def beforeAll() { - // Set the arch to 64-bit and compressedOops to true to get a deterministic test-case + // Set the arch to 64-bit and compressedOops to true to get a deterministic test-case oldArch = System.setProperty("os.arch", "amd64") oldOops = System.setProperty("spark.test.useCompressedOops", "true") } @@ -46,54 +63,54 @@ class SizeEstimatorSuite } test("simple classes") { - expect(16)(SizeEstimator.estimate(new DummyClass1)) - expect(16)(SizeEstimator.estimate(new DummyClass2)) - expect(24)(SizeEstimator.estimate(new DummyClass3)) - expect(24)(SizeEstimator.estimate(new DummyClass4(null))) - expect(48)(SizeEstimator.estimate(new DummyClass4(new DummyClass3))) + assert(SizeEstimator.estimate(new DummyClass1) === 16) + assert(SizeEstimator.estimate(new DummyClass2) === 16) + assert(SizeEstimator.estimate(new DummyClass3) === 24) + assert(SizeEstimator.estimate(new DummyClass4(null)) === 24) + assert(SizeEstimator.estimate(new DummyClass4(new DummyClass3)) === 48) } // NOTE: The String class definition varies across JDK versions (1.6 vs. 1.7) and vendors // (Sun vs IBM). Use a DummyString class to make tests deterministic. test("strings") { - expect(40)(SizeEstimator.estimate(DummyString(""))) - expect(48)(SizeEstimator.estimate(DummyString("a"))) - expect(48)(SizeEstimator.estimate(DummyString("ab"))) - expect(56)(SizeEstimator.estimate(DummyString("abcdefgh"))) + assert(SizeEstimator.estimate(DummyString("")) === 40) + assert(SizeEstimator.estimate(DummyString("a")) === 48) + assert(SizeEstimator.estimate(DummyString("ab")) === 48) + assert(SizeEstimator.estimate(DummyString("abcdefgh")) === 56) } test("primitive arrays") { - expect(32)(SizeEstimator.estimate(new Array[Byte](10))) - expect(40)(SizeEstimator.estimate(new Array[Char](10))) - expect(40)(SizeEstimator.estimate(new Array[Short](10))) - expect(56)(SizeEstimator.estimate(new Array[Int](10))) - expect(96)(SizeEstimator.estimate(new Array[Long](10))) - expect(56)(SizeEstimator.estimate(new Array[Float](10))) - expect(96)(SizeEstimator.estimate(new Array[Double](10))) - expect(4016)(SizeEstimator.estimate(new Array[Int](1000))) - expect(8016)(SizeEstimator.estimate(new Array[Long](1000))) + assert(SizeEstimator.estimate(new Array[Byte](10)) === 32) + assert(SizeEstimator.estimate(new Array[Char](10)) === 40) + assert(SizeEstimator.estimate(new Array[Short](10)) === 40) + assert(SizeEstimator.estimate(new Array[Int](10)) === 56) + assert(SizeEstimator.estimate(new Array[Long](10)) === 96) + assert(SizeEstimator.estimate(new Array[Float](10)) === 56) + assert(SizeEstimator.estimate(new Array[Double](10)) === 96) + assert(SizeEstimator.estimate(new Array[Int](1000)) === 4016) + assert(SizeEstimator.estimate(new Array[Long](1000)) === 8016) } test("object arrays") { // Arrays containing nulls should just have one pointer per element - expect(56)(SizeEstimator.estimate(new Array[String](10))) - expect(56)(SizeEstimator.estimate(new Array[AnyRef](10))) + assert(SizeEstimator.estimate(new Array[String](10)) === 56) + assert(SizeEstimator.estimate(new Array[AnyRef](10)) === 56) // For object arrays with non-null elements, each object should take one pointer plus // however many bytes that class takes. (Note that Array.fill calls the code in its // second parameter separately for each object, so we get distinct objects.) - expect(216)(SizeEstimator.estimate(Array.fill(10)(new DummyClass1))) - expect(216)(SizeEstimator.estimate(Array.fill(10)(new DummyClass2))) - expect(296)(SizeEstimator.estimate(Array.fill(10)(new DummyClass3))) - expect(56)(SizeEstimator.estimate(Array(new DummyClass1, new DummyClass2))) + assert(SizeEstimator.estimate(Array.fill(10)(new DummyClass1)) === 216) + assert(SizeEstimator.estimate(Array.fill(10)(new DummyClass2)) === 216) + assert(SizeEstimator.estimate(Array.fill(10)(new DummyClass3)) === 296) + assert(SizeEstimator.estimate(Array(new DummyClass1, new DummyClass2)) === 56) // Past size 100, our samples 100 elements, but we should still get the right size. - expect(28016)(SizeEstimator.estimate(Array.fill(1000)(new DummyClass3))) + assert(SizeEstimator.estimate(Array.fill(1000)(new DummyClass3)) === 28016) // If an array contains the *same* element many times, we should only count it once. val d1 = new DummyClass1 - expect(72)(SizeEstimator.estimate(Array.fill(10)(d1))) // 10 pointers plus 8-byte object - expect(432)(SizeEstimator.estimate(Array.fill(100)(d1))) // 100 pointers plus 8-byte object + assert(SizeEstimator.estimate(Array.fill(10)(d1)) === 72) // 10 pointers plus 8-byte object + assert(SizeEstimator.estimate(Array.fill(100)(d1)) === 432) // 100 pointers plus 8-byte object // Same thing with huge array containing the same element many times. Note that this won't // return exactly 4032 because it can't tell that *all* the elements will equal the first @@ -111,10 +128,10 @@ class SizeEstimatorSuite val initialize = PrivateMethod[Unit]('initialize) SizeEstimator invokePrivate initialize() - expect(40)(SizeEstimator.estimate(DummyString(""))) - expect(48)(SizeEstimator.estimate(DummyString("a"))) - expect(48)(SizeEstimator.estimate(DummyString("ab"))) - expect(56)(SizeEstimator.estimate(DummyString("abcdefgh"))) + assert(SizeEstimator.estimate(DummyString("")) === 40) + assert(SizeEstimator.estimate(DummyString("a")) === 48) + assert(SizeEstimator.estimate(DummyString("ab")) === 48) + assert(SizeEstimator.estimate(DummyString("abcdefgh")) === 56) resetOrClear("os.arch", arch) } @@ -128,10 +145,10 @@ class SizeEstimatorSuite val initialize = PrivateMethod[Unit]('initialize) SizeEstimator invokePrivate initialize() - expect(56)(SizeEstimator.estimate(DummyString(""))) - expect(64)(SizeEstimator.estimate(DummyString("a"))) - expect(64)(SizeEstimator.estimate(DummyString("ab"))) - expect(72)(SizeEstimator.estimate(DummyString("abcdefgh"))) + assert(SizeEstimator.estimate(DummyString("")) === 56) + assert(SizeEstimator.estimate(DummyString("a")) === 64) + assert(SizeEstimator.estimate(DummyString("ab")) === 64) + assert(SizeEstimator.estimate(DummyString("abcdefgh")) === 72) resetOrClear("os.arch", arch) resetOrClear("spark.test.useCompressedOops", oops) diff --git a/core/src/test/scala/spark/SortingSuite.scala b/core/src/test/scala/spark/SortingSuite.scala index 495f957e53..b933c4aab8 100644 --- a/core/src/test/scala/spark/SortingSuite.scala +++ b/core/src/test/scala/spark/SortingSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import org.scalatest.FunSuite @@ -5,16 +22,14 @@ import org.scalatest.BeforeAndAfter import org.scalatest.matchers.ShouldMatchers import SparkContext._ -class SortingSuite extends FunSuite with LocalSparkContext with ShouldMatchers with Logging { - +class SortingSuite extends FunSuite with SharedSparkContext with ShouldMatchers with Logging { + test("sortByKey") { - sc = new SparkContext("local", "test") val pairs = sc.parallelize(Array((1, 0), (2, 0), (0, 0), (3, 0)), 2) - assert(pairs.sortByKey().collect() === Array((0,0), (1,0), (2,0), (3,0))) + assert(pairs.sortByKey().collect() === Array((0,0), (1,0), (2,0), (3,0))) } test("large array") { - sc = new SparkContext("local", "test") val rand = new scala.util.Random() val pairArr = Array.fill(1000) { (rand.nextInt(), rand.nextInt()) } val pairs = sc.parallelize(pairArr, 2) @@ -24,7 +39,6 @@ class SortingSuite extends FunSuite with LocalSparkContext with ShouldMatchers w } test("large array with one split") { - sc = new SparkContext("local", "test") val rand = new scala.util.Random() val pairArr = Array.fill(1000) { (rand.nextInt(), rand.nextInt()) } val pairs = sc.parallelize(pairArr, 2) @@ -32,9 +46,8 @@ class SortingSuite extends FunSuite with LocalSparkContext with ShouldMatchers w