1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
|
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES 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.spark.deploy.yarn
import java.net.URI
import java.nio.ByteBuffer
import java.security.PrivilegedExceptionAction
import scala.collection.JavaConversions._
import scala.collection.mutable.HashMap
import org.apache.hadoop.conf.Configuration
import org.apache.hadoop.fs.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._
import org.apache.hadoop.yarn.api.ApplicationConstants.Environment
import org.apache.hadoop.yarn.api.records._
import org.apache.hadoop.yarn.api.protocolrecords._
import org.apache.hadoop.yarn.util.{Apps, ConverterUtils, Records}
import org.apache.spark.{SparkConf, Logging}
import org.apache.hadoop.yarn.conf.YarnConfiguration
trait WorkerRunnableUtil extends Logging {
val yarnConf: YarnConfiguration
val sparkConf: SparkConf
lazy val env = prepareEnvironment
def prepareCommand(
masterAddress: String,
slaveId: String,
hostname: String,
workerMemory: Int,
workerCores: Int) = {
// Extra options for the JVM
var JAVA_OPTS = ""
// Set the JVM memory
val workerMemoryString = workerMemory + "m"
JAVA_OPTS += "-Xms" + workerMemoryString + " -Xmx" + workerMemoryString + " "
if (env.isDefinedAt("SPARK_JAVA_OPTS")) {
JAVA_OPTS += env("SPARK_JAVA_OPTS") + " "
}
JAVA_OPTS += " -Djava.io.tmpdir=" +
new Path(Environment.PWD.$(), YarnConfiguration.DEFAULT_CONTAINER_TEMP_DIR) + " "
// Commenting it out for now - so that people can refer to the properties if required. Remove
// it once cpuset version is pushed out.
// The context is, default gc for server class machines end up using all cores to do gc - hence
// if there are multiple containers in same node, spark gc effects all other containers
// performance (which can also be other spark containers)
// Instead of using this, rely on cpusets by YARN to enforce spark behaves 'properly' in
// multi-tenant environments. Not sure how default java gc behaves if it is limited to subset
// of cores on a node.
/*
else {
// If no java_opts specified, default to using -XX:+CMSIncrementalMode
// It might be possible that other modes/config is being done in SPARK_JAVA_OPTS, so we dont
// want to mess with it.
// In our expts, using (default) throughput collector has severe perf ramnifications in
// multi-tennent machines
// The options are based on
// http://www.oracle.com/technetwork/java/gc-tuning-5-138395.html#0.0.0.%20When%20to%20Use%20the%20Concurrent%20Low%20Pause%20Collector|outline
JAVA_OPTS += " -XX:+UseConcMarkSweepGC "
JAVA_OPTS += " -XX:+CMSIncrementalMode "
JAVA_OPTS += " -XX:+CMSIncrementalPacing "
JAVA_OPTS += " -XX:CMSIncrementalDutyCycleMin=0 "
JAVA_OPTS += " -XX:CMSIncrementalDutyCycle=10 "
}
*/
var javaCommand = "java"
val javaHome = System.getenv("JAVA_HOME")
if ((javaHome != null && !javaHome.isEmpty()) || env.isDefinedAt("JAVA_HOME")) {
javaCommand = Environment.JAVA_HOME.$() + "/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.
// TODO: If the OOM is not recoverable by rescheduling it on different node, then do
// 'something' to fail job ... akin to blacklisting trackers in mapred ?
" -XX:OnOutOfMemoryError='kill %p' " +
JAVA_OPTS +
" org.apache.spark.executor.CoarseGrainedExecutorBackend " +
masterAddress + " " +
slaveId + " " +
hostname + " " +
workerCores +
" 1> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stdout" +
" 2> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stderr")
commands
}
private def setupDistributedCache(
file: String,
rtype: LocalResourceType,
localResources: HashMap[String, LocalResource],
timestamp: String,
size: String,
vis: String) = {
val uri = new URI(file)
val amJarRsrc = Records.newRecord(classOf[LocalResource]).asInstanceOf[LocalResource]
amJarRsrc.setType(rtype)
amJarRsrc.setVisibility(LocalResourceVisibility.valueOf(vis))
amJarRsrc.setResource(ConverterUtils.getYarnUrlFromURI(uri))
amJarRsrc.setTimestamp(timestamp.toLong)
amJarRsrc.setSize(size.toLong)
localResources(uri.getFragment()) = amJarRsrc
}
def prepareLocalResources: HashMap[String, LocalResource] = {
logInfo("Preparing Local resources")
val localResources = HashMap[String, LocalResource]()
if (System.getenv("SPARK_YARN_CACHE_FILES") != null) {
val timeStamps = System.getenv("SPARK_YARN_CACHE_FILES_TIME_STAMPS").split(',')
val fileSizes = System.getenv("SPARK_YARN_CACHE_FILES_FILE_SIZES").split(',')
val distFiles = System.getenv("SPARK_YARN_CACHE_FILES").split(',')
val visibilities = System.getenv("SPARK_YARN_CACHE_FILES_VISIBILITIES").split(',')
for( i <- 0 to distFiles.length - 1) {
setupDistributedCache(distFiles(i), LocalResourceType.FILE, localResources, timeStamps(i),
fileSizes(i), visibilities(i))
}
}
if (System.getenv("SPARK_YARN_CACHE_ARCHIVES") != null) {
val timeStamps = System.getenv("SPARK_YARN_CACHE_ARCHIVES_TIME_STAMPS").split(',')
val fileSizes = System.getenv("SPARK_YARN_CACHE_ARCHIVES_FILE_SIZES").split(',')
val distArchives = System.getenv("SPARK_YARN_CACHE_ARCHIVES").split(',')
val visibilities = System.getenv("SPARK_YARN_CACHE_ARCHIVES_VISIBILITIES").split(',')
for( i <- 0 to distArchives.length - 1) {
setupDistributedCache(distArchives(i), LocalResourceType.ARCHIVE, localResources,
timeStamps(i), fileSizes(i), visibilities(i))
}
}
logInfo("Prepared Local resources " + localResources)
localResources
}
def prepareEnvironment: HashMap[String, String] = {
val env = new HashMap[String, String]()
ClientBase.populateClasspath(yarnConf, sparkConf, System.getenv("SPARK_YARN_LOG4J_PATH") != null, env)
// Allow users to specify some environment variables
Apps.setEnvFromInputString(env, System.getenv("SPARK_YARN_USER_ENV"))
System.getenv().filterKeys(_.startsWith("SPARK")).foreach { case (k,v) => env(k) = v }
env
}
}
|
