From dfbdc9ddb773e2b1149e6a6c661f14b631b692d0 Mon Sep 17 00:00:00 2001 From: shane-huang Date: Mon, 23 Sep 2013 11:28:58 +0800 Subject: added spark-class and spark-executor to sbin Signed-off-by: shane-huang --- docs/running-on-yarn.md | 4 ++-- docs/spark-standalone.md | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'docs') diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index c611db0af4..767eb5cdac 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -42,7 +42,7 @@ This would be used to connect to the cluster, write to the dfs and submit jobs t The command to launch the YARN Client is as follows: - SPARK_JAR= ./spark-class org.apache.spark.deploy.yarn.Client \ + SPARK_JAR= ./sbin/spark-class org.apache.spark.deploy.yarn.Client \ --jar \ --class \ --args \ @@ -62,7 +62,7 @@ For example: # Submit Spark's ApplicationMaster to YARN's ResourceManager, and instruct Spark to run the SparkPi example $ SPARK_JAR=./assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop2.0.5-alpha.jar \ - ./spark-class org.apache.spark.deploy.yarn.Client \ + ./sbin/spark-class org.apache.spark.deploy.yarn.Client \ --jar examples/target/scala-{{site.SCALA_VERSION}}/spark-examples-assembly-{{site.SPARK_VERSION}}.jar \ --class org.apache.spark.examples.SparkPi \ --args yarn-standalone \ diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index 81cdbefd0c..9d4ad1ec8d 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -25,7 +25,7 @@ the master's web UI, which is [http://localhost:8080](http://localhost:8080) by Similarly, you can start one or more workers and connect them to the master via: - ./spark-class org.apache.spark.deploy.worker.Worker spark://IP:PORT + ./sbin/spark-class org.apache.spark.deploy.worker.Worker spark://IP:PORT Once you have started a worker, look at the master's web UI ([http://localhost:8080](http://localhost:8080) by default). You should see the new node listed there, along with its number of CPUs and memory (minus one gigabyte left for the OS). -- cgit v1.2.3 From fcfe4f920484b64b01e4e22219d59c78ffd17054 Mon Sep 17 00:00:00 2001 From: shane-huang Date: Mon, 23 Sep 2013 12:42:34 +0800 Subject: add admin scripts to sbin Signed-off-by: shane-huang --- bin/compute-classpath.cmd | 69 ------------------- bin/compute-classpath.sh | 61 ----------------- bin/slaves.sh | 74 -------------------- bin/spark-config.sh | 36 ---------- bin/spark-daemon.sh | 164 --------------------------------------------- bin/spark-daemons.sh | 35 ---------- bin/start-all.sh | 34 ---------- bin/start-master.sh | 52 -------------- bin/start-slave.sh | 35 ---------- bin/start-slaves.sh | 48 ------------- bin/stop-all.sh | 32 --------- bin/stop-master.sh | 27 -------- bin/stop-slaves.sh | 37 ---------- docs/spark-standalone.md | 12 ++-- sbin/compute-classpath.cmd | 69 +++++++++++++++++++ sbin/compute-classpath.sh | 61 +++++++++++++++++ sbin/slaves.sh | 74 ++++++++++++++++++++ sbin/spark-config.sh | 36 ++++++++++ sbin/spark-daemon.sh | 164 +++++++++++++++++++++++++++++++++++++++++++++ sbin/spark-daemons.sh | 35 ++++++++++ sbin/start-all.sh | 34 ++++++++++ sbin/start-master.sh | 52 ++++++++++++++ sbin/start-slave.sh | 35 ++++++++++ sbin/start-slaves.sh | 48 +++++++++++++ sbin/stop-all.sh | 32 +++++++++ sbin/stop-master.sh | 27 ++++++++ sbin/stop-slaves.sh | 37 ++++++++++ 27 files changed, 710 insertions(+), 710 deletions(-) delete mode 100644 bin/compute-classpath.cmd delete mode 100755 bin/compute-classpath.sh delete mode 100755 bin/slaves.sh delete mode 100755 bin/spark-config.sh delete mode 100755 bin/spark-daemon.sh delete mode 100755 bin/spark-daemons.sh delete mode 100755 bin/start-all.sh delete mode 100755 bin/start-master.sh delete mode 100755 bin/start-slave.sh delete mode 100755 bin/start-slaves.sh delete mode 100755 bin/stop-all.sh delete mode 100755 bin/stop-master.sh delete mode 100755 bin/stop-slaves.sh create mode 100644 sbin/compute-classpath.cmd create mode 100755 sbin/compute-classpath.sh create mode 100755 sbin/slaves.sh create mode 100755 sbin/spark-config.sh create mode 100755 sbin/spark-daemon.sh create mode 100755 sbin/spark-daemons.sh create mode 100755 sbin/start-all.sh create mode 100755 sbin/start-master.sh create mode 100755 sbin/start-slave.sh create mode 100755 sbin/start-slaves.sh create mode 100755 sbin/stop-all.sh create mode 100755 sbin/stop-master.sh create mode 100755 sbin/stop-slaves.sh (limited to 'docs') diff --git a/bin/compute-classpath.cmd b/bin/compute-classpath.cmd deleted file mode 100644 index cf38188c4b..0000000000 --- a/bin/compute-classpath.cmd +++ /dev/null @@ -1,69 +0,0 @@ -@echo off - -rem -rem Licensed to the Apache Software Foundation (ASF) under one or more -rem contributor license agreements. See the NOTICE file distributed with -rem this work for additional information regarding copyright ownership. -rem The ASF licenses this file to You under the Apache License, Version 2.0 -rem (the "License"); you may not use this file except in compliance with -rem the License. You may obtain a copy of the License at -rem -rem http://www.apache.org/licenses/LICENSE-2.0 -rem -rem Unless required by applicable law or agreed to in writing, software -rem distributed under the License is distributed on an "AS IS" BASIS, -rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -rem See the License for the specific language governing permissions and -rem limitations under the License. -rem - -rem This script computes Spark's classpath and prints it to stdout; it's used by both the "run" -rem script and the ExecutorRunner in standalone cluster mode. - -set SCALA_VERSION=2.9.3 - -rem Figure out where the Spark framework is installed -set FWDIR=%~dp0..\ - -rem Load environment variables from conf\spark-env.cmd, if it exists -if exist "%FWDIR%conf\spark-env.cmd" call "%FWDIR%conf\spark-env.cmd" - -rem Build up classpath -set CLASSPATH=%SPARK_CLASSPATH%;%FWDIR%conf -if exist "%FWDIR%RELEASE" ( - for %%d in ("%FWDIR%jars\spark-assembly*.jar") do ( - set ASSEMBLY_JAR=%%d - ) -) else ( - for %%d in ("%FWDIR%assembly\target\scala-%SCALA_VERSION%\spark-assembly*hadoop*.jar") do ( - set ASSEMBLY_JAR=%%d - ) -) -set CLASSPATH=%CLASSPATH%;%ASSEMBLY_JAR% - -if "x%SPARK_TESTING%"=="x1" ( - rem Add test clases to path - set CLASSPATH=%CLASSPATH%;%FWDIR%core\target\scala-%SCALA_VERSION%\test-classes - set CLASSPATH=%CLASSPATH%;%FWDIR%repl\target\scala-%SCALA_VERSION%\test-classes - set CLASSPATH=%CLASSPATH%;%FWDIR%mllib\target\scala-%SCALA_VERSION%\test-classes - set CLASSPATH=%CLASSPATH%;%FWDIR%bagel\target\scala-%SCALA_VERSION%\test-classes - set CLASSPATH=%CLASSPATH%;%FWDIR%streaming\target\scala-%SCALA_VERSION%\test-classes -) - -rem Add hadoop conf dir - else FileSystem.*, etc fail -rem Note, this assumes that there is either a HADOOP_CONF_DIR or YARN_CONF_DIR which hosts -rem the configurtion files. -if "x%HADOOP_CONF_DIR%"=="x" goto no_hadoop_conf_dir - set CLASSPATH=%CLASSPATH%;%HADOOP_CONF_DIR% -:no_hadoop_conf_dir - -if "x%YARN_CONF_DIR%"=="x" goto no_yarn_conf_dir - set CLASSPATH=%CLASSPATH%;%YARN_CONF_DIR% -:no_yarn_conf_dir - -rem A bit of a hack to allow calling this script within run2.cmd without seeing output -if "%DONT_PRINT_CLASSPATH%"=="1" goto exit - -echo %CLASSPATH% - -:exit diff --git a/bin/compute-classpath.sh b/bin/compute-classpath.sh deleted file mode 100755 index c7819d4932..0000000000 --- a/bin/compute-classpath.sh +++ /dev/null @@ -1,61 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# This script computes Spark's classpath and prints it to stdout; it's used by both the "run" -# script and the ExecutorRunner in standalone cluster mode. - -SCALA_VERSION=2.9.3 - -# Figure out where Spark is installed -FWDIR="$(cd `dirname $0`/..; pwd)" - -# Load environment variables from conf/spark-env.sh, if it exists -if [ -e $FWDIR/conf/spark-env.sh ] ; then - . $FWDIR/conf/spark-env.sh -fi - -# Build up classpath -CLASSPATH="$SPARK_CLASSPATH:$FWDIR/conf" -if [ -f "$FWDIR/RELEASE" ]; then - ASSEMBLY_JAR=`ls "$FWDIR"/jars/spark-assembly*.jar` -else - ASSEMBLY_JAR=`ls "$FWDIR"/assembly/target/scala-$SCALA_VERSION/spark-assembly*hadoop*.jar` -fi -CLASSPATH="$CLASSPATH:$ASSEMBLY_JAR" - -# Add test classes if we're running from SBT or Maven with SPARK_TESTING set to 1 -if [[ $SPARK_TESTING == 1 ]]; then - CLASSPATH="$CLASSPATH:$FWDIR/core/target/scala-$SCALA_VERSION/test-classes" - CLASSPATH="$CLASSPATH:$FWDIR/repl/target/scala-$SCALA_VERSION/test-classes" - CLASSPATH="$CLASSPATH:$FWDIR/mllib/target/scala-$SCALA_VERSION/test-classes" - CLASSPATH="$CLASSPATH:$FWDIR/bagel/target/scala-$SCALA_VERSION/test-classes" - CLASSPATH="$CLASSPATH:$FWDIR/streaming/target/scala-$SCALA_VERSION/test-classes" -fi - -# Add hadoop conf dir if given -- otherwise FileSystem.*, etc fail ! -# Note, this assumes that there is either a HADOOP_CONF_DIR or YARN_CONF_DIR which hosts -# the configurtion files. -if [ "x" != "x$HADOOP_CONF_DIR" ]; then - CLASSPATH="$CLASSPATH:$HADOOP_CONF_DIR" -fi -if [ "x" != "x$YARN_CONF_DIR" ]; then - CLASSPATH="$CLASSPATH:$YARN_CONF_DIR" -fi - -echo "$CLASSPATH" diff --git a/bin/slaves.sh b/bin/slaves.sh deleted file mode 100755 index 752565b759..0000000000 --- a/bin/slaves.sh +++ /dev/null @@ -1,74 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Run a shell command on all slave hosts. -# -# Environment Variables -# -# SPARK_SLAVES File naming remote hosts. -# Default is ${SPARK_CONF_DIR}/slaves. -# SPARK_CONF_DIR Alternate conf dir. Default is ${SPARK_HOME}/conf. -# SPARK_SLAVE_SLEEP Seconds to sleep between spawning remote commands. -# SPARK_SSH_OPTS Options passed to ssh when running remote commands. -## - -usage="Usage: slaves.sh [--config confdir] command..." - -# if no args specified, show usage -if [ $# -le 0 ]; then - echo $usage - exit 1 -fi - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -. "$bin/spark-config.sh" - -# If the slaves file is specified in the command line, -# then it takes precedence over the definition in -# spark-env.sh. Save it here. -HOSTLIST=$SPARK_SLAVES - -if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then - . "${SPARK_CONF_DIR}/spark-env.sh" -fi - -if [ "$HOSTLIST" = "" ]; then - if [ "$SPARK_SLAVES" = "" ]; then - export HOSTLIST="${SPARK_CONF_DIR}/slaves" - else - export HOSTLIST="${SPARK_SLAVES}" - fi -fi - -# By default disable strict host key checking -if [ "$SPARK_SSH_OPTS" = "" ]; then - SPARK_SSH_OPTS="-o StrictHostKeyChecking=no" -fi - -for slave in `cat "$HOSTLIST"|sed "s/#.*$//;/^$/d"`; do - ssh $SPARK_SSH_OPTS $slave $"${@// /\\ }" \ - 2>&1 | sed "s/^/$slave: /" & - if [ "$SPARK_SLAVE_SLEEP" != "" ]; then - sleep $SPARK_SLAVE_SLEEP - fi -done - -wait diff --git a/bin/spark-config.sh b/bin/spark-config.sh deleted file mode 100755 index cd2c7b7b0d..0000000000 --- a/bin/spark-config.sh +++ /dev/null @@ -1,36 +0,0 @@ -# -# 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. -# - -# included in all the spark scripts with source command -# should not be executable directly -# also should not be passed any arguments, since we need original $* - -# resolve links - $0 may be a softlink -this="${BASH_SOURCE-$0}" -common_bin=$(cd -P -- "$(dirname -- "$this")" && pwd -P) -script="$(basename -- "$this")" -this="$common_bin/$script" - -# convert relative path to absolute path -config_bin=`dirname "$this"` -script=`basename "$this"` -config_bin=`cd "$config_bin"; pwd` -this="$config_bin/$script" - -export SPARK_PREFIX=`dirname "$this"`/.. -export SPARK_HOME=${SPARK_PREFIX} -export SPARK_CONF_DIR="$SPARK_HOME/conf" diff --git a/bin/spark-daemon.sh b/bin/spark-daemon.sh deleted file mode 100755 index 6baee0c33b..0000000000 --- a/bin/spark-daemon.sh +++ /dev/null @@ -1,164 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Runs a Spark command as a daemon. -# -# Environment Variables -# -# SPARK_CONF_DIR Alternate conf dir. Default is ${SPARK_PREFIX}/conf. -# SPARK_LOG_DIR Where log files are stored. PWD by default. -# SPARK_MASTER host:path where spark code should be rsync'd from -# SPARK_PID_DIR The pid files are stored. /tmp by default. -# SPARK_IDENT_STRING A string representing this instance of spark. $USER by default -# SPARK_NICENESS The scheduling priority for daemons. Defaults to 0. -## - -usage="Usage: spark-daemon.sh [--config ] [--hosts hostlistfile] (start|stop) " - -# if no args specified, show usage -if [ $# -le 1 ]; then - echo $usage - exit 1 -fi - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -. "$bin/spark-config.sh" - -# get arguments -startStop=$1 -shift -command=$1 -shift -instance=$1 -shift - -spark_rotate_log () -{ - log=$1; - num=5; - if [ -n "$2" ]; then - num=$2 - fi - if [ -f "$log" ]; then # rotate logs - while [ $num -gt 1 ]; do - prev=`expr $num - 1` - [ -f "$log.$prev" ] && mv "$log.$prev" "$log.$num" - num=$prev - done - mv "$log" "$log.