assert(sorted.partitions.size === 1) assert(sorted.collect() === pairArr.sortBy(_._1)) } - + test("large array with many partitions") { - sc = new SparkContext("local", "test") val rand = new scala.util.Random() val pairArr = Array.fill(1000) { (rand.nextInt(), rand.nextInt()) } val pairs = sc.parallelize(pairArr, 2) @@ -42,9 +55,8 @@ class SortingSuite extends FunSuite with LocalSparkContext with ShouldMatchers w assert(sorted.partitions.size === 20) assert(sorted.collect() === pairArr.sortBy(_._1)) } - + test("sort descending") { - sc = new SparkContext("local", "test") val rand = new scala.util.Random() val pairArr = Array.fill(1000) { (rand.nextInt(), rand.nextInt()) } val pairs = sc.parallelize(pairArr, 2) @@ -52,15 +64,13 @@ class SortingSuite extends FunSuite with LocalSparkContext with ShouldMatchers w } test("sort descending with one split") { - sc = new SparkContext("local", "test") val rand = new scala.util.Random() val pairArr = Array.fill(1000) { (rand.nextInt(), rand.nextInt()) } val pairs = sc.parallelize(pairArr, 1) assert(pairs.sortByKey(false, 1).collect() === pairArr.sortWith((x, y) => x._1 > y._1)) } - + test("sort descending with many partitions") { - sc = new SparkContext("local", "test") val rand = new scala.util.Random() val pairArr = Array.fill(1000) { (rand.nextInt(), rand.nextInt()) } val pairs = sc.parallelize(pairArr, 2) @@ -68,7 +78,6 @@ class SortingSuite extends FunSuite with LocalSparkContext with ShouldMatchers w } test("more partitions than elements") { - sc = new SparkContext("local", "test") val rand = new scala.util.Random() val pairArr = Array.fill(10) { (rand.nextInt(), rand.nextInt()) } val pairs = sc.parallelize(pairArr, 30) @@ -76,14 +85,12 @@ class SortingSuite extends FunSuite with LocalSparkContext with ShouldMatchers w } test("empty RDD") { - sc = new SparkContext("local", "test") val pairArr = new Array[(Int, Int)](0) val pairs = sc.parallelize(pairArr, 2) assert(pairs.sortByKey().collect() === pairArr.sortBy(_._1)) } test("partition balancing") { - sc = new SparkContext("local", "test") val pairArr = (1 to 1000).map(x => (x, x)).toArray val sorted = sc.parallelize(pairArr, 4).sortByKey() assert(sorted.collect() === pairArr.sortBy(_._1)) @@ -99,7 +106,6 @@ class SortingSuite extends FunSuite with LocalSparkContext with ShouldMatchers w } test("partition balancing for descending sort") { - sc = new SparkContext("local", "test") val pairArr = (1 to 1000).map(x => (x, x)).toArray val sorted = sc.parallelize(pairArr, 4).sortByKey(false) assert(sorted.collect() === pairArr.sortBy(_._1).reverse) diff --git a/core/src/test/scala/spark/SparkContextInfoSuite.scala b/core/src/test/scala/spark/SparkContextInfoSuite.scala new file mode 100644 index 0000000000..6d50bf5e1b --- /dev/null +++ b/core/src/test/scala/spark/SparkContextInfoSuite.scala @@ -0,0 +1,60 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark + +import org.scalatest.FunSuite +import spark.SparkContext._ + +class SparkContextInfoSuite extends FunSuite with LocalSparkContext { + test("getPersistentRDDs only returns RDDs that are marked as cached") { + sc = new SparkContext("local", "test") + assert(sc.getPersistentRDDs.isEmpty === true) + + val rdd = sc.makeRDD(Array(1, 2, 3, 4), 2) + assert(sc.getPersistentRDDs.isEmpty === true) + + rdd.cache() + assert(sc.getPersistentRDDs.size === 1) + assert(sc.getPersistentRDDs.values.head === rdd) + } + + test("getPersistentRDDs returns an immutable map") { + sc = new SparkContext("local", "test") + val rdd1 = sc.makeRDD(Array(1, 2, 3, 4), 2).cache() + + val myRdds = sc.getPersistentRDDs + assert(myRdds.size === 1) + assert(myRdds.values.head === rdd1) + + val rdd2 = sc.makeRDD(Array(5, 6, 7, 8), 1).cache() + + // getPersistentRDDs should have 2 RDDs, but myRdds should not change + assert(sc.getPersistentRDDs.size === 2) + assert(myRdds.size === 1) + } + + test("getRDDStorageInfo only reports on RDDs that actually persist data") { + sc = new SparkContext("local", "test") + val rdd = sc.makeRDD(Array(1, 2, 3, 4), 2).cache() + + assert(sc.getRDDStorageInfo.size === 0) + + rdd.collect() + assert(sc.getRDDStorageInfo.size === 1) + } +}
\ No newline at end of file diff --git a/core/src/test/scala/spark/ThreadingSuite.scala b/core/src/test/scala/spark/ThreadingSuite.scala index ff315b6693..f2acd0bd3c 100644 --- a/core/src/test/scala/spark/ThreadingSuite.scala +++ b/core/src/test/scala/spark/ThreadingSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import java.util.concurrent.Semaphore diff --git a/core/src/test/scala/spark/UnpersistSuite.scala b/core/src/test/scala/spark/UnpersistSuite.scala new file mode 100644 index 0000000000..93977d16f4 --- /dev/null +++ b/core/src/test/scala/spark/UnpersistSuite.scala @@ -0,0 +1,47 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark + +import org.scalatest.FunSuite +import org.scalatest.concurrent.Timeouts._ +import org.scalatest.time.{Span, Millis} +import spark.SparkContext._ + +class UnpersistSuite extends FunSuite with LocalSparkContext { + test("unpersist RDD") { + sc = new SparkContext("local", "test") + val rdd = sc.makeRDD(Array(1, 2, 3, 4), 2).cache() + rdd.count + assert(sc.persistentRdds.isEmpty === false) + rdd.unpersist() + assert(sc.persistentRdds.isEmpty === true) + + failAfter(Span(3000, Millis)) { + try { + while (! sc.getRDDStorageInfo.isEmpty) { + Thread.sleep(200) + } + } catch { + case _ => { Thread.sleep(10) } + // Do nothing. We might see exceptions because block manager + // is racing this thread to remove entries from the driver. + } + } + assert(sc.getRDDStorageInfo.isEmpty === true) + } +} diff --git a/core/src/test/scala/spark/UtilsSuite.scala b/core/src/test/scala/spark/UtilsSuite.scala index ed4701574f..31c3b25c50 100644 --- a/core/src/test/scala/spark/UtilsSuite.scala +++ b/core/src/test/scala/spark/UtilsSuite.scala @@ -1,7 +1,27 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark +import com.google.common.base.Charsets +import com.google.common.io.Files +import java.io.{ByteArrayOutputStream, ByteArrayInputStream, FileOutputStream, File} import org.scalatest.FunSuite -import java.io.{ByteArrayOutputStream, ByteArrayInputStream} +import org.apache.commons.io.FileUtils import scala.util.Random class UtilsSuite extends FunSuite { @@ -27,24 +47,93 @@ class UtilsSuite extends FunSuite { assert(os.toByteArray.toList.equals(bytes.toList)) } - test("memoryStringToMb"){ - assert(Utils.memoryStringToMb("1") == 0) - assert(Utils.memoryStringToMb("1048575") == 0) - assert(Utils.memoryStringToMb("3145728") == 3) + test("memoryStringToMb") { + assert(Utils.memoryStringToMb("1") === 0) + assert(Utils.memoryStringToMb("1048575") === 0) + assert(Utils.memoryStringToMb("3145728") === 3) + + assert(Utils.memoryStringToMb("1024k") === 1) + assert(Utils.memoryStringToMb("5000k") === 4) + assert(Utils.memoryStringToMb("4024k") === Utils.memoryStringToMb("4024K")) + + assert(Utils.memoryStringToMb("1024m") === 1024) + assert(Utils.memoryStringToMb("5000m") === 5000) + assert(Utils.memoryStringToMb("4024m") === Utils.memoryStringToMb("4024M")) + + assert(Utils.memoryStringToMb("2g") === 2048) + assert(Utils.memoryStringToMb("3g") === Utils.memoryStringToMb("3G")) + + assert(Utils.memoryStringToMb("2t") === 2097152) + assert(Utils.memoryStringToMb("3t") === Utils.memoryStringToMb("3T")) + } + + test("splitCommandString") { + assert(Utils.splitCommandString("") === Seq()) + assert(Utils.splitCommandString("a") === Seq("a")) + assert(Utils.splitCommandString("aaa") === Seq("aaa")) + assert(Utils.splitCommandString("a b c") === Seq("a", "b", "c")) + assert(Utils.splitCommandString(" a b\t c ") === Seq("a", "b", "c")) + assert(Utils.splitCommandString("a 'b