$num"; - fi -} - -if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then - . "${SPARK_CONF_DIR}/spark-env.sh" -fi - -if [ "$SPARK_IDENT_STRING" = "" ]; then - export SPARK_IDENT_STRING="$USER" -fi - - -export SPARK_PRINT_LAUNCH_COMMAND="1" - -# get log directory -if [ "$SPARK_LOG_DIR" = "" ]; then - export SPARK_LOG_DIR="$SPARK_HOME/logs" -fi -mkdir -p "$SPARK_LOG_DIR" -touch $SPARK_LOG_DIR/.spark_test > /dev/null 2>&1 -TEST_LOG_DIR=$? -if [ "${TEST_LOG_DIR}" = "0" ]; then - rm -f $SPARK_LOG_DIR/.spark_test -else - chown $SPARK_IDENT_STRING $SPARK_LOG_DIR -fi - -if [ "$SPARK_PID_DIR" = "" ]; then - SPARK_PID_DIR=/tmp -fi - -# some variables -export SPARK_LOGFILE=spark-$SPARK_IDENT_STRING-$command-$instance-$HOSTNAME.log -export SPARK_ROOT_LOGGER="INFO,DRFA" -log=$SPARK_LOG_DIR/spark-$SPARK_IDENT_STRING-$command-$instance-$HOSTNAME.out -pid=$SPARK_PID_DIR/spark-$SPARK_IDENT_STRING-$command-$instance.pid - -# Set default scheduling priority -if [ "$SPARK_NICENESS" = "" ]; then - export SPARK_NICENESS=0 -fi - - -case $startStop in - - (start) - - mkdir -p "$SPARK_PID_DIR" - - if [ -f $pid ]; then - if kill -0 `cat $pid` > /dev/null 2>&1; then - echo $command running as process `cat $pid`. Stop it first. - exit 1 - fi - fi - - if [ "$SPARK_MASTER" != "" ]; then - echo rsync from $SPARK_MASTER - rsync -a -e ssh --delete --exclude=.svn --exclude='logs/*' --exclude='contrib/hod/logs/*' $SPARK_MASTER/ "$SPARK_HOME" - fi - - spark_rotate_log "$log" - echo starting $command, logging to $log - cd "$SPARK_PREFIX" - nohup nice -n $SPARK_NICENESS "$SPARK_PREFIX"/sbin/spark-class $command "$@" >> "$log" 2>&1 < /dev/null & - newpid=$! - echo $newpid > $pid - sleep 2 - # Check if the process has died; in that case we'll tail the log so the user can see - if ! kill -0 $newpid >/dev/null 2>&1; then - echo "failed to launch $command:" - tail -2 "$log" | sed 's/^/ /' - echo "full log in $log" - fi - ;; - - (stop) - - if [ -f $pid ]; then - if kill -0 `cat $pid` > /dev/null 2>&1; then - echo stopping $command - kill `cat $pid` - else - echo no $command to stop - fi - else - echo no $command to stop - fi - ;; - - (*) - echo $usage - exit 1 - ;; - -esac - - diff --git a/bin/spark-daemons.sh b/bin/spark-daemons.sh deleted file mode 100755 index 354eb905a1..0000000000 --- a/bin/spark-daemons.sh +++ /dev/null @@ -1,35 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Run a Spark command on all slave hosts. - -usage="Usage: spark-daemons.sh [--config confdir] [--hosts hostlistfile] [start|stop] command instance-number args..." - -# if no args specified, show usage -if [ $# -le 1 ]; then - echo $usage - exit 1 -fi - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -. "$bin/spark-config.sh" - -exec "$bin/slaves.sh" cd "$SPARK_HOME" \; "$bin/spark-daemon.sh" "$@" diff --git a/bin/start-all.sh b/bin/start-all.sh deleted file mode 100755 index 0182f1ab24..0000000000 --- a/bin/start-all.sh +++ /dev/null @@ -1,34 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Start all spark daemons. -# Starts the master on this node. -# Starts a worker on each node specified in conf/slaves - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -# Load the Spark configuration -. "$bin/spark-config.sh" - -# Start Master -"$bin"/start-master.sh - -# Start Workers -"$bin"/start-slaves.sh diff --git a/bin/start-master.sh b/bin/start-master.sh deleted file mode 100755 index 648c7ae75f..0000000000 --- a/bin/start-master.sh +++ /dev/null @@ -1,52 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Starts the master on the machine this script is executed on. - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -. "$bin/spark-config.sh" - -if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then - . "${SPARK_CONF_DIR}/spark-env.sh" -fi - -if [ "$SPARK_MASTER_PORT" = "" ]; then - SPARK_MASTER_PORT=7077 -fi - -if [ "$SPARK_MASTER_IP" = "" ]; then - SPARK_MASTER_IP=`hostname` -fi - -if [ "$SPARK_MASTER_WEBUI_PORT" = "" ]; then - SPARK_MASTER_WEBUI_PORT=8080 -fi - -# Set SPARK_PUBLIC_DNS so the master report the correct webUI address to the slaves -if [ "$SPARK_PUBLIC_DNS" = "" ]; then - # If we appear to be running on EC2, use the public address by default: - # NOTE: ec2-metadata is installed on Amazon Linux AMI. Check based on that and hostname - if command -v ec2-metadata > /dev/null || [[ `hostname` == *ec2.internal ]]; then - export SPARK_PUBLIC_DNS=`wget -q -O - http://instance-data.ec2.internal/latest/meta-data/public-hostname` - fi -fi - -"$bin"/spark-daemon.sh start org.apache.spark.deploy.master.Master 1 --ip $SPARK_MASTER_IP --port $SPARK_MASTER_PORT --webui-port $SPARK_MASTER_WEBUI_PORT diff --git a/bin/start-slave.sh b/bin/start-slave.sh deleted file mode 100755 index 4eefa20944..0000000000 --- a/bin/start-slave.sh +++ /dev/null @@ -1,35 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Usage: start-slave.sh -# where is like "spark://localhost:7077" - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -# Set SPARK_PUBLIC_DNS so slaves can be linked in master web UI -if [ "$SPARK_PUBLIC_DNS" = "" ]; then - # If we appear to be running on EC2, use the public address by default: - # NOTE: ec2-metadata is installed on Amazon Linux AMI. Check based on that and hostname - if command -v ec2-metadata > /dev/null || [[ `hostname` == *ec2.internal ]]; then - export SPARK_PUBLIC_DNS=`wget -q -O - http://instance-data.ec2.internal/latest/meta-data/public-hostname` - fi -fi - -"$bin"/spark-daemon.sh start org.apache.spark.deploy.worker.Worker "$@" diff --git a/bin/start-slaves.sh b/bin/start-slaves.sh deleted file mode 100755 index 00dc4888b2..0000000000 --- a/bin/start-slaves.sh +++ /dev/null @@ -1,48 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -. "$bin/spark-config.sh" - -if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then - . "${SPARK_CONF_DIR}/spark-env.sh" -fi - -# Find the port number for the master -if [ "$SPARK_MASTER_PORT" = "" ]; then - SPARK_MASTER_PORT=7077 -fi - -if [ "$SPARK_MASTER_IP" = "" ]; then - SPARK_MASTER_IP=`hostname` -fi - -# Launch the slaves -if [ "$SPARK_WORKER_INSTANCES" = "" ]; then - exec "$bin/slaves.sh" cd "$SPARK_HOME" \; "$bin/start-slave.sh" 1 spark://$SPARK_MASTER_IP:$SPARK_MASTER_PORT -else - if [ "$SPARK_WORKER_WEBUI_PORT" = "" ]; then - SPARK_WORKER_WEBUI_PORT=8081 - fi - for ((i=0; i<$SPARK_WORKER_INSTANCES; i++)); do - "$bin/slaves.sh" cd "$SPARK_HOME" \; "$bin/start-slave.sh" $(( $i + 1 )) spark://$SPARK_MASTER_IP:$SPARK_MASTER_PORT --webui-port $(( $SPARK_WORKER_WEBUI_PORT + $i )) - done -fi diff --git a/bin/stop-all.sh b/bin/stop-all.sh deleted file mode 100755 index b6c83a7ba4..0000000000 --- a/bin/stop-all.sh +++ /dev/null @@ -1,32 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Start all spark daemons. -# Run this on the master nde - - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -# Load the Spark configuration -. "$bin/spark-config.sh" - -# Stop the slaves, then the master -"$bin"/stop-slaves.sh -"$bin"/stop-master.sh diff --git a/bin/stop-master.sh b/bin/stop-master.sh deleted file mode 100755 index 310e33bedc..0000000000 --- a/bin/stop-master.sh +++ /dev/null @@ -1,27 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Starts the master on the machine this script is executed on. - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -. "$bin/spark-config.sh" - -"$bin"/spark-daemon.sh stop org.apache.spark.deploy.master.Master 1 diff --git a/bin/stop-slaves.sh b/bin/stop-slaves.sh deleted file mode 100755 index 03e416a132..0000000000 --- a/bin/stop-slaves.sh +++ /dev/null @@ -1,37 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -# Starts the master on the machine this script is executed on. - -bin=`dirname "$0"` -bin=`cd "$bin"; pwd` - -. "$bin/spark-config.sh" - -if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then - . "${SPARK_CONF_DIR}/spark-env.sh" -fi - -if [ "$SPARK_WORKER_INSTANCES" = "" ]; then - "$bin"/spark-daemons.sh stop org.apache.spark.deploy.worker.Worker 1 -else - for ((i=0; i<$SPARK_WORKER_INSTANCES; i++)); do - "$bin"/spark-daemons.sh stop org.apache.spark.deploy.worker.Worker $(( $i + 1 )) - done -fi diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index 9d4ad1ec8d..b3f9160673 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -67,12 +67,12 @@ To launch a Spark standalone cluster with the launch scripts, you need to create Once you've set up this file, you can launch or stop your cluster with the following shell scripts, based on Hadoop's deploy scripts, and available in `SPARK_HOME/bin`: -- `bin/start-master.sh` - Starts a master instance on the machine the script is executed on. -- `bin/start-slaves.sh` - Starts a slave instance on each machine specified in the `conf/slaves` file. -- `bin/start-all.sh` - Starts both a master and a number of slaves as described above. -- `bin/stop-master.sh` - Stops the master that was started via the `bin/start-master.sh` script. -- `bin/stop-slaves.sh` - Stops the slave instances that were started via `bin/start-slaves.sh`. -- `bin/stop-all.sh` - Stops both the master and the slaves as described above. +- `sbin/start-master.sh` - Starts a master instance on the machine the script is executed on. +- `sbin/start-slaves.sh` - Starts a slave instance on each machine specified in the `conf/slaves` file. +- `sbin/start-all.sh` - Starts both a master and a number of slaves as described above. +- `sbin/stop-master.sh` - Stops the master that was started via the `bin/start-master.sh` script. +- `sbin/stop-slaves.sh` - Stops the slave instances that were started via `bin/start-slaves.sh`. +- `sbin/stop-all.sh` - Stops both the master and the slaves as described above. Note that these scripts must be executed on the machine you want to run the Spark master on, not your local machine. diff --git a/sbin/compute-classpath.cmd b/sbin/compute-classpath.cmd new file mode 100644 index 0000000000..cf38188c4b --- /dev/null +++ b/sbin/compute-classpath.cmd @@ -0,0 +1,69 @@ +@echo off + +rem +rem Licensed to the Apache Software Foundation (ASF) under one or more +rem contributor license agreements. See the NOTICE file distributed with +rem this work for additional information regarding copyright ownership. +rem The ASF licenses this file to You under the Apache License, Version 2.0 +rem (the "License"); you may not use this file except in compliance with +rem the License. You may obtain a copy of the License at +rem +rem http://www.apache.org/licenses/LICENSE-2.0 +rem +rem Unless required by applicable law or agreed to in writing, software +rem distributed under the License is distributed on an "AS IS" BASIS, +rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +rem See the License for the specific language governing permissions and +rem limitations under the License. +rem + +rem This script computes Spark's classpath and prints it to stdout; it's used by both the "run" +rem script and the ExecutorRunner in standalone cluster mode. + +set SCALA_VERSION=2.9.3 + +rem Figure out where the Spark framework is installed +set FWDIR=%~dp0..\ + +rem Load environment variables from conf\spark-env.cmd, if it exists +if exist "%FWDIR%conf\spark-env.cmd" call "%FWDIR%conf\spark-env.cmd" + +rem Build up classpath +set CLASSPATH=%SPARK_CLASSPATH%;%FWDIR%conf +if exist "%FWDIR%RELEASE" ( + for %%d in ("%FWDIR%jars\spark-assembly*.jar") do ( + set ASSEMBLY_JAR=%%d + ) +) else ( + for %%d in ("%FWDIR%assembly\target\scala-%SCALA_VERSION%\spark-assembly*hadoop*.jar") do ( + set ASSEMBLY_JAR=%%d + ) +) +set CLASSPATH=%CLASSPATH%;%ASSEMBLY_JAR% + +if "x%SPARK_TESTING%"=="x1" ( + rem Add test clases to path + set CLASSPATH=%CLASSPATH%;%FWDIR%core\target\scala-%SCALA_VERSION%\test-classes + set CLASSPATH=%CLASSPATH%;%FWDIR%repl\target\scala-%SCALA_VERSION%\test-classes + set CLASSPATH=%CLASSPATH%;%FWDIR%mllib\target\scala-%SCALA_VERSION%\test-classes + set CLASSPATH=%CLASSPATH%;%FWDIR%bagel\target\scala-%SCALA_VERSION%\test-classes + set CLASSPATH=%CLASSPATH%;%FWDIR%streaming\target\scala-%SCALA_VERSION%\test-classes +) + +rem Add hadoop conf dir - else FileSystem.*, etc fail +rem Note, this assumes that there is either a HADOOP_CONF_DIR or YARN_CONF_DIR which hosts +rem the configurtion files. +if "x%HADOOP_CONF_DIR%"=="x" goto no_hadoop_conf_dir + set CLASSPATH=%CLASSPATH%;%HADOOP_CONF_DIR% +:no_hadoop_conf_dir + +if "x%YARN_CONF_DIR%"=="x" goto no_yarn_conf_dir + set CLASSPATH=%CLASSPATH%;%YARN_CONF_DIR% +:no_yarn_conf_dir + +rem A bit of a hack to allow calling this script within run2.cmd without seeing output +if "%DONT_PRINT_CLASSPATH%"=="1" goto exit + +echo %CLASSPATH% + +:exit diff --git a/sbin/compute-classpath.sh b/sbin/compute-classpath.sh new file mode 100755 index 0000000000..c7819d4932 --- /dev/null +++ b/sbin/compute-classpath.sh @@ -0,0 +1,61 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# This script computes Spark's classpath and prints it to stdout; it's used by both the "run" +# script and the ExecutorRunner in standalone cluster mode. + +SCALA_VERSION=2.9.3 + +# Figure out where Spark is installed +FWDIR="$(cd `dirname $0`/..; pwd)" + +# Load environment variables from conf/spark-env.sh, if it exists +if [ -e $FWDIR/conf/spark-env.sh ] ; then + . $FWDIR/conf/spark-env.sh +fi + +# Build up classpath +CLASSPATH="$SPARK_CLASSPATH:$FWDIR/conf" +if [ -f "$FWDIR/RELEASE" ]; then + ASSEMBLY_JAR=`ls "$FWDIR"/jars/spark-assembly*.jar` +else + ASSEMBLY_JAR=`ls "$FWDIR"/assembly/target/scala-$SCALA_VERSION/spark-assembly*hadoop*.jar` +fi +CLASSPATH="$CLASSPATH:$ASSEMBLY_JAR" + +# Add test classes if we're running from SBT or Maven with SPARK_TESTING set to 1 +if [[ $SPARK_TESTING == 1 ]]; then + CLASSPATH="$CLASSPATH:$FWDIR/core/target/scala-$SCALA_VERSION/test-classes" + CLASSPATH="$CLASSPATH:$FWDIR/repl/target/scala-$SCALA_VERSION/test-classes" + CLASSPATH="$CLASSPATH:$FWDIR/mllib/target/scala-$SCALA_VERSION/test-classes" + CLASSPATH="$CLASSPATH:$FWDIR/bagel/target/scala-$SCALA_VERSION/test-classes" + CLASSPATH="$CLASSPATH:$FWDIR/streaming/target/scala-$SCALA_VERSION/test-classes" +fi + +# Add hadoop conf dir if given -- otherwise FileSystem.