c'") === Seq("a", "b c")) + assert(Utils.splitCommandString("a 'b c' d") === Seq("a", "b c", "d")) + assert(Utils.splitCommandString("'b c'") === Seq("b c")) + assert(Utils.splitCommandString("a \"b c\"") === Seq("a", "b c")) + assert(Utils.splitCommandString("a \"b c\" d") === Seq("a", "b c", "d")) + assert(Utils.splitCommandString("\"b c\"") === Seq("b c")) + assert(Utils.splitCommandString("a 'b\" c' \"d' e\"") === Seq("a", "b\" c", "d' e")) + assert(Utils.splitCommandString("a\t'b\nc'\nd") === Seq("a", "b\nc", "d")) + assert(Utils.splitCommandString("a \"b\\\\c\"") === Seq("a", "b\\c")) + assert(Utils.splitCommandString("a \"b\\\"c\"") === Seq("a", "b\"c")) + assert(Utils.splitCommandString("a 'b\\\"c'") === Seq("a", "b\\\"c")) + assert(Utils.splitCommandString("'a'b") === Seq("ab")) + assert(Utils.splitCommandString("'a''b'") === Seq("ab")) + assert(Utils.splitCommandString("\"a\"b") === Seq("ab")) + assert(Utils.splitCommandString("\"a\"\"b\"") === Seq("ab")) + assert(Utils.splitCommandString("''") === Seq("")) + assert(Utils.splitCommandString("\"\"") === Seq("")) + } + + test("string formatting of time durations") { + val second = 1000 + val minute = second * 60 + val hour = minute * 60 + def str = Utils.msDurationToString(_) + + assert(str(123) === "123 ms") + assert(str(second) === "1.0 s") + assert(str(second + 462) === "1.5 s") + assert(str(hour) === "1.00 h") + assert(str(minute) === "1.0 m") + assert(str(minute + 4 * second + 34) === "1.1 m") + assert(str(10 * hour + minute + 4 * second) === "10.02 h") + assert(str(10 * hour + 59 * minute + 59 * second + 999) === "11.00 h") + } + + test("reading offset bytes of a file") { + val tmpDir2 = Files.createTempDir() + val f1Path = tmpDir2 + "/f1" + val f1 = new FileOutputStream(f1Path) + f1.write("1\n2\n3\n4\n5\n6\n7\n8\n9\n".getBytes(Charsets.UTF_8)) + f1.close() + + // Read first few bytes + assert(Utils.offsetBytes(f1Path, 0, 5) === "1\n2\n3") + + // Read some middle bytes + assert(Utils.offsetBytes(f1Path, 4, 11) === "3\n4\n5\n6") + + // Read last few bytes + assert(Utils.offsetBytes(f1Path, 12, 18) === "7\n8\n9\n") - assert(Utils.memoryStringToMb("1024k") == 1) - assert(Utils.memoryStringToMb("5000k") == 4) - assert(Utils.memoryStringToMb("4024k") == Utils.memoryStringToMb("4024K")) + // Read some nonexistent bytes in the beginning + assert(Utils.offsetBytes(f1Path, -5, 5) === "1\n2\n3") - assert(Utils.memoryStringToMb("1024m") == 1024) - assert(Utils.memoryStringToMb("5000m") == 5000) - assert(Utils.memoryStringToMb("4024m") == Utils.memoryStringToMb("4024M")) + // Read some nonexistent bytes at the end + assert(Utils.offsetBytes(f1Path, 12, 22) === "7\n8\n9\n") - assert(Utils.memoryStringToMb("2g") == 2048) - assert(Utils.memoryStringToMb("3g") == Utils.memoryStringToMb("3G")) + // Read some nonexistent bytes on both ends + assert(Utils.offsetBytes(f1Path, -3, 25) === "1\n2\n3\n4\n5\n6\n7\n8\n9\n") - assert(Utils.memoryStringToMb("2t") == 2097152) - assert(Utils.memoryStringToMb("3t") == Utils.memoryStringToMb("3T")) + FileUtils.deleteDirectory(tmpDir2) } } diff --git a/core/src/test/scala/spark/ZippedPartitionsSuite.scala b/core/src/test/scala/spark/ZippedPartitionsSuite.scala index 5f60aa75d7..5e6d7b09d8 100644 --- a/core/src/test/scala/spark/ZippedPartitionsSuite.scala +++ b/core/src/test/scala/spark/ZippedPartitionsSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark import scala.collection.immutable.NumericRange @@ -17,9 +34,8 @@ object ZippedPartitionsSuite { } } -class ZippedPartitionsSuite extends FunSuite with LocalSparkContext { +class ZippedPartitionsSuite extends FunSuite with SharedSparkContext { test("print sizes") { - sc = new SparkContext("local", "test") val data1 = sc.makeRDD(Array(1, 2, 3, 4), 2) val data2 = sc.makeRDD(Array("1", "2", "3", "4", "5", "6"), 2) val data3 = sc.makeRDD(Array(1.0, 2.0), 2) diff --git a/core/src/test/scala/spark/metrics/MetricsConfigSuite.scala b/core/src/test/scala/spark/metrics/MetricsConfigSuite.scala new file mode 100644 index 0000000000..87cd2ffad2 --- /dev/null +++ b/core/src/test/scala/spark/metrics/MetricsConfigSuite.scala @@ -0,0 +1,64 @@ +package spark.metrics + +import java.util.Properties +import java.io.{File, FileOutputStream} + +import org.scalatest.{BeforeAndAfter, FunSuite} + +import spark.metrics._ + +class MetricsConfigSuite extends FunSuite with BeforeAndAfter { + var filePath: String = _ + + before { + filePath = getClass.getClassLoader.getResource("test_metrics_config.properties").getFile() + } + + test("MetricsConfig with default properties") { + val conf = new MetricsConfig(Option("dummy-file")) + conf.initialize() + + assert(conf.properties.size() === 0) + assert(conf.properties.getProperty("test-for-dummy") === null) + + val property = conf.getInstance("random") + assert(property.size() === 0) + } + + test("MetricsConfig with properties set") { + val conf = new MetricsConfig(Option(filePath)) + conf.initialize() + + val masterProp = conf.getInstance("master") + assert(masterProp.size() === 3) + assert(masterProp.getProperty("sink.console.period") === "20") + assert(masterProp.getProperty("sink.console.unit") === "minutes") + assert(masterProp.getProperty("source.jvm.class") === "spark.metrics.source.JvmSource") + + val workerProp = conf.getInstance("worker") + assert(workerProp.size() === 3) + assert(workerProp.getProperty("sink.console.period") === "10") + assert(workerProp.getProperty("sink.console.unit") === "seconds") + assert(masterProp.getProperty("source.jvm.class") === "spark.metrics.source.JvmSource") + } + + test("MetricsConfig with subProperties") { + val conf = new MetricsConfig(Option(filePath)) + conf.initialize() + + val propCategories = conf.propertyCategories + assert(propCategories.size === 2) + + val masterProp = conf.getInstance("master") + val sourceProps = conf.subProperties(masterProp, MetricsSystem.SOURCE_REGEX) + assert(sourceProps.size === 1) + assert(sourceProps("jvm").getProperty("class") === "spark.metrics.source.JvmSource") + + val sinkProps = conf.subProperties(masterProp, MetricsSystem.SINK_REGEX) + assert(sinkProps.size === 1) + assert(sinkProps.contains("console")) + + val consoleProps = sinkProps("console") + assert(consoleProps.size() === 2) + } +} diff --git a/core/src/test/scala/spark/metrics/MetricsSystemSuite.scala b/core/src/test/scala/spark/metrics/MetricsSystemSuite.scala new file mode 100644 index 0000000000..c189996417 --- /dev/null +++ b/core/src/test/scala/spark/metrics/MetricsSystemSuite.scala @@ -0,0 +1,39 @@ +package spark.metrics + +import java.util.Properties +import java.io.{File, FileOutputStream} + +import org.scalatest.{BeforeAndAfter, FunSuite} + +import spark.metrics._ + +class MetricsSystemSuite extends FunSuite with BeforeAndAfter { + var filePath: String = _ + + before { + filePath = getClass.getClassLoader.getResource("test_metrics_system.properties").getFile() + System.setProperty("spark.metrics.conf", filePath) + } + + test("MetricsSystem with default config") { + val metricsSystem = MetricsSystem.createMetricsSystem("default") + val sources = metricsSystem.sources + val sinks = metricsSystem.sinks + + assert(sources.length === 0) + assert(sinks.length === 0) + } + + test("MetricsSystem with sources add") { + val metricsSystem = MetricsSystem.createMetricsSystem("test") + val sources = metricsSystem.sources + val sinks = metricsSystem.sinks + + assert(sources.length === 0) + assert(sinks.length === 1) + + val source = new spark.deploy.master.MasterSource(null) + metricsSystem.registerSource(source) + assert(sources.length === 1) + } +} diff --git a/core/src/test/scala/spark/rdd/JdbcRDDSuite.scala b/core/src/test/scala/spark/rdd/JdbcRDDSuite.scala new file mode 100644 index 0000000000..dc8ca941c1 --- /dev/null +++ b/core/src/test/scala/spark/rdd/JdbcRDDSuite.scala @@ -0,0 +1,73 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark + +import org.scalatest.