*, etc fail ! +# Note, this assumes that there is either a HADOOP_CONF_DIR or YARN_CONF_DIR which hosts +# the configurtion files. +if [ "x" != "x$HADOOP_CONF_DIR" ]; then + CLASSPATH="$CLASSPATH:$HADOOP_CONF_DIR" +fi +if [ "x" != "x$YARN_CONF_DIR" ]; then + CLASSPATH="$CLASSPATH:$YARN_CONF_DIR" +fi + +echo "$CLASSPATH" diff --git a/sbin/slaves.sh b/sbin/slaves.sh new file mode 100755 index 0000000000..68408bcad8 --- /dev/null +++ b/sbin/slaves.sh @@ -0,0 +1,74 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Run a shell command on all slave hosts. +# +# Environment Variables +# +# SPARK_SLAVES File naming remote hosts. +# Default is ${SPARK_CONF_DIR}/slaves. +# SPARK_CONF_DIR Alternate conf dir. Default is ${SPARK_HOME}/conf. +# SPARK_SLAVE_SLEEP Seconds to sleep between spawning remote commands. +# SPARK_SSH_OPTS Options passed to ssh when running remote commands. +## + +usage="Usage: slaves.sh [--config confdir] command..." + +# if no args specified, show usage +if [ $# -le 0 ]; then + echo $usage + exit 1 +fi + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +. "$sbin/spark-config.sh" + +# If the slaves file is specified in the command line, +# then it takes precedence over the definition in +# spark-env.sh. Save it here. +HOSTLIST=$SPARK_SLAVES + +if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then + . "${SPARK_CONF_DIR}/spark-env.sh" +fi + +if [ "$HOSTLIST" = "" ]; then + if [ "$SPARK_SLAVES" = "" ]; then + export HOSTLIST="${SPARK_CONF_DIR}/slaves" + else + export HOSTLIST="${SPARK_SLAVES}" + fi +fi + +# By default disable strict host key checking +if [ "$SPARK_SSH_OPTS" = "" ]; then + SPARK_SSH_OPTS="-o StrictHostKeyChecking=no" +fi + +for slave in `cat "$HOSTLIST"|sed "s/#.*$//;/^$/d"`; do + ssh $SPARK_SSH_OPTS $slave $"${@// /\\ }" \ + 2>&1 | sed "s/^/$slave: /" & + if [ "$SPARK_SLAVE_SLEEP" != "" ]; then + sleep $SPARK_SLAVE_SLEEP + fi +done + +wait diff --git a/sbin/spark-config.sh b/sbin/spark-config.sh new file mode 100755 index 0000000000..cd2c7b7b0d --- /dev/null +++ b/sbin/spark-config.sh @@ -0,0 +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. +# + +# included in all the spark scripts with source command +# should not be executable directly +# also should not be passed any arguments, since we need original $* + +# resolve links - $0 may be a softlink +this="${BASH_SOURCE-$0}" +common_bin=$(cd -P -- "$(dirname -- "$this")" && pwd -P) +script="$(basename -- "$this")" +this="$common_bin/$script" + +# convert relative path to absolute path +config_bin=`dirname "$this"` +script=`basename "$this"` +config_bin=`cd "$config_bin"; pwd` +this="$config_bin/$script" + +export SPARK_PREFIX=`dirname "$this"`/.. +export SPARK_HOME=${SPARK_PREFIX} +export SPARK_CONF_DIR="$SPARK_HOME/conf" diff --git a/sbin/spark-daemon.sh b/sbin/spark-daemon.sh new file mode 100755 index 0000000000..ae82349cc6 --- /dev/null +++ b/sbin/spark-daemon.sh @@ -0,0 +1,164 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Runs a Spark command as a daemon. +# +# Environment Variables +# +# SPARK_CONF_DIR Alternate conf dir. Default is ${SPARK_PREFIX}/conf. +# SPARK_LOG_DIR Where log files are stored. PWD by default. +# SPARK_MASTER host:path where spark code should be rsync'd from +# SPARK_PID_DIR The pid files are stored. /tmp by default. +# SPARK_IDENT_STRING A string representing this instance of spark. $USER by default +# SPARK_NICENESS The scheduling priority for daemons. Defaults to 0. +## + +usage="Usage: spark-daemon.sh [--config ] [--hosts hostlistfile] (start|stop) " + +# if no args specified, show usage +if [ $# -le 1 ]; then + echo $usage + exit 1 +fi + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +. "$sbin/spark-config.sh" + +# get arguments +startStop=$1 +shift +command=$1 +shift +instance=$1 +shift + +spark_rotate_log () +{ + log=$1; + num=5; + if [ -n "$2" ]; then + num=$2 + fi + if [ -f "$log" ]; then # rotate logs + while [ $num -gt 1 ]; do + prev=`expr $num - 1` + [ -f "$log.$prev" ] && mv "$log.$prev" "$log.$num" + num=$prev + done + mv "$log" "$log.$num"; + fi +} + +if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then + . "${SPARK_CONF_DIR}/spark-env.sh" +fi + +if [ "$SPARK_IDENT_STRING" = "" ]; then + export SPARK_IDENT_STRING="$USER" +fi + + +export SPARK_PRINT_LAUNCH_COMMAND="1" + +# get log directory +if [ "$SPARK_LOG_DIR" = "" ]; then + export SPARK_LOG_DIR="$SPARK_HOME/logs" +fi +mkdir -p "$SPARK_LOG_DIR" +touch $SPARK_LOG_DIR/.spark_test > /dev/null 2>&1 +TEST_LOG_DIR=$? +if [ "${TEST_LOG_DIR}" = "0" ]; then + rm -f $SPARK_LOG_DIR/.spark_test +else + chown $SPARK_IDENT_STRING $SPARK_LOG_DIR +fi + +if [ "$SPARK_PID_DIR" = "" ]; then + SPARK_PID_DIR=/tmp +fi + +# some variables +export SPARK_LOGFILE=spark-$SPARK_IDENT_STRING-$command-$instance-$HOSTNAME.log +export SPARK_ROOT_LOGGER="INFO,DRFA" +log=$SPARK_LOG_DIR/spark-$SPARK_IDENT_STRING-$command-$instance-$HOSTNAME.out +pid=$SPARK_PID_DIR/spark-$SPARK_IDENT_STRING-$command-$instance.pid + +# Set default scheduling priority +if [ "$SPARK_NICENESS" = "" ]; then + export SPARK_NICENESS=0 +fi + + +case $startStop in + + (start) + + mkdir -p "$SPARK_PID_DIR" + + if [ -f $pid ]; then + if kill -0 `cat $pid` > /dev/null 2>&1; then + echo $command running as process `cat $pid`. Stop it first. + exit 1 + fi + fi + + if [ "$SPARK_MASTER" != "" ]; then + echo rsync from $SPARK_MASTER + rsync -a -e ssh --delete --exclude=.svn --exclude='logs/*' --exclude='contrib/hod/logs/*' $SPARK_MASTER/ "$SPARK_HOME" + fi + + spark_rotate_log "$log" + echo starting $command, logging to $log + cd "$SPARK_PREFIX" + nohup nice -n $SPARK_NICENESS "$SPARK_PREFIX"/sbin/spark-class $command "$@" >> "$log" 2>&1 < /dev/null & + newpid=$! + echo $newpid > $pid + sleep 2 + # Check if the process has died; in that case we'll tail the log so the user can see + if ! kill -0 $newpid >/dev/null 2>&1; then + echo "failed to launch $command:" + tail -2 "$log" | sed 's/^/ /' + echo "full log in $log" + fi + ;; + + (stop) + + if [ -f $pid ]; then + if kill -0 `cat $pid` > /dev/null 2>&1; then + echo stopping $command + kill `cat $pid` + else + echo no $command to stop + fi + else + echo no $command to stop + fi + ;; + + (*) + echo $usage + exit 1 + ;; + +esac + + diff --git a/sbin/spark-daemons.sh b/sbin/spark-daemons.sh new file mode 100755 index 0000000000..d91254b690 --- /dev/null +++ b/sbin/spark-daemons.sh @@ -0,0 +1,35 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Run a Spark command on all slave hosts. + +usage="Usage: spark-daemons.sh [--config confdir] [--hosts hostlistfile] [start|stop] command instance-number args..." + +# if no args specified, show usage +if [ $# -le 1 ]; then + echo $usage + exit 1 +fi + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +. "$sbin/spark-config.sh" + +exec "$sbin/slaves.sh" cd "$SPARK_HOME" \; "$sbin/spark-daemon.sh" "$@" diff --git a/sbin/start-all.sh b/sbin/start-all.sh new file mode 100755 index 0000000000..2daf49db35 --- /dev/null +++ b/sbin/start-all.sh @@ -0,0 +1,34 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Start all spark daemons. +# Starts the master on this node. +# Starts a worker on each node specified in conf/slaves + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +# Load the Spark configuration +. "$sbin/spark-config.sh" + +# Start Master +"$sbin"/start-master.sh + +# Start Workers +"$sbin"/start-slaves.sh diff --git a/sbin/start-master.sh b/sbin/start-master.sh new file mode 100755 index 0000000000..3dcf7cc348 --- /dev/null +++ b/sbin/start-master.sh @@ -0,0 +1,52 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Starts the master on the machine this script is executed on. + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +. "$sbin/spark-config.sh" + +if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then + . "${SPARK_CONF_DIR}/spark-env.sh" +fi + +if [ "$SPARK_MASTER_PORT" = "" ]; then + SPARK_MASTER_PORT=7077 +fi + +if [ "$SPARK_MASTER_IP" = "" ]; then + SPARK_MASTER_IP=`hostname` +fi + +if [ "$SPARK_MASTER_WEBUI_PORT" = "" ]; then + SPARK_MASTER_WEBUI_PORT=8080 +fi + +# Set SPARK_PUBLIC_DNS so the master report the correct webUI address to the slaves +if [ "$SPARK_PUBLIC_DNS" = "" ]; then + # If we appear to be running on EC2, use the public address by default: + # NOTE: ec2-metadata is installed on Amazon Linux AMI. Check based on that and hostname + if command -v ec2-metadata > /dev/null || [[ `hostname` == *ec2.internal ]]; then + export SPARK_PUBLIC_DNS=`wget -q -O - http://instance-data.ec2.internal/latest/meta-data/public-hostname` + fi +fi + +"$sbin"/spark-daemon.sh start org.apache.spark.deploy.master.Master 1 --ip $SPARK_MASTER_IP --port $SPARK_MASTER_PORT --webui-port $SPARK_MASTER_WEBUI_PORT diff --git a/sbin/start-slave.sh b/sbin/start-slave.sh new file mode 100755 index 0000000000..524be38c62 --- /dev/null +++ b/sbin/start-slave.sh @@ -0,0 +1,35 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Usage: start-slave.sh +# where is like "spark://localhost:7077" + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +# Set SPARK_PUBLIC_DNS so slaves can be linked in master web UI +if [ "$SPARK_PUBLIC_DNS" = "" ]; then + # If we appear to be running on EC2, use the public address by default: + # NOTE: ec2-metadata is installed on Amazon Linux AMI. Check based on that and hostname + if command -v ec2-metadata > /dev/null || [[ `hostname` == *ec2.internal ]]; then + export SPARK_PUBLIC_DNS=`wget -q -O - http://instance-data.ec2.internal/latest/meta-data/public-hostname` + fi +fi + +"$sbin"/spark-daemon.sh start org.apache.spark.deploy.worker.Worker "$@" diff --git a/sbin/start-slaves.sh b/sbin/start-slaves.sh new file mode 100755 index 0000000000..fd5cdeb1e6 --- /dev/null +++ b/sbin/start-slaves.sh @@ -0,0 +1,48 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +. "$sbin/spark-config.sh" + +if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then + . "${SPARK_CONF_DIR}/spark-env.sh" +fi + +# Find the port number for the master +if [ "$SPARK_MASTER_PORT" = "" ]; then + SPARK_MASTER_PORT=7077 +fi + +if [ "$SPARK_MASTER_IP" = "" ]; then + SPARK_MASTER_IP=`hostname` +fi + +# Launch the slaves +if [ "$SPARK_WORKER_INSTANCES" = "" ]; then + exec "$sbin/slaves.sh" cd "$SPARK_HOME" \; "$sbin/start-slave.sh" 1 spark://$SPARK_MASTER_IP:$SPARK_MASTER_PORT +else + if [ "$SPARK_WORKER_WEBUI_PORT" = "" ]; then + SPARK_WORKER_WEBUI_PORT=8081 + fi + for ((i=0; i<$SPARK_WORKER_INSTANCES; i++)); do + "$sbin/slaves.sh" cd "$SPARK_HOME" \; "$sbin/start-slave.sh" $(( $i + 1 )) spark://$SPARK_MASTER_IP:$SPARK_MASTER_PORT --webui-port $(( $SPARK_WORKER_WEBUI_PORT + $i )) + done +fi diff --git a/sbin/stop-all.sh b/sbin/stop-all.sh new file mode 100755 index 0000000000..60b358d374 --- /dev/null +++ b/sbin/stop-all.sh @@ -0,0 +1,32 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Start all spark daemons. +# Run this on the master nde + + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +# Load the Spark configuration +. "$sbin/spark-config.sh" + +# Stop the slaves, then the master +"$sbin"/stop-slaves.sh +"$sbin"/stop-master.sh diff --git a/sbin/stop-master.sh b/sbin/stop-master.sh new file mode 100755 index 0000000000..2adabd4265 --- /dev/null +++ b/sbin/stop-master.sh @@ -0,0 +1,27 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Starts the master on the machine this script is executed on. + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +. "$sbin/spark-config.sh" + +"$sbin"/spark-daemon.sh stop org.apache.spark.deploy.master.Master 1 diff --git a/sbin/stop-slaves.sh b/sbin/stop-slaves.sh new file mode 100755 index 0000000000..63802e6df5 --- /dev/null +++ b/sbin/stop-slaves.sh @@ -0,0 +1,37 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +# Starts the master on the machine this script is executed on. + +sbin=`dirname "$0"` +sbin=`cd "$sbin"; pwd` + +. "$sbin/spark-config.sh" + +if [ -f "${SPARK_CONF_DIR}/spark-env.sh" ]; then + . "${SPARK_CONF_DIR}/spark-env.sh" +fi + +if [ "$SPARK_WORKER_INSTANCES" = "" ]; then + "$sbin"/spark-daemons.sh stop org.apache.spark.deploy.worker.Worker 1 +else + for ((i=0; i<$SPARK_WORKER_INSTANCES; i++)); do + "$sbin"/spark-daemons.sh stop org.apache.spark.deploy.worker.Worker $(( $i + 1 )) + done +fi -- cgit v1.2.3 From 6ffa9bb226ac9ceec4a34f0011c35d2d9710f8f8 Mon Sep 17 00:00:00 2001 From: Patrick Wendell Date: Sun, 29 Dec 2013 11:26:56 -0800 Subject: Documentation and adding supervise option --- .../apache/spark/deploy/DriverDescription.scala | 1 + .../apache/spark/deploy/client/DriverClient.scala | 3 +- .../deploy/client/DriverClientArguments.scala | 12 ++++--- .../apache/spark/deploy/worker/DriverRunner.scala | 15 +++++---- docs/spark-standalone.md | 38 +++++++++++++++++++--- 5 files changed, 51 insertions(+), 18 deletions(-) (limited to 'docs') diff --git a/core/src/main/scala/org/apache/spark/deploy/DriverDescription.scala b/core/src/main/scala/org/apache/spark/deploy/DriverDescription.scala index aba81ec27c..58c95dc4f9 100644 --- a/core/src/main/scala/org/apache/spark/deploy/DriverDescription.scala +++ b/core/src/main/scala/org/apache/spark/deploy/DriverDescription.scala @@ -21,6 +21,7 @@ private[spark] class DriverDescription( val jarUrl: String, val mem: Int, val cores: Int, + val supervise: Boolean, val command: Command) extends Serializable { diff --git a/core/src/main/scala/org/apache/spark/deploy/client/DriverClient.scala b/core/src/main/scala/org/apache/spark/deploy/client/DriverClient.scala index 7e75563e60..6257303830 100644 --- a/core/src/main/scala/org/apache/spark/deploy/client/DriverClient.scala +++ b/core/src/main/scala/org/apache/spark/deploy/client/DriverClient.scala @@ -71,7 +71,7 @@ object DriverClient extends Logging { driverArgs.cmd match { case "launch" => // TODO: Could modify env here to pass a flag indicating Spark is in deploy-driver mode - // then use that to load jars locally + // then use that to load jars locally (e.g. truncate the filesystem path) val env = Map[String, String]() System.getenv().foreach{case (k, v) => env(k) = v} @@ -83,6 +83,7 @@ object DriverClient extends Logging { driverArgs.jarUrl, driverArgs.memory, driverArgs.cores, + driverArgs.supervise, command) driver ! RequestSubmitDriver(driverDescription) diff --git a/core/src/main/scala/org/apache/spark/deploy/client/DriverClientArguments.scala b/core/src/main/scala/org/apache/spark/deploy/client/DriverClientArguments.scala index 3875838319..6a15422c6c 100644 --- a/core/src/main/scala/org/apache/spark/deploy/client/DriverClientArguments.scala +++ b/core/src/main/scala/org/apache/spark/deploy/client/DriverClientArguments.scala @@ -29,6 +29,7 @@ private[spark] class DriverClientArguments(args: Array[String]) { var master: String = "" var jarUrl: String = "" var mainClass: String = "" + var supervise: Boolean = false var memory: Int = 512 var cores: Int = 1 private var _driverOptions = ListBuffer[String]() @@ -48,6 +49,10 @@ private[spark] class DriverClientArguments(args: Array[String]) { memory = value.toInt parse(tail) + case ("--supervise" | "-s") :: tail => + supervise = true + parse(tail) + case ("--help" | "-h") :: tail => printUsageAndExit(0) @@ -71,10 +76,6 @@ private[spark] class DriverClientArguments(args: Array[String]) { * Print usage and exit JVM with the given exit code. */ def printUsageAndExit(exitCode: Int) { - // TODO: Document the submission approach here. It is: - // 1) Create an uber jar with your application and dependencies (excluding Spark) - // 2) You'll need to add this jar using addJar(X) inside of your spark context - // TODO: It wouldn't be too hard to allow users to submit their app and dependency jars // separately similar to in the YARN client. System.err.println( @@ -83,7 +84,8 @@ private[spark] class DriverClientArguments(args: Array[String]) { "usage: DriverClient kill \n\n" + "Options:\n" + " -c CORES, --cores CORES Number of cores to request \n" + - " -m MEMORY, --memory MEMORY Megabytes of memory to request\n") + " -m MEMORY, --memory MEMORY Megabytes of memory to request\n" + + " -s, --supervise Whether to restart the driver on failure\n") System.exit(exitCode) } } diff --git a/core/src/main/scala/org/apache/spark/deploy/worker/DriverRunner.scala b/core/src/main/scala/org/apache/spark/deploy/worker/DriverRunner.scala index 8950fb71a1..41500bb28d 100644 --- a/core/src/main/scala/org/apache/spark/deploy/worker/DriverRunner.scala +++ b/core/src/main/scala/org/apache/spark/deploy/worker/DriverRunner.scala @@ -66,7 +66,7 @@ private[spark] class DriverRunner( val command = CommandUtils.buildCommandSeq(newCommand, driverDesc.mem, sparkHome.getAbsolutePath) - runCommandWithRetry(command, env, driverDir) + runCommand(command, env, driverDir, driverDesc.supervise) } catch { case e: Exception => exn = Some(e) @@ -137,13 +137,14 @@ private[spark] class DriverRunner( localJarFilename } - /** Continue launching the supplied command until it exits zero or is killed. */ - private def runCommandWithRetry(command: Seq[String], envVars: Map[String, String], baseDir: File) { + /** Launch the supplied command. */ + private def runCommand(command: Seq[String], envVars: Map[String, String], baseDir: File, + supervise: Boolean) { // Time to wait between submission retries. var waitSeconds = 1 - var cleanExit = false + var keepTrying = !killed - while (!cleanExit && !killed) { + while (keepTrying) { logInfo("Launch Command: " + command.mkString("\"", "\" \"", "\"")) val builder = new ProcessBuilder(command: _*).directory(baseDir) envVars.map{ case(k,v) => builder.environment().put(k, v) } @@ -166,8 +167,8 @@ private[spark] class DriverRunner( val exitCode = process.get.waitFor() - cleanExit = exitCode == 0 - if (!cleanExit && !killed) { + keepTrying = supervise && exitCode != 0 && !killed + if (keepTrying) { waitSeconds = waitSeconds * 2 // exponential back-off logInfo(s"Command exited with status $exitCode, re-launching after $waitSeconds s.") (0 until waitSeconds).takeWhile(f => {Thread.sleep(1000); !killed}) diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index b822265b5a..59adbce156 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -10,11 +10,7 @@ In addition to running on the Mesos or YARN cluster managers, Spark also provide # Installing Spark Standalone to a Cluster -The easiest way to deploy Spark is by running the `./make-distribution.sh` script to create a binary distribution. -This distribution can be deployed to any machine with the Java runtime installed; there is no need to install Scala. - -The recommended procedure is to deploy and start the master on one node first, get the master spark URL, -then modify `conf/spark-env.sh` in the `dist/` directory before deploying to all the other nodes. +To install Spark Standlone mode, you simply place a compiled version of Spark on each node on the cluster. You can obtain pre-built versions of Spark with each release or [build it yourself](index.html#building). # Starting a Cluster Manually @@ -150,6 +146,38 @@ automatically set MASTER from the `SPARK_MASTER_IP` and `SPARK_MASTER_PORT` vari You can also pass an option `-c ` to control the number of cores that spark-shell uses on the cluster. +# Launching Applications Inside the Cluster + +You may also run your application entirely inside of the cluster by submitting your application driver using the submission client. The syntax for submitting applications is as follows: + + + ./spark-class org.apache.spark.deploy.client.DriverClient launch + [client-options] \ + \ + [application-options] + + cluster-url: The URL of the master node. + application-jar-url: Path to a bundled jar including your application and all dependencies. + Accepts hdfs://, file://, and http:// paths. + main-class: The entry point for your application. + + Client Options: + --memory (amount of memory, in MB, allocated for your driver program) + --cores (number of cores allocated for your driver program) + --supervise (whether to automatically restart your driver on application or node failure) + +Keep in mind that your driver program will be executed on a remote worker machine. You can control the execution environment in the following ways: + + * _Environment variables_: These will be captured from the environment in which you launch the client and applied when launching the driver program. + * _Java options_: You can add java options by setting `SPARK_JAVA_OPTS` in the environment in which you launch the submission client. + * _Dependencies_: You'll still need to call `sc.addJar` inside of your driver program to add your application jar and any dependencies. If you submit a local application jar to the client (e.g one with a `file://` URL), it will be uploaded into the working directory of your driver program. Then, you can add it using `sc.addJar("jar-name.jar")`. + +Once you submit a driver program, it will appear in the cluster management UI at port 8080 and +be assigned an identifier. If you'd like to prematurely terminate the program, you can do so using +the same client: + + ./spark-class org.apache.spark.deploy.client.DriverClient kill + # Resource Scheduling The standalone cluster mode currently only supports a simple FIFO scheduler across applications. -- cgit v1.2.3 From 6be4c1119493dea2af9734ad8b59fcded31f2676 Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Thu, 2 Jan 2014 14:09:37 +0530 Subject: Removed sbt folder and changed docs accordingly --- README.md | 30 ++++++++++++++++----- docs/README.md | 4 +-- docs/_plugins/copy_api_dirs.rb | 4 +-- docs/api.md | 2 +- docs/hadoop-third-party-distributions.md | 2 +- docs/index.md | 6 ++--- docs/python-programming-guide.md | 2 +- docs/quick-start.md | 2 +- docs/running-on-yarn.md | 6 ++--- docs/scala-programming-guide.md | 2 +- make-distribution.sh | 12 +++++++-- pyspark | 2 +- run-example | 2 +- sbt/sbt | 43 ------------------------------- sbt/sbt-launch-0.11.3-2.jar | Bin 1096763 -> 0 bytes sbt/sbt.cmd | 25 ------------------ spark-class | 2 +- 17 files changed, 51 insertions(+), 95 deletions(-) delete mode 100755 sbt/sbt delete mode 100644 sbt/sbt-launch-0.11.3-2.jar delete mode 100644 sbt/sbt.cmd (limited to 'docs') diff --git a/README.md b/README.md index 1550a8b551..22e7ab8245 100644 --- a/README.md +++ b/README.md @@ -13,9 +13,9 @@ This README file only contains basic setup instructions. ## Building Spark requires Scala 2.10. The project is built using Simple Build Tool (SBT), -which is packaged with it. To build Spark and its example programs, run: +which can be obtained from [here](http://www.scala-sbt.org/release/docs/Getting-Started/Setup.html). To build Spark and its example programs, run: - sbt/sbt assembly + sbt assembly Once you've built Spark, the easiest way to start using it is the shell: @@ -36,6 +36,22 @@ All of the Spark samples take a `` parameter that is the cluster URL to connect to. This can be a mesos:// or spark:// URL, or "local" to run locally with one thread, or "local[N]" to run locally with N threads. +## Running tests + +### With sbt. (you need sbt installed) +Once you have built spark with `sbt assembly` mentioned in [Building](#Building) section. Test suits can be run as follows on *nix based systems using sbt. + +`SPARK_HOME=$(pwd) SPARK_TESTING=1 sbt test` + +TODO: figure out instructions for windows. + +### With maven. + +1. Build assembly by +`mvn package -DskipTests` + +2. Run tests +`mvn test` ## A Note About Hadoop Versions @@ -49,22 +65,22 @@ For Apache Hadoop versions 1.x, Cloudera CDH MRv1, and other Hadoop versions without YARN, use: # Apache Hadoop 1.2.1 - $ SPARK_HADOOP_VERSION=1.2.1 sbt/sbt assembly + $ SPARK_HADOOP_VERSION=1.2.1 sbt assembly # Cloudera CDH 4.2.0 with MapReduce v1 - $ SPARK_HADOOP_VERSION=2.0.0-mr1-cdh4.2.0 sbt/sbt assembly + $ SPARK_HADOOP_VERSION=2.0.0-mr1-cdh4.2.0 sbt assembly For Apache Hadoop 2.2.X, 2.1.X, 2.0.X, 0.23.x, Cloudera CDH MRv2, and other Hadoop versions with YARN, also set `SPARK_YARN=true`: # Apache Hadoop 2.0.5-alpha - $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt/sbt assembly + $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt assembly # Cloudera CDH 4.2.0 with MapReduce v2 - $ SPARK_HADOOP_VERSION=2.0.0-cdh4.2.0 SPARK_YARN=true sbt/sbt assembly + $ SPARK_HADOOP_VERSION=2.0.0-cdh4.2.0 SPARK_YARN=true sbt assembly # Apache Hadoop 2.2.X and newer - $ SPARK_HADOOP_VERSION=2.2.0 SPARK_YARN=true sbt/sbt assembly + $ SPARK_HADOOP_VERSION=2.2.0 SPARK_YARN=true sbt assembly When developing a Spark application, specify the Hadoop version by adding the "hadoop-client" artifact to your project's dependencies. For example, if you're diff --git a/docs/README.md b/docs/README.md index dfcf753553..e3d6c9a5bc 100644 --- a/docs/README.md +++ b/docs/README.md @@ -27,10 +27,10 @@ To mark a block of code in your markdown to be syntax highlighted by jekyll duri ## API Docs (Scaladoc and Epydoc) -You can build just the Spark scaladoc by running `sbt/sbt doc` from the SPARK_PROJECT_ROOT directory. +You can build just the Spark scaladoc by running `sbt doc` from the SPARK_PROJECT_ROOT directory. Similarly, you can build just the PySpark epydoc by running `epydoc --config epydoc.conf` from the SPARK_PROJECT_ROOT/pyspark directory. -When you run `jekyll` in the docs directory, it will also copy over the scaladoc for the various Spark subprojects into the docs directory (and then also into the _site directory). We use a jekyll plugin to run `sbt/sbt doc` before building the site so if you haven't run it (recently) it may take some time as it generates all of the scaladoc. The jekyll plugin also generates the PySpark docs using [epydoc](http://epydoc.sourceforge.net/). +When you run `jekyll` in the docs directory, it will also copy over the scaladoc for the various Spark subprojects into the docs directory (and then also into the _site directory). We use a jekyll plugin to run `sbt doc` before building the site so if you haven't run it (recently) it may take some time as it generates all of the scaladoc. The jekyll plugin also generates the PySpark docs using [epydoc](http://epydoc.sourceforge.net/). NOTE: To skip the step of building and copying over the Scala and Python API docs, run `SKIP_API=1 jekyll`. diff --git a/docs/_plugins/copy_api_dirs.rb b/docs/_plugins/copy_api_dirs.rb index 431de909cb..ef9912c808 100644 --- a/docs/_plugins/copy_api_dirs.rb +++ b/docs/_plugins/copy_api_dirs.rb @@ -26,8 +26,8 @@ if not (ENV['SKIP_API'] == '1' or ENV['SKIP_SCALADOC'] == '1') curr_dir = pwd cd("..") - puts "Running sbt/sbt doc from " + pwd + "; this may take a few minutes..." - puts `sbt/sbt doc` + puts "Running sbt doc from " + pwd + "; this may take a few minutes..." + puts `sbt doc` puts "Moving back into docs dir." cd("docs") diff --git a/docs/api.md b/docs/api.md index e86d07770a..11e2c15324 100644 --- a/docs/api.md +++ b/docs/api.md @@ -3,7 +3,7 @@ layout: global title: Spark API documentation (Scaladoc) --- -Here you can find links to the Scaladoc generated for the Spark sbt subprojects. If the following links don't work, try running `sbt/sbt doc` from the Spark project home directory. +Here you can find links to the Scaladoc generated for the Spark sbt subprojects. If the following links don't work, try running `sbt doc` from the Spark project home directory. - [Spark](api/core/index.html) - [Spark Examples](api/examples/index.html) diff --git a/docs/hadoop-third-party-distributions.md b/docs/hadoop-third-party-distributions.md index de6a2b0a43..141d475ba6 100644 --- a/docs/hadoop-third-party-distributions.md +++ b/docs/hadoop-third-party-distributions.md @@ -12,7 +12,7 @@ with these distributions: When compiling Spark, you'll need to [set the SPARK_HADOOP_VERSION flag](index.html#a-note-about-hadoop-versions): - SPARK_HADOOP_VERSION=1.0.4 sbt/sbt assembly + SPARK_HADOOP_VERSION=1.0.4 sbt assembly The table below lists the corresponding `SPARK_HADOOP_VERSION` code for each CDH/HDP release. Note that some Hadoop releases are binary compatible across client versions. This means the pre-built Spark diff --git a/docs/index.md b/docs/index.md index d3ac696d1e..5278e33e1c 100644 --- a/docs/index.md +++ b/docs/index.md @@ -17,7 +17,7 @@ Spark runs on both Windows and UNIX-like systems (e.g. Linux, Mac OS). All you n Spark uses [Simple Build Tool](http://www.scala-sbt.org), which is bundled with it. To compile the code, go into the top-level Spark directory and run - sbt/sbt assembly + sbt assembly For its Scala API, Spark {{site.SPARK_VERSION}} depends on Scala {{site.SCALA_VERSION}}. If you write applications in Scala, you will need to use this same version of Scala in your own program -- newer major versions may not work. You can get the right version of Scala from [scala-lang.org](http://www.scala-lang.org/download/). @@ -56,12 +56,12 @@ Hadoop, you must build Spark against the same version that your cluster uses. By default, Spark links to Hadoop 1.0.4. You can change this by setting the `SPARK_HADOOP_VERSION` variable when compiling: - SPARK_HADOOP_VERSION=2.2.0 sbt/sbt assembly + SPARK_HADOOP_VERSION=2.2.0 sbt assembly In addition, if you wish to run Spark on [YARN](running-on-yarn.html), set `SPARK_YARN` to `true`: - SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt/sbt assembly + SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt assembly Note that on Windows, you need to set the environment variables on separate lines, e.g., `set SPARK_HADOOP_VERSION=1.2.1`. diff --git a/docs/python-programming-guide.md b/docs/python-programming-guide.md index 55e39b1de1..a33977ed82 100644 --- a/docs/python-programming-guide.md +++ b/docs/python-programming-guide.md @@ -69,7 +69,7 @@ The script automatically adds the `pyspark` package to the `PYTHONPATH`. The `pyspark` script launches a Python interpreter that is configured to run PySpark applications. To use `pyspark` interactively, first build Spark, then launch it directly from the command line without any options: {% highlight bash %} -$ sbt/sbt assembly +$ sbt assembly $ ./pyspark {% endhighlight %} diff --git a/docs/quick-start.md b/docs/quick-start.md index 8f782db5b8..5c55def398 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -12,7 +12,7 @@ See the [programming guide](scala-programming-guide.html) for a more complete re To follow along with this guide, you only need to have successfully built Spark on one machine. Simply go into your Spark directory and run: {% highlight bash %} -$ sbt/sbt assembly +$ sbt assembly {% endhighlight %} # Interactive Analysis with the Spark Shell diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index aa75ca4324..13d5fd3685 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -12,7 +12,7 @@ was added to Spark in version 0.6.0, and improved in 0.7.0 and 0.8.0. We need a consolidated Spark JAR (which bundles all the required dependencies) to run Spark jobs on a YARN cluster. This can be built by setting the Hadoop version and `SPARK_YARN` environment variable, as follows: - SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true ./sbt/sbt assembly + SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true ./sbt assembly The assembled JAR will be something like this: `./assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly_{{site.SPARK_VERSION}}-hadoop2.0.5.jar`. @@ -25,7 +25,7 @@ The build process now also supports new YARN versions (2.2.x). See below. - The assembled jar can be installed into HDFS or used locally. - Your application code must be packaged into a separate JAR file. -If you want to test out the YARN deployment mode, you can use the current Spark examples. A `spark-examples_{{site.SCALA_VERSION}}-{{site.SPARK_VERSION}}` file can be generated by running `sbt/sbt assembly`. NOTE: since the documentation you're reading is for Spark version {{site.SPARK_VERSION}}, we are assuming here that you have downloaded Spark {{site.SPARK_VERSION}} or checked it out of source control. If you are using a different version of Spark, the version numbers in the jar generated by the sbt package command will obviously be different. +If you want to test out the YARN deployment mode, you can use the current Spark examples. A `spark-examples_{{site.SCALA_VERSION}}-{{site.SPARK_VERSION}}` file can be generated by running `sbt assembly`. NOTE: since the documentation you're reading is for Spark version {{site.SPARK_VERSION}}, we are assuming here that you have downloaded Spark {{site.SPARK_VERSION}} or checked it out of source control. If you are using a different version of Spark, the version numbers in the jar generated by the sbt package command will obviously be different. # Configuration @@ -72,7 +72,7 @@ The command to launch the YARN Client is as follows: For example: # Build the Spark assembly JAR and the Spark examples JAR - $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true ./sbt/sbt assembly + $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true ./sbt assembly # Configure logging $ cp conf/log4j.properties.template conf/log4j.properties diff --git a/docs/scala-programming-guide.md b/docs/scala-programming-guide.md index 56d2a3a4a0..3e7075c382 100644 --- a/docs/scala-programming-guide.md +++ b/docs/scala-programming-guide.md @@ -31,7 +31,7 @@ In addition, if you wish to access an HDFS cluster, you need to add a dependency artifactId = hadoop-client version = -For other build systems, you can run `sbt/sbt assembly` to pack Spark and its dependencies into one JAR (`assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop*.jar`), then add this to your CLASSPATH. Set the HDFS version as described [here](index.html#a-note-about-hadoop-versions). +For other build systems, you can run `sbt assembly` to pack Spark and its dependencies into one JAR (`assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop*.jar`), then add this to your CLASSPATH. Set the HDFS version as described [here](index.html#a-note-about-hadoop-versions). Finally, you need to import some Spark classes and implicit conversions into your program. Add the following lines: diff --git a/make-distribution.sh b/make-distribution.sh index 32bbdb90a5..a2c8e64597 100755 --- a/make-distribution.sh +++ b/make-distribution.sh @@ -43,7 +43,13 @@ DISTDIR="$FWDIR/dist" # Get version from SBT export TERM=dumb # Prevents color codes in SBT output -VERSION=$($FWDIR/sbt/sbt "show version" | tail -1 | cut -f 2 | sed 's/^\([a-zA-Z0-9.-]*\).*/\1/') + +if ! test `which sbt` ;then + echo -e "You need sbt installed and available on path, please follow the instructions here: http://www.scala-sbt.org/release/docs/Getting-Started/Setup.html" + exit -1; +fi + +VERSION=$(sbt "show version" | tail -1 | cut -f 2 | sed 's/^\([a-zA-Z0-9.-]*\).*/\1/') # Initialize defaults SPARK_HADOOP_VERSION=1.0.4 @@ -83,7 +89,9 @@ fi # Build fat JAR export SPARK_HADOOP_VERSION export SPARK_YARN -"$FWDIR/sbt/sbt" "assembly/assembly" +cd $FWDIR + +"sbt" "assembly/assembly" # Make directories rm -rf "$DISTDIR" diff --git a/pyspark b/pyspark index 12cc926dda..1d003e2a00 100755 --- a/pyspark +++ b/pyspark @@ -31,7 +31,7 @@ if [ ! -f "$FWDIR/RELEASE" ]; then ls "$FWDIR"/assembly/target/scala-$SCALA_VERSION/spark-assembly*hadoop*.jar >& /dev/null if [[ $? != 0 ]]; then echo "Failed to find Spark assembly in $FWDIR/assembly/target" >&2 - echo "You need to build Spark with sbt/sbt assembly before running this program" >&2 + echo "You need to build Spark with sbt assembly before running this program" >&2 exit 1 fi fi diff --git a/run-example b/run-example index a78192d31d..fbd81fe6f3 100755 --- a/run-example +++ b/run-example @@ -55,7 +55,7 @@ if [ -e "$EXAMPLES_DIR"/target/spark-examples*[0-9Tg].jar ]; then fi if [[ -z $SPARK_EXAMPLES_JAR ]]; then echo "Failed to find Spark examples assembly in $FWDIR/examples/target" >&2 - echo "You need to build Spark with sbt/sbt assembly before running this program" >&2 + echo "You need to build Spark with sbt assembly before running this program" >&2 exit 1 fi diff --git a/sbt/sbt b/sbt/sbt deleted file mode 100755 index 5942280585..0000000000 --- a/sbt/sbt +++ /dev/null @@ -1,43 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -cygwin=false -case "`uname`" in - CYGWIN*) cygwin=true;; -esac - -EXTRA_ARGS="-Xmx1200m -XX:MaxPermSize=350m -XX:ReservedCodeCacheSize=256m" -if [ "$MESOS_HOME" != "" ]; then - EXTRA_ARGS="$EXTRA_ARGS -Djava.library.path=$MESOS_HOME/lib/java" -fi - -export SPARK_HOME=$(cd "$(dirname $0)/.." 2>&1 >/dev/null ; pwd) -export SPARK_TESTING=1 # To put test classes on classpath - -SBT_JAR="$SPARK_HOME"/sbt/sbt-launch-*.jar -if $cygwin; then - SBT_JAR=`cygpath -w $SBT_JAR` - export SPARK_HOME=`cygpath -w $SPARK_HOME` - EXTRA_ARGS="$EXTRA_ARGS -Djline.terminal=jline.UnixTerminal -Dsbt.cygwin=true" - stty -icanon min 1 -echo > /dev/null 2>&1 - java $EXTRA_ARGS $SBT_OPTS -jar $SBT_JAR "$@" - stty icanon echo > /dev/null 2>&1 -else - java $EXTRA_ARGS $SBT_OPTS -jar $SBT_JAR "$@" -fi \ No newline at end of file diff --git a/sbt/sbt-launch-0.11.3-2.jar b/sbt/sbt-launch-0.11.3-2.jar deleted file mode 100644 index 23e5c3f311..0000000000 Binary files a/sbt/sbt-launch-0.11.3-2.jar and /dev/null differ diff --git a/sbt/sbt.cmd b/sbt/sbt.cmd deleted file mode 100644 index 681fe00f92..0000000000 --- a/sbt/sbt.cmd +++ /dev/null @@ -1,25 +0,0 @@ -@echo off - -rem -rem Licensed to the Apache Software Foundation (ASF) under one or more -rem contributor license agreements. See the NOTICE file distributed with -rem this work for additional information regarding copyright ownership. -rem The ASF licenses this file to You under the Apache License, Version 2.0 -rem (the "License"); you may not use this file except in compliance with -rem the License. You may obtain a copy of the License at -rem -rem http://www.apache.org/licenses/LICENSE-2.0 -rem -rem Unless required by applicable law or agreed to in writing, software -rem distributed under the License is distributed on an "AS IS" BASIS, -rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -rem See the License for the specific language governing permissions and -rem limitations under the License. -rem - -set EXTRA_ARGS= -if not "%MESOS_HOME%x"=="x" set EXTRA_ARGS=-Djava.library.path=%MESOS_HOME%\lib\java - -set SPARK_HOME=%~dp0.. - -java -Xmx1200M -XX:MaxPermSize=200m -XX:ReservedCodeCacheSize=256m %EXTRA_ARGS% -jar %SPARK_HOME%\sbt\sbt-launch-0.11.3-2.jar "%*" diff --git a/spark-class b/spark-class index 1858ea6247..254ddee04a 100755 --- a/spark-class +++ b/spark-class @@ -104,7 +104,7 @@ if [ ! -f "$FWDIR/RELEASE" ]; then jars_list=$(ls "$FWDIR"/assembly/target/scala-$SCALA_VERSION/ | grep "spark-assembly.*hadoop.*.jar") if [ "$num_jars" -eq "0" ]; then echo "Failed to find Spark assembly in $FWDIR/assembly/target/scala-$SCALA_VERSION/" >&2 - echo "You need to build Spark with 'sbt/sbt assembly' before running this program." >&2 + echo "You need to build Spark with 'sbt assembly' before running this program." >&2 exit 1 fi if [ "$num_jars" -gt "1" ]; then -- cgit v1.2.3 From b810a85cdddb247e1a104f4daad905b97222ad85 Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Thu, 2 Jan 2014 18:37:40 +0530 Subject: spark-shell -> bin/spark-shell --- README.md | 2 +- docs/index.md | 2 +- docs/mllib-guide.md | 2 +- docs/quick-start.md | 4 ++-- docs/running-on-yarn.md | 2 +- docs/scala-programming-guide.md | 10 +++++----- docs/spark-debugger.md | 2 +- docs/spark-standalone.md | 4 ++-- make-distribution.sh | 2 +- 9 files changed, 15 insertions(+), 15 deletions(-) (limited to 'docs') diff --git a/README.md b/README.md index 1550a8b551..170e964851 100644 --- a/README.md +++ b/README.md @@ -19,7 +19,7 @@ which is packaged with it. To build Spark and its example programs, run: Once you've built Spark, the easiest way to start using it is the shell: - ./spark-shell + ./bin/spark-shell Or, for the Python API, the Python shell (`./pyspark`). diff --git a/docs/index.md b/docs/index.md index d3ac696d1e..a5c179c19c 100644 --- a/docs/index.md +++ b/docs/index.md @@ -34,7 +34,7 @@ to connect to. This can be a [URL for a distributed cluster](scala-programming-g or `local` to run locally with one thread, or `local[N]` to run locally with N threads. You should start by using `local` for testing. -Finally, you can run Spark interactively through modified versions of the Scala shell (`./spark-shell`) or +Finally, you can run Spark interactively through modified versions of the Scala shell (`./bin/spark-shell`) or Python interpreter (`./pyspark`). These are a great way to learn the framework. # Launching on a Cluster diff --git a/docs/mllib-guide.md b/docs/mllib-guide.md index c1ff9c417c..95537ef185 100644 --- a/docs/mllib-guide.md +++ b/docs/mllib-guide.md @@ -87,7 +87,7 @@ svmAlg.optimizer.setNumIterations(200) val modelL1 = svmAlg.run(parsedData) {% endhighlight %} -Both of the code snippets above can be executed in `spark-shell` to generate a +Both of the code snippets above can be executed in `bin/spark-shell` to generate a classifier for the provided dataset. Available algorithms for binary classification: diff --git a/docs/quick-start.md b/docs/quick-start.md index bdf127d404..912f332236 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -20,7 +20,7 @@ $ sbt/sbt assembly ## Basics Spark's interactive