{ BeforeAndAfter, FunSuite } +import spark.SparkContext._ +import spark.rdd.JdbcRDD +import java.sql._ + +class JdbcRDDSuite extends FunSuite with BeforeAndAfter with LocalSparkContext { + + before { + Class.forName("org.apache.derby.jdbc.EmbeddedDriver") + val conn = DriverManager.getConnection("jdbc:derby:target/JdbcRDDSuiteDb;create=true") + try { + val create = conn.createStatement + create.execute(""" + CREATE TABLE FOO( + ID INTEGER NOT NULL GENERATED ALWAYS AS IDENTITY (START WITH 1, INCREMENT BY 1), + DATA INTEGER + )""") + create.close + val insert = conn.prepareStatement("INSERT INTO FOO(DATA) VALUES(?)") + (1 to 100).foreach { i => + insert.setInt(1, i * 2) + insert.executeUpdate + } + insert.close + } catch { + case e: SQLException if e.getSQLState == "X0Y32" => + // table exists + } finally { + conn.close + } + } + + test("basic functionality") { + sc = new SparkContext("local", "test") + val rdd = new JdbcRDD( + sc, + () => { DriverManager.getConnection("jdbc:derby:target/JdbcRDDSuiteDb") }, + "SELECT DATA FROM FOO WHERE ? <= ID AND ID <= ?", + 1, 100, 3, + (r: ResultSet) => { r.getInt(1) } ).cache + + assert(rdd.count === 100) + assert(rdd.reduce(_+_) === 10100) + } + + after { + try { + DriverManager.getConnection("jdbc:derby:;shutdown=true") + } catch { + case se: SQLException if se.getSQLState == "XJ015" => + // normal shutdown + } + } +} diff --git a/core/src/test/scala/spark/rdd/ParallelCollectionSplitSuite.scala b/core/src/test/scala/spark/rdd/ParallelCollectionSplitSuite.scala index d27a2538e4..d1276d541f 100644 --- a/core/src/test/scala/spark/rdd/ParallelCollectionSplitSuite.scala +++ b/core/src/test/scala/spark/rdd/ParallelCollectionSplitSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.rdd import scala.collection.immutable.NumericRange diff --git a/core/src/test/scala/spark/scheduler/ClusterSchedulerSuite.scala b/core/src/test/scala/spark/scheduler/ClusterSchedulerSuite.scala new file mode 100644 index 0000000000..05afcd6567 --- /dev/null +++ b/core/src/test/scala/spark/scheduler/ClusterSchedulerSuite.scala @@ -0,0 +1,267 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler + +import org.scalatest.FunSuite +import org.scalatest.BeforeAndAfter + +import spark._ +import spark.scheduler._ +import spark.scheduler.cluster._ +import scala.collection.mutable.ArrayBuffer + +import java.util.Properties + +class DummyTaskSetManager( + initPriority: Int, + initStageId: Int, + initNumTasks: Int, + clusterScheduler: ClusterScheduler, + taskSet: TaskSet) + extends ClusterTaskSetManager(clusterScheduler, taskSet) { + + parent = null + weight = 1 + minShare = 2 + runningTasks = 0 + priority = initPriority + stageId = initStageId + name = "TaskSet_"+stageId + override val numTasks = initNumTasks + tasksFinished = 0 + + override def increaseRunningTasks(taskNum: Int) { + runningTasks += taskNum + if (parent != null) { + parent.increaseRunningTasks(taskNum) + } + } + + override def decreaseRunningTasks(taskNum: Int) { + runningTasks -= taskNum + if (parent != null) { + parent.decreaseRunningTasks(taskNum) + } + } + + override def addSchedulable(schedulable: Schedulable) { + } + + override def removeSchedulable(schedulable: Schedulable) { + } + + override def getSchedulableByName(name: String): Schedulable = { + return null + } + + override def executorLost(executorId: String, host: String): Unit = { + } + + override def slaveOffer(execId: String, host: String, avaiableCpus: Double, overrideLocality: TaskLocality.TaskLocality = null): Option[TaskDescription] = { + if (tasksFinished + runningTasks < numTasks) { + increaseRunningTasks(1) + return Some(new TaskDescription(0, execId, "task 0:0", null)) + } + return None + } + + override def checkSpeculatableTasks(): Boolean = { + return true + } + + def taskFinished() { + decreaseRunningTasks(1) + tasksFinished +=1 + if (tasksFinished == numTasks) { + parent.removeSchedulable(this) + } + } + + def abort() { + decreaseRunningTasks(runningTasks) + parent.removeSchedulable(this) + } +} + +class DummyTask(stageId: Int) extends Task[Int](stageId) +{ + def run(attemptId: Long): Int = { + return 0 + } +} + +class ClusterSchedulerSuite extends FunSuite with LocalSparkContext with Logging { + + def createDummyTaskSetManager(priority: Int, stage: Int, numTasks: Int, cs: ClusterScheduler, taskSet: TaskSet): DummyTaskSetManager = { + new DummyTaskSetManager(priority, stage, numTasks, cs , taskSet) + } + + def resourceOffer(rootPool: Pool): Int = { + val taskSetQueue = rootPool.getSortedTaskSetQueue() + /* Just for Test*/ + for (manager <- taskSetQueue) { + logInfo("parentName:%s, parent running tasks:%d, name:%s,runningTasks:%d".format(manager.parent.name, manager.parent.runningTasks, manager.name, manager.runningTasks)) + } + for (taskSet <- taskSetQueue) { + taskSet.slaveOffer("execId_1", "hostname_1", 1) match { + case Some(task) => + return taskSet.stageId + case None => {} + } + } + -1 + } + + def checkTaskSetId(rootPool: Pool, expectedTaskSetId: Int) { + assert(resourceOffer(rootPool) === expectedTaskSetId) + } + + test("FIFO Scheduler Test") { + sc = new SparkContext("local", "ClusterSchedulerSuite") + val clusterScheduler = new ClusterScheduler(sc) + var tasks = ArrayBuffer[Task[_]]() + val task = new DummyTask(0) + tasks += task + val taskSet = new TaskSet(tasks.toArray,0,0,0,null) + + val rootPool = new Pool("", SchedulingMode.FIFO, 0, 0) + val schedulableBuilder = new FIFOSchedulableBuilder(rootPool) + schedulableBuilder.buildPools() + + val taskSetManager0 = createDummyTaskSetManager(0, 0, 2, clusterScheduler, taskSet) + val taskSetManager1 = createDummyTaskSetManager(0, 1, 2, clusterScheduler, taskSet) + val taskSetManager2 = createDummyTaskSetManager(0, 2, 2, clusterScheduler, taskSet) + schedulableBuilder.addTaskSetManager(taskSetManager0, null) + schedulableBuilder.addTaskSetManager(taskSetManager1, null) + schedulableBuilder.addTaskSetManager(taskSetManager2, null) + + checkTaskSetId(rootPool, 0) + resourceOffer(rootPool) + checkTaskSetId(rootPool, 1) + resourceOffer(rootPool) + taskSetManager1.abort() + checkTaskSetId(rootPool, 2) + } + + test("Fair Scheduler Test") { + sc = new SparkContext("local", "ClusterSchedulerSuite") + val clusterScheduler = new ClusterScheduler(sc) + var tasks = ArrayBuffer[Task[_]]() + val task = new DummyTask(0) + tasks += task + val taskSet = new TaskSet(tasks.toArray,0,0,0,null) + + val xmlPath = getClass.getClassLoader.getResource("fairscheduler.xml").getFile() + System.setProperty("spark.fairscheduler.allocation.file", xmlPath) + val rootPool = new Pool("", SchedulingMode.FAIR, 0, 0) + val schedulableBuilder = new FairSchedulableBuilder(rootPool) + schedulableBuilder.buildPools() + + assert(rootPool.getSchedulableByName("default") != null) + assert(rootPool.getSchedulableByName("1") != null) + assert(rootPool.getSchedulableByName("2") != null) + assert(rootPool.getSchedulableByName("3") != null) + assert(rootPool.getSchedulableByName("1").minShare === 2) + assert(rootPool.getSchedulableByName("1").weight === 1) + assert(rootPool.getSchedulableByName("2").minShare === 3) + assert(rootPool.getSchedulableByName("2").weight === 1) + assert(rootPool.getSchedulableByName("3").minShare === 2) + assert(rootPool.getSchedulableByName("3").weight === 1) + + val properties1 = new Properties() + properties1.setProperty("spark.scheduler.cluster.fair.pool","1") + val properties2 = new Properties() + properties2.setProperty("spark.scheduler.cluster.fair.pool","2") + + val taskSetManager10 = createDummyTaskSetManager(1, 0, 1, clusterScheduler, taskSet) + val taskSetManager11 = createDummyTaskSetManager(1, 1, 1, clusterScheduler, taskSet) + val taskSetManager12 = createDummyTaskSetManager(1, 2, 2, clusterScheduler, taskSet) + schedulableBuilder.addTaskSetManager(taskSetManager10, properties1) + schedulableBuilder.addTaskSetManager(taskSetManager11, properties1) + schedulableBuilder.addTaskSetManager(taskSetManager12, properties1) + + val taskSetManager23 = createDummyTaskSetManager(2, 3, 2, clusterScheduler, taskSet) + val taskSetManager24 = createDummyTaskSetManager(2, 4, 2, clusterScheduler, taskSet) + schedulableBuilder.addTaskSetManager(taskSetManager23, properties2) + schedulableBuilder.addTaskSetManager(taskSetManager24, properties2) + + checkTaskSetId(rootPool, 0) + checkTaskSetId(rootPool, 3) + checkTaskSetId(rootPool, 3) + checkTaskSetId(rootPool, 1) + checkTaskSetId(rootPool, 4) + checkTaskSetId(rootPool, 2) + checkTaskSetId(rootPool, 2) + checkTaskSetId(rootPool, 4) + + taskSetManager12.taskFinished() + assert(rootPool.getSchedulableByName("1").runningTasks === 3) + taskSetManager24.abort() + assert(rootPool.getSchedulableByName("2").runningTasks === 2) + } + + test("Nested Pool Test") { + sc = new SparkContext("local", "ClusterSchedulerSuite") + val clusterScheduler = new ClusterScheduler(sc) + var tasks = ArrayBuffer[Task[_]]() + val task = new DummyTask(0) + tasks += task + val taskSet = new TaskSet(tasks.toArray,0,0,0,null) + + val rootPool = new Pool("", SchedulingMode.FAIR, 0, 0) + val pool0 = new Pool("0", SchedulingMode.FAIR, 3, 1) + val pool1 = new Pool("1", SchedulingMode.FAIR, 4, 1) + rootPool.addSchedulable(pool0) + rootPool.addSchedulable(pool1) + + val pool00 = new Pool("00", SchedulingMode.FAIR, 2, 2) + val pool01 = new Pool("01", SchedulingMode.FAIR, 1, 1) + pool0.addSchedulable(pool00) + pool0.addSchedulable(pool01) + + val pool10 = new Pool("10", SchedulingMode.FAIR, 2, 2) + val pool11 = new Pool("11", SchedulingMode.FAIR, 2, 1) + pool1.addSchedulable(pool10) + pool1.addSchedulable(pool11) + + val taskSetManager000 = createDummyTaskSetManager(0, 0, 5, clusterScheduler, taskSet) + val taskSetManager001 = createDummyTaskSetManager(0, 1, 5, clusterScheduler, taskSet) + pool00.addSchedulable(taskSetManager000) + pool00.addSchedulable(taskSetManager001) + + val taskSetManager010 = createDummyTaskSetManager(1, 2, 5, clusterScheduler, taskSet) + val taskSetManager011 = createDummyTaskSetManager(1, 3, 5, clusterScheduler, taskSet) + pool01.addSchedulable(taskSetManager010) + pool01.addSchedulable(taskSetManager011) + + val taskSetManager100 = createDummyTaskSetManager(2, 4, 5, clusterScheduler, taskSet) + val taskSetManager101 = createDummyTaskSetManager(2, 5, 5, clusterScheduler, taskSet) + pool10.addSchedulable(taskSetManager100) + pool10.addSchedulable(taskSetManager101) + + val taskSetManager110 = createDummyTaskSetManager(3, 6, 5, clusterScheduler, taskSet) + val taskSetManager111 = createDummyTaskSetManager(3, 7, 5, clusterScheduler, taskSet) + pool11.addSchedulable(taskSetManager110) + pool11.addSchedulable(taskSetManager111) + + checkTaskSetId(rootPool, 0) + checkTaskSetId(rootPool, 4) + checkTaskSetId(rootPool, 6) + checkTaskSetId(rootPool, 2) + } +} diff --git a/core/src/test/scala/spark/scheduler/DAGSchedulerSuite.scala b/core/src/test/scala/spark/scheduler/DAGSchedulerSuite.scala index c0f8986de8..a8b88d7936 100644 --- a/core/src/test/scala/spark/scheduler/DAGSchedulerSuite.scala +++ b/core/src/test/scala/spark/scheduler/DAGSchedulerSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import scala.collection.mutable.{Map, HashMap} @@ -6,21 +23,14 @@ import org.scalatest.FunSuite import org.scalatest.BeforeAndAfter import spark.LocalSparkContext - -import spark.storage.BlockManager -import spark.storage.BlockManagerId -import spark.storage.BlockManagerMaster -import spark.{Dependency, ShuffleDependency, OneToOneDependency} -import spark.FetchFailedException import spark.MapOutputTracker import spark.RDD import spark.SparkContext -import spark.SparkException import spark.Partition import spark.TaskContext -import spark.TaskEndReason - -import spark.{FetchFailed, Success} +import spark.{Dependency, ShuffleDependency, OneToOneDependency} +import spark.{FetchFailed, Success, TaskEndReason} +import spark.storage.{BlockManagerId, BlockManagerMaster} /** * Tests for DAGScheduler. These tests directly call the event processing functions in DAGScheduler @@ -44,7 +54,7 @@ class DAGSchedulerSuite extends FunSuite with BeforeAndAfter with LocalSparkCont override def submitTasks(taskSet: TaskSet) = { // normally done by TaskSetManager taskSet.tasks.foreach(_.generation = mapOutputTracker.getGeneration) - taskSets += taskSet + taskSets += taskSet } override def setListener(listener: TaskSchedulerListener) = {} override def defaultParallelism() = 2 @@ -164,7 +174,7 @@ class DAGSchedulerSuite extends FunSuite with BeforeAndAfter with LocalSparkCont } } } - + /** Sends the rdd to the scheduler for scheduling. */ private def submit( rdd: RDD[_], @@ -174,7 +184,7 @@ class DAGSchedulerSuite extends FunSuite with BeforeAndAfter with LocalSparkCont listener: JobListener = listener) { runEvent(JobSubmitted(rdd, func, partitions, allowLocal, null, listener)) } - + /** Sends TaskSetFailed to the scheduler. */ private def failed(taskSet: TaskSet, message: String) { runEvent(TaskSetFailed(taskSet, message)) @@ -209,11 +219,11 @@ class DAGSchedulerSuite extends FunSuite with BeforeAndAfter with LocalSparkCont runEvent(JobSubmitted(rdd, jobComputeFunc, Array(0), true, null, listener)) assert(results === Map(0 -> 42)) } - + test("run trivial job w/ dependency") { val baseRdd = makeRdd(1, Nil) val finalRdd = makeRdd(1, List(new OneToOneDependency(baseRdd))) - submit(finalRdd, Array(0)) + submit(finalRdd, Array(0)) complete(taskSets(0), Seq((Success, 42))) assert(results === Map(0 -> 42)) } @@ -250,7 +260,7 @@ class DAGSchedulerSuite extends FunSuite with BeforeAndAfter with LocalSparkCont complete(taskSets(1), Seq((Success, 42))) assert(results === Map(0 -> 42)) } - + test("run trivial shuffle with fetch failure") { val shuffleMapRdd = makeRdd(2, Nil) val shuffleDep = new ShuffleDependency(shuffleMapRdd, null) @@ -385,12 +395,12 @@ class DAGSchedulerSuite extends FunSuite with BeforeAndAfter with LocalSparkCont assert(results === Map(0 -> 42)) } - /** Assert that the supplied TaskSet has exactly the given preferredLocations. */ + /** Assert that the supplied TaskSet has exactly the given preferredLocations. Note, converts taskSet's locations to host only. */ private def assertLocations(taskSet: TaskSet, locations: Seq[Seq[String]]) { assert(locations.size === taskSet.tasks.size) for ((expectLocs, taskLocs) <- taskSet.tasks.map(_.preferredLocations).zip(locations)) { - assert(expectLocs === taskLocs) + assert(expectLocs.map(loc => spark.Utils.parseHostPort(loc)._1) === taskLocs) } } @@ -398,6 +408,6 @@ class DAGSchedulerSuite extends FunSuite with BeforeAndAfter with LocalSparkCont new MapStatus(makeBlockManagerId(host), Array.fill[Byte](reduces)(2)) private def makeBlockManagerId(host: String): BlockManagerId = - BlockManagerId("exec-" + host, host, 12345) + BlockManagerId("exec-" + host, host, 12345, 0) } diff --git a/core/src/test/scala/spark/scheduler/JobLoggerSuite.scala b/core/src/test/scala/spark/scheduler/JobLoggerSuite.scala new file mode 100644 index 0000000000..0f855c38da --- /dev/null +++ b/core/src/test/scala/spark/scheduler/JobLoggerSuite.scala @@ -0,0 +1,121 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler + +import java.util.Properties +import java.util.concurrent.LinkedBlockingQueue +import org.scalatest.FunSuite +import org.scalatest.matchers.ShouldMatchers +import scala.collection.mutable +import spark._ +import spark.SparkContext._ + + +class JobLoggerSuite extends FunSuite with LocalSparkContext with ShouldMatchers { + + test("inner method") { + sc = new SparkContext("local", "joblogger") + val joblogger = new JobLogger { + def createLogWriterTest(jobID: Int) = createLogWriter(jobID) + def closeLogWriterTest(jobID: Int) = closeLogWriter(jobID) + def getRddNameTest(rdd: RDD[_]) = getRddName(rdd) + def buildJobDepTest(jobID: Int, stage: Stage) = buildJobDep(jobID, stage) + } + type MyRDD = RDD[(Int, Int)] + def makeRdd( + numPartitions: Int, + dependencies: List[Dependency[_]] + ): MyRDD = { + val maxPartition = numPartitions - 1 + return new MyRDD(sc, dependencies) { + override def compute(split: Partition, context: TaskContext): Iterator[(Int, Int)] = + throw new RuntimeException("should not be reached") + override def getPartitions = (0 to maxPartition).map(i => new Partition { + override def index = i + }).toArray + } + } + val jobID = 5 + val parentRdd = makeRdd(4, Nil) + val shuffleDep = new ShuffleDependency(parentRdd, null) + val rootRdd = makeRdd(4, List(shuffleDep)) + val shuffleMapStage = new Stage(1, parentRdd, Some(shuffleDep), Nil, jobID, None) + val rootStage = new Stage(0, rootRdd, None, List(shuffleMapStage), jobID, None) + + joblogger.onStageSubmitted(SparkListenerStageSubmitted(rootStage, 4)) + joblogger.getRddNameTest(parentRdd) should be (parentRdd.getClass.getName) + parentRdd.setName("MyRDD") + joblogger.getRddNameTest(parentRdd) should be ("MyRDD") + joblogger.createLogWriterTest(jobID) + joblogger.getJobIDtoPrintWriter.size should be (1) + joblogger.buildJobDepTest(jobID, rootStage) + joblogger.getJobIDToStages.get(jobID).get.size should be (2) + joblogger.getStageIDToJobID.get(0) should be (Some(jobID)) + joblogger.getStageIDToJobID.get(1) should be (Some(jobID)) + joblogger.closeLogWriterTest(jobID) + joblogger.getStageIDToJobID.size should be (0) + joblogger.getJobIDToStages.size should be (0) + joblogger.getJobIDtoPrintWriter.size should be (0) + } + + test("inner variables") { + sc = new SparkContext("local[4]", "joblogger") + val joblogger = new JobLogger { + override protected def closeLogWriter(jobID: Int) = + getJobIDtoPrintWriter.get(jobID).foreach { fileWriter => + fileWriter.close() + } + } + sc.addSparkListener(joblogger) + val rdd = sc.parallelize(1 to 1e2.toInt, 4).map{ i => (i % 12, 2 * i) } + rdd.reduceByKey(_+_).collect() + + joblogger.getLogDir should be ("/tmp/spark") + joblogger.getJobIDtoPrintWriter.size should be (1) + joblogger.getStageIDToJobID.size should be (2) + joblogger.getStageIDToJobID.get(0) should be (Some(0)) + joblogger.getStageIDToJobID.get(1) should be (Some(0)) + joblogger.getJobIDToStages.size should be (1) + } + + + test("interface functions") { + sc = new SparkContext("local[4]", "joblogger") + val joblogger = new JobLogger { + var onTaskEndCount = 0 + var onJobEndCount = 0 + var onJobStartCount = 0 + var onStageCompletedCount = 0 + var onStageSubmittedCount = 0 + override def onTaskEnd(taskEnd: SparkListenerTaskEnd) = onTaskEndCount += 1 + override def onJobEnd(jobEnd: SparkListenerJobEnd) = onJobEndCount += 1 + override def onJobStart(jobStart: SparkListenerJobStart) = onJobStartCount += 1 + override def onStageCompleted(stageCompleted: StageCompleted) = onStageCompletedCount += 1 + override def onStageSubmitted(stageSubmitted: SparkListenerStageSubmitted) = onStageSubmittedCount += 1 + } + sc.addSparkListener(joblogger) + val rdd = sc.parallelize(1 to 1e2.toInt, 4).map{ i => (i % 12, 2 * i) } + rdd.reduceByKey(_+_).collect() + + joblogger.onJobStartCount should be (1) + joblogger.onJobEndCount should be (1) + joblogger.onTaskEndCount should be (8) + joblogger.onStageSubmittedCount should be (2) + joblogger.onStageCompletedCount should be (2) + } +} diff --git a/core/src/test/scala/spark/scheduler/LocalSchedulerSuite.scala b/core/src/test/scala/spark/scheduler/LocalSchedulerSuite.scala new file mode 100644 index 0000000000..14bb58731b --- /dev/null +++ b/core/src/test/scala/spark/scheduler/LocalSchedulerSuite.scala @@ -0,0 +1,223 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.scheduler + +import org.scalatest.FunSuite +import org.scalatest.BeforeAndAfter + +import spark._ +import spark.scheduler._ +import spark.scheduler.cluster._ +import scala.collection.mutable.ArrayBuffer +import scala.collection.mutable.{ConcurrentMap, HashMap} +import java.util.concurrent.Semaphore +import java.util.concurrent.CountDownLatch +import java.util.Properties + +class Lock() { + var finished = false + def jobWait() = { + synchronized { + while(!finished) { + this.wait() + } + } + } + + def jobFinished() = { + synchronized { + finished = true + this.notifyAll() + } + } +} + +object TaskThreadInfo { + val threadToLock = HashMap[Int, Lock]() + val threadToRunning = HashMap[Int, Boolean]() + val threadToStarted = HashMap[Int, CountDownLatch]() +} + +/* + * 1. each thread contains one job. + * 2. each job contains one stage. + * 3. each stage only contains one task. + * 4. each task(launched) must be lanched orderly(using threadToStarted) to make sure + * it will get cpu core resource, and will wait to finished after user manually + * release "Lock" and then cluster will contain another free cpu cores. + * 5. each task(pending) must use "sleep" to make sure it has been added to taskSetManager queue, + * thus it will be scheduled later when cluster has free cpu cores. + */ +class LocalSchedulerSuite extends FunSuite with LocalSparkContext { + + def createThread(threadIndex: Int, poolName: String, sc: SparkContext, sem: Semaphore) { + + TaskThreadInfo.threadToRunning(threadIndex) = false + val nums = sc.parallelize(threadIndex to threadIndex, 1) + TaskThreadInfo.threadToLock(threadIndex) = new Lock() + TaskThreadInfo.threadToStarted(threadIndex) = new CountDownLatch(1) + new Thread { + if (poolName != null) { + sc.addLocalProperties("spark.scheduler.cluster.fair.pool",poolName) + } + override def run() { + val ans = nums.map(number => { + TaskThreadInfo.threadToRunning(number) = true + TaskThreadInfo.threadToStarted(number).countDown() + TaskThreadInfo.threadToLock(number).jobWait() + TaskThreadInfo.threadToRunning(number) = false + number + }).collect() + assert(ans.toList === List(threadIndex)) + sem.release() + } + }.start() + } + + test("Local FIFO scheduler end-to-end test") { + System.setProperty("spark.cluster.schedulingmode", "FIFO") + sc = new SparkContext("local[4]", "test") + val sem = new Semaphore(0) + + createThread(1,null,sc,sem) + TaskThreadInfo.threadToStarted(1).await() + createThread(2,null,sc,sem) + TaskThreadInfo.threadToStarted(2).await() + createThread(3,null,sc,sem) + TaskThreadInfo.threadToStarted(3).await() + createThread(4,null,sc,sem) + TaskThreadInfo.threadToStarted(4).await() + // thread 5 and 6 (stage pending)must meet following two points + // 1. stages (taskSetManager) of jobs in thread 5 and 6 should be add to taskSetManager + // queue before executing TaskThreadInfo.threadToLock(1).jobFinished() + // 2. priority of stage in thread 5 should be prior to priority of stage in thread 6 + // So I just use "sleep" 1s here for each thread. + // TODO: any better