shell provides a simple way to learn the API, as well as a powerful tool to analyze datasets interactively. -Start the shell by running `./spark-shell` in the Spark directory. +Start the shell by running `./bin/spark-shell` in the Spark directory. Spark's primary abstraction is a distributed collection of items called a Resilient Distributed Dataset (RDD). RDDs can be created from Hadoop InputFormats (such as HDFS files) or by transforming other RDDs. Let's make a new RDD from the text of the README file in the Spark source directory: @@ -99,7 +99,7 @@ scala> linesWithSpark.count() res9: Long = 15 {% endhighlight %} -It may seem silly to use Spark to explore and cache a 30-line text file. The interesting part is that these same functions can be used on very large data sets, even when they are striped across tens or hundreds of nodes. You can also do this interactively by connecting `spark-shell` to a cluster, as described in the [programming guide](scala-programming-guide.html#initializing-spark). +It may seem silly to use Spark to explore and cache a 30-line text file. The interesting part is that these same functions can be used on very large data sets, even when they are striped across tens or hundreds of nodes. You can also do this interactively by connecting `bin/spark-shell` to a cluster, as described in the [programming guide](scala-programming-guide.html#initializing-spark). # A Standalone App in Scala Now say we wanted to write a standalone application using the Spark API. We will walk through a simple application in both Scala (with SBT), Java (with Maven), and Python. If you are using other build systems, consider using the Spark assembly JAR described in the developer guide. diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index 00cad99881..be32034b3e 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -112,7 +112,7 @@ For example: SPARK_JAR=./assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop2.0.5-alpha.jar \ SPARK_YARN_APP_JAR=examples/target/scala-{{site.SCALA_VERSION}}/spark-examples-assembly-{{site.SPARK_VERSION}}.jar \ - MASTER=yarn-client ./spark-shell + MASTER=yarn-client ./bin/spark-shell # Building Spark for Hadoop/YARN 2.2.x diff --git a/docs/scala-programming-guide.md b/docs/scala-programming-guide.md index 1db255ca53..16ae15046a 100644 --- a/docs/scala-programming-guide.md +++ b/docs/scala-programming-guide.md @@ -13,7 +13,7 @@ At a high level, every Spark application consists of a *driver program* that run A second abstraction in Spark is *shared variables* that can be used in parallel operations. By default, when Spark runs a function in parallel as a set of tasks on different nodes, it ships a copy of each variable used in the function to each task. Sometimes, a variable needs to be shared across tasks, or between tasks and the driver program. Spark supports two types of shared variables: *broadcast variables*, which can be used to cache a value in memory on all nodes, and *accumulators*, which are variables that are only "added" to, such as counters and sums. -This guide shows each of these features and walks through some samples. It assumes some familiarity with Scala, especially with the syntax for [closures](http://www.scala-lang.org/node/133). Note that you can also run Spark interactively using the `spark-shell` script. We highly recommend doing that to follow along! +This guide shows each of these features and walks through some samples. It assumes some familiarity with Scala, especially with the syntax for [closures](http://www.scala-lang.org/node/133). Note that you can also run Spark interactively using the `bin/spark-shell` script. We highly recommend doing that to follow along! # Linking with Spark @@ -54,16 +54,16 @@ object for more advanced configuration. The `master` parameter is a string specifying a [Spark or Mesos cluster URL](#master-urls) to connect to, or a special "local" string to run in local mode, as described below. `appName` is a name for your application, which will be shown in the cluster web UI. Finally, the last two parameters are needed to deploy your code to a cluster if running in distributed mode, as described later. -In the Spark shell, a special interpreter-aware SparkContext is already created for you, in the variable called `sc`. Making your own SparkContext will not work. You can set which master the context connects to using the `MASTER` environment variable, and you can add JARs to the classpath with the `ADD_JARS` variable. For example, to run `spark-shell` on four cores, use +In the Spark shell, a special interpreter-aware SparkContext is already created for you, in the variable called `sc`. Making your own SparkContext will not work. You can set which master the context connects to using the `MASTER` environment variable, and you can add JARs to the classpath with the `ADD_JARS` variable. For example, to run `bin/spark-shell` on four cores, use {% highlight bash %} -$ MASTER=local[4] ./spark-shell +$ MASTER=local[4] ./bin/spark-shell {% endhighlight %} Or, to also add `code.jar` to its classpath, use: {% highlight bash %} -$ MASTER=local[4] ADD_JARS=code.jar ./spark-shell +$ MASTER=local[4] ADD_JARS=code.jar ./bin/spark-shell {% endhighlight %} ### Master URLs @@ -95,7 +95,7 @@ If you want to run your application on a cluster, you will need to specify the t * `sparkHome`: The path at which Spark is installed on your worker machines (it should be the same on all of them). * `jars`: A list of JAR files on the local machine containing your application's code and any dependencies, which Spark will deploy to all the worker nodes. You'll need to package your application into a set of JARs using your build system. For example, if you're using SBT, the [sbt-assembly](https://github.com/sbt/sbt-assembly) plugin is a good way to make a single JAR with your code and dependencies. -If you run `spark-shell` on a cluster, you can add JARs to it by specifying the `ADD_JARS` environment variable before you launch it. This variable should contain a comma-separated list of JARs. For example, `ADD_JARS=a.jar,b.jar ./spark-shell` will launch a shell with `a.jar` and `b.jar` on its classpath. In addition, any new classes you define in the shell will automatically be distributed. +If you run `bin/spark-shell` on a cluster, you can add JARs to it by specifying the `ADD_JARS` environment variable before you launch it. This variable should contain a comma-separated list of JARs. For example, `ADD_JARS=a.jar,b.jar ./bin/spark-shell` will launch a shell with `a.jar` and `b.jar` on its classpath. In addition, any new classes you define in the shell will automatically be distributed. # Resilient Distributed Datasets (RDDs) diff --git a/docs/spark-debugger.md b/docs/spark-debugger.md index d6315d97f4..11c51d5cde 100644 --- a/docs/spark-debugger.md +++ b/docs/spark-debugger.md @@ -39,7 +39,7 @@ where `path/to/event-log` is where you want the event log to go relative to `$SP ### Loading the event log into the debugger -1. Run a Spark shell with `MASTER=host ./spark-shell`. +1. Run a Spark shell with `MASTER=host ./bin/spark-shell`. 2. Use `EventLogReader` to load the event log as follows: {% highlight scala %} spark> val r = new spark.EventLogReader(sc, Some("path/to/event-log")) diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index c2ce23081a..ec5ae5b4f7 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -143,9 +143,9 @@ constructor](scala-programming-guide.html#initializing-spark). To run an interactive Spark shell against the cluster, run the following command: - MASTER=spark://IP:PORT ./spark-shell + MASTER=spark://IP:PORT ./bin/spark-shell -Note that if you are running spark-shell from one of the spark cluster machines, the `spark-shell` script will +Note that if you are running spark-shell from one of the spark cluster machines, the `bin/spark-shell` script will automatically set MASTER from the `SPARK_MASTER_IP` and `SPARK_MASTER_PORT` variables in `conf/spark-env.sh`. You can also pass an option `-c ` to control the number of cores that spark-shell uses on the cluster. diff --git a/make-distribution.sh b/make-distribution.sh index 82638a23a0..8765c7e620 100755 --- a/make-distribution.sh +++ b/make-distribution.sh @@ -34,7 +34,7 @@ # 2) cd to deploy dir; ./bin/start-master.sh # 3) Verify master is up by visiting web page, ie http://master-ip:8080. Note the spark:// URL. # 4) ./bin/start-slave.sh 1 <> -# 5) MASTER="spark://my-master-ip:7077" ./spark-shell +# 5) MASTER="spark://my-master-ip:7077" ./bin/spark-shell # # Figure out where the Spark framework is installed -- cgit v1.2.3 From 94b7a7fe37a4b1459bfdbece2a4162451d6a8ac2 Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Thu, 2 Jan 2014 18:41:21 +0530 Subject: run-example -> bin/run-example --- README.md | 4 ++-- bin/run-example | 2 +- bin/run-example2.cmd | 2 +- docs/bagel-programming-guide.md | 4 ++-- docs/index.md | 6 +++--- docs/java-programming-guide.md | 4 ++-- docs/running-on-yarn.md | 2 +- docs/scala-programming-guide.md | 4 ++-- docs/streaming-programming-guide.md | 4 ++-- .../org/apache/spark/streaming/examples/JavaKafkaWordCount.java | 2 +- .../scala/org/apache/spark/streaming/examples/ActorWordCount.scala | 4 ++-- .../scala/org/apache/spark/streaming/examples/HdfsWordCount.scala | 2 +- .../scala/org/apache/spark/streaming/examples/KafkaWordCount.scala | 2 +- .../scala/org/apache/spark/streaming/examples/MQTTWordCount.scala | 4 ++-- .../org/apache/spark/streaming/examples/NetworkWordCount.scala | 2 +- .../apache/spark/streaming/examples/StatefulNetworkWordCount.scala | 2 +- .../scala/org/apache/spark/streaming/examples/ZeroMQWordCount.scala | 4 ++-- .../spark/streaming/examples/clickstream/PageViewGenerator.scala | 4 ++-- .../spark/streaming/examples/clickstream/PageViewStream.scala | 4 ++-- 19 files changed, 31 insertions(+), 31 deletions(-) (limited to 'docs') diff --git a/README.md b/README.md index 170e964851..7154165ab1 100644 --- a/README.md +++ b/README.md @@ -24,9 +24,9 @@ Once you've built Spark, the easiest way to start using it is the shell: Or, for the Python API, the Python shell (`./pyspark`). Spark also comes with several sample programs in the `examples` directory. -To run one of them, use `./run-example `. For example: +To run one of them, use `./bin/run-example `. For example: - ./run-example org.apache.spark.examples.SparkLR local[2] + ./bin/run-example org.apache.spark.examples.SparkLR local[2] will run the Logistic Regression example locally on 2 CPUs. diff --git a/bin/run-example b/bin/run-example index f2699c38a9..a3976beb12 100755 --- a/bin/run-example +++ b/bin/run-example @@ -36,7 +36,7 @@ if [ -e "$FWDIR/conf/spark-env.sh" ] ; then fi if [ -z "$1" ]; then - echo "Usage: run-example []" >&2 + echo "Usage: bin/run-example []" >&2 exit 1 fi diff --git a/bin/run-example2.cmd b/bin/run-example2.cmd index 6861334cb0..d9c10545e7 100644 --- a/bin/run-example2.cmd +++ b/bin/run-example2.cmd @@ -30,7 +30,7 @@ if exist "%FWDIR%conf\spark-env.cmd" call "%FWDIR%conf\spark-env.cmd" rem Test that an argument was given if not "x%1"=="x" goto arg_given - echo Usage: run-example ^ [^] + echo Usage: bin/run-example ^ [^] goto exit :arg_given diff --git a/docs/bagel-programming-guide.md b/docs/bagel-programming-guide.md index de001e6c52..c4f1f6d6ad 100644 --- a/docs/bagel-programming-guide.md +++ b/docs/bagel-programming-guide.md @@ -157,8 +157,8 @@ trait Message[K] { # Where to Go from Here -Two example jobs, PageRank and shortest path, are included in `examples/src/main/scala/org/apache/spark/examples/bagel`. You can run them by passing the class name to the `run-example` script included in Spark; e.g.: +Two example jobs, PageRank and shortest path, are included in `examples/src/main/scala/org/apache/spark/examples/bagel`. You can run them by passing the class name to the `bin/run-example` script included in Spark; e.g.: - ./run-example org.apache.spark.examples.bagel.WikipediaPageRank + ./bin/run-example org.apache.spark.examples.bagel.WikipediaPageRank Each example program prints usage help when run without any arguments. diff --git a/docs/index.md b/docs/index.md index a5c179c19c..f0990fff73 100644 --- a/docs/index.md +++ b/docs/index.md @@ -24,9 +24,9 @@ For its Scala API, Spark {{site.SPARK_VERSION}} depends on Scala {{site.SCALA_VE # Running the Examples and Shell Spark comes with several sample programs in the `examples` directory. -To run one of the samples, use `./run-example ` in the top-level Spark directory -(the `run-example` script sets up the appropriate paths and launches that program). -For example, try `./run-example org.apache.spark.examples.SparkPi local`. +To run one of the samples, use `./bin/run-example ` in the top-level Spark directory +(the `bin/run-example` script sets up the appropriate paths and launches that program). +For example, try `./bin/run-example org.apache.spark.examples.SparkPi local`. Each example prints usage help when run with no parameters. Note that all of the sample programs take a `` parameter specifying the cluster URL diff --git a/docs/java-programming-guide.md b/docs/java-programming-guide.md index 53085cc671..07732fa122 100644 --- a/docs/java-programming-guide.md +++ b/docs/java-programming-guide.md @@ -190,9 +190,9 @@ We hope to generate documentation with Java-style syntax in the future. Spark includes several sample programs using the Java API in [`examples/src/main/java`](https://github.com/apache/incubator-spark/tree/master/examples/src/main/java/org/apache/spark/examples). You can run them by passing the class name to the -`run-example` script included in Spark; for example: +`bin/run-example` script included in Spark; for example: - ./run-example org.apache.spark.examples.JavaWordCount + ./bin/run-example org.apache.spark.examples.JavaWordCount Each example program prints usage help when run without any arguments. diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index be32034b3e..e2b21f9cde 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -107,7 +107,7 @@ For example: SPARK_JAR=./assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop2.0.5-alpha.jar \ SPARK_YARN_APP_JAR=examples/target/scala-{{site.SCALA_VERSION}}/spark-examples-assembly-{{site.SPARK_VERSION}}.jar \ - ./run-example org.apache.spark.examples.SparkPi yarn-client + ./bin/run-example org.apache.spark.examples.SparkPi yarn-client SPARK_JAR=./assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop2.0.5-alpha.jar \ diff --git a/docs/scala-programming-guide.md b/docs/scala-programming-guide.md index 16ae15046a..c1ef46a1cd 100644 --- a/docs/scala-programming-guide.md +++ b/docs/scala-programming-guide.md @@ -366,9 +366,9 @@ res2: Int = 10 # Where to Go from Here You can see some [example Spark programs](http://spark.incubator.apache.org/examples.html) on the Spark website. -In addition, Spark includes several samples in `examples/src/main/scala`. Some of them have both Spark versions and local (non-parallel) versions, allowing you to see what had to be changed to make the program run on a cluster. You can run them using by passing the class name to the `run-example` script included in Spark; for example: +In addition, Spark includes several samples in `examples/src/main/scala`. Some of them have both Spark versions and local (non-parallel) versions, allowing you to see what had to be changed to make the program run on a cluster. You can run them using by passing the class name to the `bin/run-example` script included in Spark; for example: - ./run-example org.apache.spark.examples.SparkPi + ./bin/run-example org.apache.spark.examples.SparkPi Each example program prints usage help when run without any arguments. diff --git a/docs/streaming-programming-guide.md b/docs/streaming-programming-guide.md index 6f97db6a3e..1c9ece6270 100644 --- a/docs/streaming-programming-guide.md +++ b/docs/streaming-programming-guide.md @@ -245,7 +245,7 @@ $ nc -lk 9999 Then, in a different terminal, you can start NetworkWordCount by using {% highlight bash %} -$ ./run-example org.apache.spark.streaming.examples.NetworkWordCount local[2] localhost 9999 +$ ./bin/run-example org.apache.spark.streaming.examples.NetworkWordCount local[2] localhost 9999 {% endhighlight %} This will make NetworkWordCount connect to the netcat server. Any lines typed in the terminal running the netcat server will be counted and printed on screen. @@ -283,7 +283,7 @@ Time: 1357008430000 ms -You can find more examples in `/streaming/src/main/scala/org/apache/spark/streaming/examples/`. They can be run in the similar manner using `./run-example org.apache.spark.streaming.examples....` . Executing without any parameter would give the required parameter list. Further explanation to run them can be found in comments in the files. +You can find more examples in `/streaming/src/main/scala/org/apache/spark/streaming/examples/`. They can be run in the similar manner using `./bin/run-example org.apache.spark.streaming.examples....` . Executing without any parameter would give the required parameter list. Further explanation to run them can be found in comments in the files. # DStream Persistence Similar to RDDs, DStreams also allow developers to persist the stream's data in memory. That is, using `persist()` method on a DStream would automatically persist every RDD of that DStream in memory. This is useful if the data in the DStream will be computed multiple times (e.g., multiple operations on the same data). For window-based operations like `reduceByWindow` and `reduceByKeyAndWindow` and state-based operations like `updateStateByKey`, this is implicitly true. Hence, DStreams generated by window-based operations are automatically persisted in memory, without the developer calling `persist()`. diff --git a/examples/src/main/java/org/apache/spark/streaming/examples/JavaKafkaWordCount.java b/examples/src/main/java/org/apache/spark/streaming/examples/JavaKafkaWordCount.java index 22994fb2ec..75b588e4b8 100644 --- a/examples/src/main/java/org/apache/spark/streaming/examples/JavaKafkaWordCount.java +++ b/examples/src/main/java/org/apache/spark/streaming/examples/JavaKafkaWordCount.java @@ -41,7 +41,7 @@ import scala.Tuple2; * is the number of threads the kafka consumer should use * * Example: - * `./run-example org.apache.spark.streaming.examples.JavaKafkaWordCount local[2] zoo01,zoo02, + * `./bin/run-example org.apache.spark.streaming.examples.JavaKafkaWordCount local[2] zoo01,zoo02, * zoo03 my-consumer-group topic1,topic2 1` */ diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/ActorWordCount.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/ActorWordCount.scala index 2402409e6e..3641517934 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/ActorWordCount.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/ActorWordCount.scala @@ -134,9 +134,9 @@ object FeederActor { * and describe the AkkaSystem that Spark Sample feeder is running on. * * To run this example locally, you may run Feeder Actor as - * `$ ./run-example spark.streaming.examples.FeederActor 127.0.1.1 9999` + * `$ ./bin/run-example spark.streaming.examples.FeederActor 127.0.1.1 9999` * and then run the example - * `$ ./run-example spark.streaming.examples.ActorWordCount local[2] 127.0.1.1 9999` + * `$ ./bin/run-example spark.streaming.examples.ActorWordCount local[2] 127.0.1.1 9999` */ object ActorWordCount { def main(args: Array[String]) { diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/HdfsWordCount.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/HdfsWordCount.scala index bc8564b3ba..61be1ce4b1 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/HdfsWordCount.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/HdfsWordCount.scala @@ -28,7 +28,7 @@ import org.apache.spark.streaming.StreamingContext._ * is the directory that Spark Streaming will use to find and read new text files. * * To run this on your local machine on directory `localdir`, run this example - * `$ ./run-example spark.streaming.examples.HdfsWordCount local[2] localdir` + * `$ ./bin/run-example spark.streaming.examples.HdfsWordCount local[2] localdir` * Then create a text file in `localdir` and the words in the file will get counted. */ object HdfsWordCount { diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/KafkaWordCount.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/KafkaWordCount.scala index 570ba4c81a..8dc8a3531a 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/KafkaWordCount.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/KafkaWordCount.scala @@ -35,7 +35,7 @@ import org.apache.spark.streaming.util.RawTextHelper._ * is the number of threads the kafka consumer should use * * Example: - * `./run-example spark.streaming.examples.KafkaWordCount local[2] zoo01,zoo02,zoo03 my-consumer-group topic1,topic2 1` + * `./bin/run-example spark.streaming.examples.KafkaWordCount local[2] zoo01,zoo02,zoo03 my-consumer-group topic1,topic2 1` */ object KafkaWordCount { def main(args: Array[String]) { diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/MQTTWordCount.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/MQTTWordCount.scala index ff332a0282..ea138f55e8 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/MQTTWordCount.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/MQTTWordCount.scala @@ -79,9 +79,9 @@ object MQTTPublisher { * and describe where Mqtt publisher is running. * * To run this example locally, you may run publisher as - * `$ ./run-example org.apache.spark.streaming.examples.MQTTPublisher tcp://localhost:1883 foo` + * `$ ./bin/run-example org.apache.spark.streaming.examples.MQTTPublisher tcp://localhost:1883 foo` * and run the example as - * `$ ./run-example org.apache.spark.streaming.examples.MQTTWordCount local[2] tcp://localhost:1883 foo` + * `$ ./bin/run-example org.apache.spark.streaming.examples.MQTTWordCount local[2] tcp://localhost:1883 foo` */ object MQTTWordCount { diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/NetworkWordCount.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/NetworkWordCount.scala index e2487dca5f..ce8df8c502 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/NetworkWordCount.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/NetworkWordCount.scala @@ -29,7 +29,7 @@ import org.apache.spark.streaming.StreamingContext._ * To run this on your local machine, you need to first run a Netcat server * `$ nc -lk 9999` * and then run the example - * `$ ./run-example spark.streaming.examples.NetworkWordCount local[2] localhost 9999` + * `$ ./bin/run-example spark.streaming.examples.NetworkWordCount local[2] localhost 9999` */ object NetworkWordCount { def main(args: Array[String]) { diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/StatefulNetworkWordCount.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/StatefulNetworkWordCount.scala index cb30c4edb3..e55d71edfc 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/StatefulNetworkWordCount.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/StatefulNetworkWordCount.scala @@ -29,7 +29,7 @@ import org.apache.spark.streaming.StreamingContext._ * To run this on your local machine, you need to first run a Netcat server * `$ nc -lk 9999` * and then run the example - * `$ ./run-example spark.streaming.examples.StatefulNetworkWordCount local[2] localhost 9999` + * `$ ./bin/run-example spark.streaming.examples.StatefulNetworkWordCount local[2] localhost 9999` */ object StatefulNetworkWordCount { def main(args: Array[String]) { diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/ZeroMQWordCount.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/ZeroMQWordCount.scala index e83ce78aa5..2948aa7cc4 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/ZeroMQWordCount.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/ZeroMQWordCount.scala @@ -62,9 +62,9 @@ object SimpleZeroMQPublisher { * and describe where zeroMq publisher is running. * * To run this example locally, you may run publisher as - * `$ ./run-example spark.streaming.examples.SimpleZeroMQPublisher tcp://127.0.1.1:1234 foo.bar` + * `$ ./bin/run-example spark.streaming.examples.SimpleZeroMQPublisher tcp://127.0.1.1:1234 foo.bar` * and run the example as - * `$ ./run-example spark.streaming.examples.ZeroMQWordCount local[2] tcp://127.0.1.1:1234 foo` + * `$ ./bin/run-example spark.streaming.examples.ZeroMQWordCount local[2] tcp://127.0.1.1:1234 foo` */ object ZeroMQWordCount { def main(args: Array[String]) { diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/clickstream/PageViewGenerator.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/clickstream/PageViewGenerator.scala index de70c50473..1a40fdb9a3 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/clickstream/PageViewGenerator.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/clickstream/PageViewGenerator.scala @@ -39,8 +39,8 @@ object PageView extends Serializable { /** Generates streaming events to simulate page views on a website. * * This should be used in tandem with PageViewStream.scala. Example: - * $ ./run-example spark.streaming.examples.clickstream.PageViewGenerator 44444 10 - * $ ./run-example spark.streaming.examples.clickstream.PageViewStream errorRatePerZipCode localhost 44444 + * $ ./bin/run-example spark.streaming.examples.clickstream.PageViewGenerator 44444 10 + * $ ./bin/run-example spark.streaming.examples.clickstream.PageViewStream errorRatePerZipCode localhost 44444 * * When running this, you may want to set the root logging level to ERROR in * conf/log4j.properties to reduce the verbosity of the output. diff --git a/examples/src/main/scala/org/apache/spark/streaming/examples/clickstream/PageViewStream.scala b/examples/src/main/scala/org/apache/spark/streaming/examples/clickstream/PageViewStream.scala index 8282cc9269..968b578487 100644 --- a/examples/src/main/scala/org/apache/spark/streaming/examples/clickstream/PageViewStream.scala +++ b/examples/src/main/scala/org/apache/spark/streaming/examples/clickstream/PageViewStream.scala @@ -25,8 +25,8 @@ import org.apache.spark.SparkContext._ * operators available in Spark streaming. * * This should be used in tandem with PageViewStream.scala. Example: - * $ ./run-example spark.streaming.examples.clickstream.PageViewGenerator 44444 10 - * $ ./run-example spark.streaming.examples.clickstream.PageViewStream errorRatePerZipCode localhost 44444 + * $ ./bin/run-example spark.streaming.examples.clickstream.PageViewGenerator 44444 10 + * $ ./bin/run-example spark.streaming.examples.clickstream.PageViewStream errorRatePerZipCode localhost 44444 */ object PageViewStream { def main(args: Array[String]) { -- cgit v1.2.3 From a3f90a2ecf14a01aa27fc95c133b1ff375583adb Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Thu, 2 Jan 2014 18:50:12 +0530 Subject: pyspark -> bin/pyspark --- README.md | 2 +- docs/index.md | 2 +- docs/python-programming-guide.md | 28 ++++++++++++++-------------- docs/quick-start.md | 4 ++-- python/pyspark/shell.py | 2 +- 5 files changed, 19 insertions(+), 19 deletions(-) (limited to 'docs') diff --git a/README.md b/README.md index 7154165ab1..f7ad26de5b 100644 --- a/README.md +++ b/README.md @@ -21,7 +21,7 @@ Once you've built Spark, the easiest