solution? + createThread(5,null,sc,sem) + Thread.sleep(1000) + createThread(6,null,sc,sem) + Thread.sleep(1000) + + assert(TaskThreadInfo.threadToRunning(1) === true) + assert(TaskThreadInfo.threadToRunning(2) === true) + assert(TaskThreadInfo.threadToRunning(3) === true) + assert(TaskThreadInfo.threadToRunning(4) === true) + assert(TaskThreadInfo.threadToRunning(5) === false) + assert(TaskThreadInfo.threadToRunning(6) === false) + + TaskThreadInfo.threadToLock(1).jobFinished() + TaskThreadInfo.threadToStarted(5).await() + + assert(TaskThreadInfo.threadToRunning(1) === false) + assert(TaskThreadInfo.threadToRunning(2) === true) + assert(TaskThreadInfo.threadToRunning(3) === true) + assert(TaskThreadInfo.threadToRunning(4) === true) + assert(TaskThreadInfo.threadToRunning(5) === true) + assert(TaskThreadInfo.threadToRunning(6) === false) + + TaskThreadInfo.threadToLock(3).jobFinished() + TaskThreadInfo.threadToStarted(6).await() + + assert(TaskThreadInfo.threadToRunning(1) === false) + assert(TaskThreadInfo.threadToRunning(2) === true) + assert(TaskThreadInfo.threadToRunning(3) === false) + assert(TaskThreadInfo.threadToRunning(4) === true) + assert(TaskThreadInfo.threadToRunning(5) === true) + assert(TaskThreadInfo.threadToRunning(6) === true) + + TaskThreadInfo.threadToLock(2).jobFinished() + TaskThreadInfo.threadToLock(4).jobFinished() + TaskThreadInfo.threadToLock(5).jobFinished() + TaskThreadInfo.threadToLock(6).jobFinished() + sem.acquire(6) + } + + test("Local fair scheduler end-to-end test") { + sc = new SparkContext("local[8]", "LocalSchedulerSuite") + val sem = new Semaphore(0) + System.setProperty("spark.cluster.schedulingmode", "FAIR") + val xmlPath = getClass.getClassLoader.getResource("fairscheduler.xml").getFile() + System.setProperty("spark.fairscheduler.allocation.file", xmlPath) + + createThread(10,"1",sc,sem) + TaskThreadInfo.threadToStarted(10).await() + createThread(20,"2",sc,sem) + TaskThreadInfo.threadToStarted(20).await() + createThread(30,"3",sc,sem) + TaskThreadInfo.threadToStarted(30).await() + + assert(TaskThreadInfo.threadToRunning(10) === true) + assert(TaskThreadInfo.threadToRunning(20) === true) + assert(TaskThreadInfo.threadToRunning(30) === true) + + createThread(11,"1",sc,sem) + TaskThreadInfo.threadToStarted(11).await() + createThread(21,"2",sc,sem) + TaskThreadInfo.threadToStarted(21).await() + createThread(31,"3",sc,sem) + TaskThreadInfo.threadToStarted(31).await() + + assert(TaskThreadInfo.threadToRunning(11) === true) + assert(TaskThreadInfo.threadToRunning(21) === true) + assert(TaskThreadInfo.threadToRunning(31) === true) + + createThread(12,"1",sc,sem) + TaskThreadInfo.threadToStarted(12).await() + createThread(22,"2",sc,sem) + TaskThreadInfo.threadToStarted(22).await() + createThread(32,"3",sc,sem) + + assert(TaskThreadInfo.threadToRunning(12) === true) + assert(TaskThreadInfo.threadToRunning(22) === true) + assert(TaskThreadInfo.threadToRunning(32) === false) + + TaskThreadInfo.threadToLock(10).jobFinished() + TaskThreadInfo.threadToStarted(32).await() + + assert(TaskThreadInfo.threadToRunning(32) === true) + + //1. Similar with above scenario, sleep 1s for stage of 23 and 33 to be added to taskSetManager + // queue so that cluster will assign free cpu core to stage 23 after stage 11 finished. + //2. priority of 23 and 33 will be meaningless as using fair scheduler here. + createThread(23,"2",sc,sem) + createThread(33,"3",sc,sem) + Thread.sleep(1000) + + TaskThreadInfo.threadToLock(11).jobFinished() + TaskThreadInfo.threadToStarted(23).await() + + assert(TaskThreadInfo.threadToRunning(23) === true) + assert(TaskThreadInfo.threadToRunning(33) === false) + + TaskThreadInfo.threadToLock(12).jobFinished() + TaskThreadInfo.threadToStarted(33).await() + + assert(TaskThreadInfo.threadToRunning(33) === true) + + TaskThreadInfo.threadToLock(20).jobFinished() + TaskThreadInfo.threadToLock(21).jobFinished() + TaskThreadInfo.threadToLock(22).jobFinished() + TaskThreadInfo.threadToLock(23).jobFinished() + TaskThreadInfo.threadToLock(30).jobFinished() + TaskThreadInfo.threadToLock(31).jobFinished() + TaskThreadInfo.threadToLock(32).jobFinished() + TaskThreadInfo.threadToLock(33).jobFinished() + + sem.acquire(11) + } +} diff --git a/core/src/test/scala/spark/scheduler/SparkListenerSuite.scala b/core/src/test/scala/spark/scheduler/SparkListenerSuite.scala index 42a87d8b90..392d67d67b 100644 --- a/core/src/test/scala/spark/scheduler/SparkListenerSuite.scala +++ b/core/src/test/scala/spark/scheduler/SparkListenerSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import org.scalatest.FunSuite @@ -77,7 +94,7 @@ class SparkListenerSuite extends FunSuite with LocalSparkContext with ShouldMatc class SaveStageInfo extends SparkListener { val stageInfos = mutable.Buffer[StageInfo]() - def onStageCompleted(stage: StageCompleted) { + override def onStageCompleted(stage: StageCompleted) { stageInfos += stage.stageInfo } } diff --git a/core/src/test/scala/spark/scheduler/TaskContextSuite.scala b/core/src/test/scala/spark/scheduler/TaskContextSuite.scala index 647bcaf860..95a6eee2fc 100644 --- a/core/src/test/scala/spark/scheduler/TaskContextSuite.scala +++ b/core/src/test/scala/spark/scheduler/TaskContextSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.scheduler import org.scalatest.FunSuite diff --git a/core/src/test/scala/spark/storage/BlockManagerSuite.scala b/core/src/test/scala/spark/storage/BlockManagerSuite.scala index 5a11a4483b..b719d65342 100644 --- a/core/src/test/scala/spark/storage/BlockManagerSuite.scala +++ b/core/src/test/scala/spark/storage/BlockManagerSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.storage import java.nio.ByteBuffer @@ -15,10 +32,10 @@ import org.scalatest.time.SpanSugar._ import spark.JavaSerializer import spark.KryoSerializer import spark.SizeEstimator -import spark.Utils import spark.util.AkkaUtils import spark.util.ByteBufferInputStream + class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodTester { var store: BlockManager = null var store2: BlockManager = null @@ -47,6 +64,8 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT oldHeartBeat = System.setProperty("spark.storage.disableBlockManagerHeartBeat", "true") val initialize = PrivateMethod[Unit]('initialize) SizeEstimator invokePrivate initialize() + // Set some value ... + System.setProperty("spark.hostPort", spark.Utils.localHostName() + ":" + 1111) } after { @@ -97,9 +116,9 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT } test("BlockManagerId object caching") { - val id1 = BlockManagerId("e1", "XXX", 1) - val id2 = BlockManagerId("e1", "XXX", 1) // this should return the same object as id1 - val id3 = BlockManagerId("e1", "XXX", 2) // this should return a different object + val id1 = BlockManagerId("e1", "XXX", 1, 0) + val id2 = BlockManagerId("e1", "XXX", 1, 0) // this should return the same object as id1 + val id3 = BlockManagerId("e1", "XXX", 2, 0) // this should return a different object assert(id2 === id1, "id2 is not same as id1") assert(id2.eq(id1), "id2 is not the same object as id1") assert(id3 != id1, "id3 is same as id1") @@ -122,7 +141,7 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT // Putting a1, a2 and a3 in memory and telling master only about a1 and a2 store.putSingle("a1", a1, StorageLevel.MEMORY_ONLY) store.putSingle("a2", a2, StorageLevel.MEMORY_ONLY) - store.putSingle("a3", a3, StorageLevel.MEMORY_ONLY, false) + store.putSingle("a3", a3, StorageLevel.MEMORY_ONLY, tellMaster = false) // Checking whether blocks are in memory assert(store.getSingle("a1") != None, "a1 was not in store") @@ -168,7 +187,7 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT // Putting a1, a2 and a3 in memory and telling master only about a1 and a2 store.putSingle("a1-to-remove", a1, StorageLevel.MEMORY_ONLY) store.putSingle("a2-to-remove", a2, StorageLevel.MEMORY_ONLY) - store.putSingle("a3-to-remove", a3, StorageLevel.MEMORY_ONLY, false) + store.putSingle("a3-to-remove", a3, StorageLevel.MEMORY_ONLY, tellMaster = false) // Checking whether blocks are in memory and memory size val memStatus = master.getMemoryStatus.head._2 @@ -207,6 +226,39 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT } } + test("removing rdd") { + store = new BlockManager("<driver>", actorSystem, master, serializer, 2000) + val a1 = new Array[Byte](400) + val a2 = new Array[Byte](400) + val a3 = new Array[Byte](400) + // Putting a1, a2 and a3 in memory. + store.putSingle("rdd_0_0", a1, StorageLevel.MEMORY_ONLY) + store.putSingle("rdd_0_1", a2, StorageLevel.MEMORY_ONLY) + store.putSingle("nonrddblock", a3, StorageLevel.MEMORY_ONLY) + master.removeRdd(0, blocking = false) + + eventually(timeout(1000 milliseconds), interval(10 milliseconds)) { + store.getSingle("rdd_0_0") should be (None) + master.getLocations("rdd_0_0") should have size 0 + } + eventually(timeout(1000 milliseconds), interval(10 milliseconds)) { + store.getSingle("rdd_0_1") should be (None) + master.getLocations("rdd_0_1") should have size 0 + } + eventually(timeout(1000 milliseconds), interval(10 milliseconds)) { + store.getSingle("nonrddblock") should not be (None) + master.getLocations("nonrddblock") should have size (1) + } + + store.putSingle("rdd_0_0", a1, StorageLevel.MEMORY_ONLY) + store.putSingle("rdd_0_1", a2, StorageLevel.MEMORY_ONLY) + master.removeRdd(0, blocking = true) + store.getSingle("rdd_0_0") should be (None) + master.getLocations("rdd_0_0") should have size 0 + store.getSingle("rdd_0_1") should be (None) + master.getLocations("rdd_0_1") should have size 0 + } + test("reregistration on heart beat") { val heartBeat = PrivateMethod[Unit]('heartBeat) store = new BlockManager("<driver>", actorSystem, master, serializer, 2000) @@ -235,7 +287,7 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT master.removeExecutor(store.blockManagerId.executorId) assert(master.getLocations("a1").size == 0, "a1 was not removed from master") - store.putSingle("a2", a1, StorageLevel.MEMORY_ONLY) + store.putSingle("a2", a2, StorageLevel.MEMORY_ONLY) store.waitForAsyncReregister() assert(master.getLocations("a1").size > 0, "a1 was not reregistered with master") @@ -253,7 +305,7 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT master.removeExecutor(store.blockManagerId.executorId) val t1 = new Thread { override def run() { - store.put("a2", a2.iterator, StorageLevel.MEMORY_ONLY, true) + store.put("a2", a2.iterator, StorageLevel.MEMORY_ONLY, tellMaster = true) } } val t2 = new Thread { @@ -463,9 +515,9 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT val list1 = List(new Array[Byte](200), new Array[Byte](200)) val list2 = List(new Array[Byte](200), new Array[Byte](200)) val list3 = List(new Array[Byte](200), new Array[Byte](200)) - store.put("list1", list1.iterator, StorageLevel.MEMORY_ONLY, true) - store.put("list2", list2.iterator, StorageLevel.MEMORY_ONLY, true) - store.put("list3", list3.iterator, StorageLevel.MEMORY_ONLY, true) + store.put("list1", list1.iterator, StorageLevel.MEMORY_ONLY, tellMaster = true) + store.put("list2", list2.iterator, StorageLevel.MEMORY_ONLY, tellMaster = true) + store.put("list3", list3.iterator, StorageLevel.MEMORY_ONLY, tellMaster = true) assert(store.get("list2") != None, "list2 was not in store") assert(store.get("list2").get.size == 2) assert(store.get("list3") != None, "list3 was not in store") @@ -474,7 +526,7 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT assert(store.get("list2") != None, "list2 was not in store") assert(store.get("list2").get.size == 2) // At this point list2 was gotten last, so LRU will getSingle rid of list3 - store.put("list1", list1.iterator, StorageLevel.MEMORY_ONLY, true) + store.put("list1", list1.iterator, StorageLevel.MEMORY_ONLY, tellMaster = true) assert(store.get("list1") != None, "list1 was not in store") assert(store.get("list1").get.size == 2) assert(store.get("list2") != None, "list2 was not in store") @@ -489,9 +541,9 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT val list3 = List(new Array[Byte](200), new Array[Byte](200)) val list4 = List(new Array[Byte](200), new Array[Byte](200)) // First store list1 and list2, both in memory, and list3, on disk only - store.put("list1", list1.iterator, StorageLevel.MEMORY_ONLY_SER, true) - store.put("list2", list2.iterator, StorageLevel.MEMORY_ONLY_SER, true) - store.put("list3", list3.iterator, StorageLevel.DISK_ONLY, true) + store.put("list1", list1.iterator, StorageLevel.MEMORY_ONLY_SER, tellMaster = true) + store.put("list2", list2.iterator, StorageLevel.MEMORY_ONLY_SER, tellMaster = true) + store.put("list3", list3.iterator, StorageLevel.DISK_ONLY, tellMaster = true) // At this point LRU should not kick in because list3 is only on disk assert(store.get("list1") != None, "list2 was not in store") assert(store.get("list1").get.size === 2) @@ -506,7 +558,7 @@ class BlockManagerSuite extends FunSuite with BeforeAndAfter with PrivateMethodT assert(store.get("list3") != None, "list1 was not in store") assert(store.get("list3").get.size === 2) // Now let's add in list4, which uses both disk and memory; list1 should drop out - store.put("list4", list4.iterator, StorageLevel.MEMORY_AND_DISK_SER, true) + store.put("list4", list4.iterator, StorageLevel.MEMORY_AND_DISK_SER, tellMaster = true) assert(store.get("list1") === None, "list1 was in store") assert(store.get("list2") != None, "list3 was not in store") assert(store.get("list2").get.size === 2) diff --git a/core/src/test/scala/spark/ui/UISuite.scala b/core/src/test/scala/spark/ui/UISuite.scala new file mode 100644 index 0000000000..56c1fed6ad --- /dev/null +++ b/core/src/test/scala/spark/ui/UISuite.scala @@ -0,0 +1,46 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package spark.ui + +import scala.util.{Failure, Success, Try} +import java.net.ServerSocket +import org.scalatest.FunSuite +import org.eclipse.jetty.server.Server + +class UISuite extends FunSuite { + test("jetty port increases under contention") { + val startPort = 33333 + val server = new Server(startPort) + server.start() + val (jettyServer1, boundPort1) = JettyUtils.startJettyServer("localhost", startPort, Seq()) + val (jettyServer2, boundPort2) = JettyUtils.startJettyServer("localhost", startPort, Seq()) + + assert(boundPort1 === startPort + 1) + assert(boundPort2 === startPort + 2) + } + + test("jetty binds to port 0 correctly") { + val (jettyServer, boundPort) = JettyUtils.startJettyServer("localhost", 0, Seq()) + assert(jettyServer.getState === "STARTED") + assert(boundPort != 0) + Try {new ServerSocket(boundPort)} match { + case Success(s) => fail("Port %s doesn't seem used by jetty server".format(boundPort)) + case Failure (e) => + } + } +} diff --git a/core/src/test/scala/spark/util/DistributionSuite.scala b/core/src/test/scala/spark/util/DistributionSuite.scala index cc6249b1dd..6578b55e82 100644 --- a/core/src/test/scala/spark/util/DistributionSuite.scala +++ b/core/src/test/scala/spark/util/DistributionSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import org.scalatest.FunSuite diff --git a/core/src/test/scala/spark/util/NextIteratorSuite.scala b/core/src/test/scala/spark/util/NextIteratorSuite.scala index ed5b36da73..fdbd43d941 100644 --- a/core/src/test/scala/spark/util/NextIteratorSuite.scala +++ b/core/src/test/scala/spark/util/NextIteratorSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import org.scalatest.FunSuite diff --git a/core/src/test/scala/spark/util/RateLimitedOutputStreamSuite.scala b/core/src/test/scala/spark/util/RateLimitedOutputStreamSuite.scala index 794063fb6d..4c0044202f 100644 --- a/core/src/test/scala/spark/util/RateLimitedOutputStreamSuite.scala +++ b/core/src/test/scala/spark/util/RateLimitedOutputStreamSuite.scala @@ -1,3 +1,20 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + package spark.util import org.scalatest.FunSuite |