way to start using it is the shell: ./bin/spark-shell -Or, for the Python API, the Python shell (`./pyspark`). +Or, for the Python API, the Python shell (`./bin/pyspark`). Spark also comes with several sample programs in the `examples` directory. To run one of them, use `./bin/run-example `. For example: diff --git a/docs/index.md b/docs/index.md index f0990fff73..86d574daaa 100644 --- a/docs/index.md +++ b/docs/index.md @@ -35,7 +35,7 @@ or `local` to run locally with one thread, or `local[N]` to run locally with N t `local` for testing. Finally, you can run Spark interactively through modified versions of the Scala shell (`./bin/spark-shell`) or -Python interpreter (`./pyspark`). These are a great way to learn the framework. +Python interpreter (`./bin/pyspark`). These are a great way to learn the framework. # Launching on a Cluster diff --git a/docs/python-programming-guide.md b/docs/python-programming-guide.md index 96f93e24fe..dc187b3efe 100644 --- a/docs/python-programming-guide.md +++ b/docs/python-programming-guide.md @@ -47,7 +47,7 @@ PySpark will automatically ship these functions to workers, along with any objec Instances of classes will be serialized and shipped to workers by PySpark, but classes themselves cannot be automatically distributed to workers. The [Standalone Use](#standalone-use) section describes how to ship code dependencies to workers. -In addition, PySpark fully supports interactive use---simply run `./pyspark` to launch an interactive shell. +In addition, PySpark fully supports interactive use---simply run `./bin/pyspark` to launch an interactive shell. # Installing and Configuring PySpark @@ -60,17 +60,17 @@ By default, PySpark requires `python` to be available on the system `PATH` and u All of PySpark's library dependencies, including [Py4J](http://py4j.sourceforge.net/), are bundled with PySpark and automatically imported. -Standalone PySpark applications should be run using the `pyspark` script, which automatically configures the Java and Python environment using the settings in `conf/spark-env.sh` or `.cmd`. -The script automatically adds the `pyspark` package to the `PYTHONPATH`. +Standalone PySpark applications should be run using the `bin/pyspark` script, which automatically configures the Java and Python environment using the settings in `conf/spark-env.sh` or `.cmd`. +The script automatically adds the `bin/pyspark` package to the `PYTHONPATH`. # Interactive Use -The `pyspark` script launches a Python interpreter that is configured to run PySpark applications. To use `pyspark` interactively, first build Spark, then launch it directly from the command line without any options: +The `bin/pyspark` script launches a Python interpreter that is configured to run PySpark applications. To use `pyspark` interactively, first build Spark, then launch it directly from the command line without any options: {% highlight bash %} $ sbt/sbt assembly -$ ./pyspark +$ ./bin/pyspark {% endhighlight %} The Python shell can be used explore data interactively and is a simple way to learn the API: @@ -82,35 +82,35 @@ The Python shell can be used explore data interactively and is a simple way to l >>> help(pyspark) # Show all pyspark functions {% endhighlight %} -By default, the `pyspark` shell creates SparkContext that runs applications locally on a single core. +By default, the `bin/pyspark` shell creates SparkContext that runs applications locally on a single core. To connect to a non-local cluster, or use multiple cores, set the `MASTER` environment variable. -For example, to use the `pyspark` shell with a [standalone Spark cluster](spark-standalone.html): +For example, to use the `bin/pyspark` shell with a [standalone Spark cluster](spark-standalone.html): {% highlight bash %} -$ MASTER=spark://IP:PORT ./pyspark +$ MASTER=spark://IP:PORT ./bin/pyspark {% endhighlight %} Or, to use four cores on the local machine: {% highlight bash %} -$ MASTER=local[4] ./pyspark +$ MASTER=local[4] ./bin/pyspark {% endhighlight %} ## IPython It is also possible to launch PySpark in [IPython](http://ipython.org), the enhanced Python interpreter. -To do this, set the `IPYTHON` variable to `1` when running `pyspark`: +To do this, set the `IPYTHON` variable to `1` when running `bin/pyspark`: {% highlight bash %} -$ IPYTHON=1 ./pyspark +$ IPYTHON=1 ./bin/pyspark {% endhighlight %} Alternatively, you can customize the `ipython` command by setting `IPYTHON_OPTS`. For example, to launch the [IPython Notebook](http://ipython.org/notebook.html) with PyLab graphing support: {% highlight bash %} -$ IPYTHON_OPTS="notebook --pylab inline" ./pyspark +$ IPYTHON_OPTS="notebook --pylab inline" ./bin/pyspark {% endhighlight %} IPython also works on a cluster or on multiple cores if you set the `MASTER` environment variable. @@ -118,7 +118,7 @@ IPython also works on a cluster or on multiple cores if you set the `MASTER` env # Standalone Programs -PySpark can also be used from standalone Python scripts by creating a SparkContext in your script and running the script using `pyspark`. +PySpark can also be used from standalone Python scripts by creating a SparkContext in your script and running the script using `bin/pyspark`. The Quick Start guide includes a [complete example](quick-start.html#a-standalone-app-in-python) of a standalone Python application. Code dependencies can be deployed by listing them in the `pyFiles` option in the SparkContext constructor: @@ -153,6 +153,6 @@ Many of the methods also contain [doctests](http://docs.python.org/2/library/doc PySpark also includes several sample programs in the [`python/examples` folder](https://github.com/apache/incubator-spark/tree/master/python/examples). You can run them by passing the files to `pyspark`; e.g.: - ./pyspark python/examples/wordcount.py + ./bin/pyspark python/examples/wordcount.py Each program prints usage help when run without arguments. diff --git a/docs/quick-start.md b/docs/quick-start.md index 912f332236..2fa2bd718b 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -277,11 +277,11 @@ We can pass Python functions to Spark, which are automatically serialized along For applications that use custom classes or third-party libraries, we can add those code dependencies to SparkContext to ensure that they will be available on remote machines; this is described in more detail in the [Python programming guide](python-programming-guide.html). `SimpleApp` is simple enough that we do not need to specify any code dependencies. -We can run this application using the `pyspark` script: +We can run this application using the `bin/pyspark` script: {% highlight python %} $ cd $SPARK_HOME -$ ./pyspark SimpleApp.py +$ ./bin/pyspark SimpleApp.py ... Lines with a: 46, Lines with b: 23 {% endhighlight python %} diff --git a/python/pyspark/shell.py b/python/pyspark/shell.py index ef07eb437b..1602227a27 100644 --- a/python/pyspark/shell.py +++ b/python/pyspark/shell.py @@ -47,7 +47,7 @@ print "Spark context available as sc." if add_files != None: print "Adding files: [%s]" % ", ".join(add_files) -# The ./pyspark script stores the old PYTHONSTARTUP value in OLD_PYTHONSTARTUP, +# The ./bin/pyspark script stores the old PYTHONSTARTUP value in OLD_PYTHONSTARTUP, # which allows us to execute the user's PYTHONSTARTUP file: _pythonstartup = os.environ.get('OLD_PYTHONSTARTUP') if _pythonstartup and os.path.isfile(_pythonstartup): -- cgit v1.2.3 From 59e8009b8d5e51b6f776720de8c9ecb09e1072dc Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Thu, 2 Jan 2014 18:54:08 +0530 Subject: a few left over document change --- docs/spark-standalone.md | 2 +- ec2/spark_ec2.py | 2 +- make-distribution.sh | 4 ++-- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'docs') diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index ec5ae5b4f7..baa0a062f7 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -20,7 +20,7 @@ then modify `conf/spark-env.sh` in the `dist/` directory before deploying to all You can start a standalone master server by executing: - ./bin/start-master.sh + ./sbin/start-master.sh Once started, the master will print out a `spark://HOST:PORT` URL for itself, which you can use to connect workers to it, or pass as the "master" argument to `SparkContext`. You can also find this URL on diff --git a/ec2/spark_ec2.py b/ec2/spark_ec2.py index a2b0e7e7f4..d82a1e1490 100755 --- a/ec2/spark_ec2.py +++ b/ec2/spark_ec2.py @@ -436,7 +436,7 @@ def setup_cluster(conn, master_nodes, slave_nodes, opts, deploy_ssh_key): def setup_standalone_cluster(master, slave_nodes, opts): slave_ips = '\n'.join([i.public_dns_name for i in slave_nodes]) ssh(master, opts, "echo \"%s\" > spark/conf/slaves" % (slave_ips)) - ssh(master, opts, "/root/spark/bin/start-all.sh") + ssh(master, opts, "/root/spark/sbin/start-all.sh") def setup_spark_cluster(master, opts): ssh(master, opts, "chmod u+x spark-ec2/setup.sh") diff --git a/make-distribution.sh b/make-distribution.sh index 8765c7e620..0463d14762 100755 --- a/make-distribution.sh +++ b/make-distribution.sh @@ -31,9 +31,9 @@ # # Recommended deploy/testing procedure (standalone mode): # 1) Rsync / deploy the dist/ dir to one host -# 2) cd to deploy dir; ./bin/start-master.sh +# 2) cd to deploy dir; ./sbin/start-master.sh # 3) Verify master is up by visiting web page, ie http://master-ip:8080. Note the spark:// URL. -# 4) ./bin/start-slave.sh 1 <> +# 4) ./sbin/start-slave.sh 1 <> # 5) MASTER="spark://my-master-ip:7077" ./bin/spark-shell # -- cgit v1.2.3 From c189c8362caeaa7a0f46af1c8e0d8d37fd171d7b Mon Sep 17 00:00:00 2001 From: Hossein Falaki Date: Thu, 2 Jan 2014 15:22:20 -0800 Subject: Added Scala and Python examples for mllib --- docs/mllib-guide.md | 313 +++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 261 insertions(+), 52 deletions(-) (limited to 'docs') diff --git a/docs/mllib-guide.md b/docs/mllib-guide.md index c1ff9c417c..e9d3785427 100644 --- a/docs/mllib-guide.md +++ b/docs/mllib-guide.md @@ -39,56 +39,9 @@ underlying gradient descent primitive (described parameter (*regParam*) along with various parameters associated with gradient descent (*stepSize*, *numIterations*, *miniBatchFraction*). -The following code snippet illustrates how to load a sample dataset, execute a -training algorithm on this training data using a static method in the algorithm -object, and make predictions with the resulting model to compute the training -error. - -{% highlight scala %} -import org.apache.spark.SparkContext -import org.apache.spark.mllib.classification.SVMWithSGD -import org.apache.spark.mllib.regression.LabeledPoint - -// Load and parse the data file -val data = sc.textFile("mllib/data/sample_svm_data.txt") -val parsedData = data.map { line => - val parts = line.split(' ') - LabeledPoint(parts(0).toDouble, parts.tail.map(x => x.toDouble).toArray) -} - -// Run training algorithm -val numIterations = 20 -val model = SVMWithSGD.train(parsedData, numIterations) - -// Evaluate model on training examples and compute training error -val labelAndPreds = parsedData.map { point => - val prediction = model.predict(point.features) - (point.label, prediction) -} -val trainErr = labelAndPreds.filter(r => r._1 != r._2).count.toDouble / parsedData.count -println("trainError = " + trainErr) -{% endhighlight %} - -The `SVMWithSGD.train()` method by default performs L2 regularization with the -regularization parameter set to 1.0. If we want to configure this algorithm, we -can customize `SVMWithSGD` further by creating a new object directly and -calling setter methods. All other MLlib algorithms support customization in -this way as well. For example, the following code produces an L1 regularized -variant of SVMs with regularization parameter set to 0.1, and runs the training -algorithm for 200 iterations. -{% highlight scala %} -import org.apache.spark.mllib.optimization.L1Updater -val svmAlg = new SVMWithSGD() -svmAlg.optimizer.setNumIterations(200) - .setRegParam(0.1) - .setUpdater(new L1Updater) -val modelL1 = svmAlg.run(parsedData) -{% endhighlight %} -Both of the code snippets above can be executed in `spark-shell` to generate a -classifier for the provided dataset. Available algorithms for binary classification: @@ -121,14 +74,14 @@ of entities with one another based on some notion of similarity. Clustering is often used for exploratory analysis and/or as a component of a hierarchical supervised learning pipeline (in which distinct classifiers or regression models are trained for each cluster). MLlib supports -[k-means](http://en.wikipedia.org/wiki/K-means_clustering) clustering, arguably -the most commonly used clustering approach that clusters the data points into -*k* clusters. The MLlib implementation includes a parallelized +[k-means](http://en.wikipedia.org/wiki/K-means_clustering) clustering, one of +the most commonly used clustering algorithms that clusters the data points into +predfined number of clusters. The MLlib implementation includes a parallelized variant of the [k-means++](http://en.wikipedia.org/wiki/K-means%2B%2B) method called [kmeans||](http://theory.stanford.edu/~sergei/papers/vldb12-kmpar.pdf). The implementation in MLlib has the following parameters: -* *k* is the number of clusters. +* *k* is the number of desired clusters. * *maxIterations* is the maximum number of iterations to run. * *initializationMode* specifies either random initialization or initialization via k-means\|\|. @@ -169,7 +122,7 @@ the entries in the user-item matrix as *explicit* preferences given by the user It is common in many real-world use cases to only have access to *implicit feedback* (e.g. views, clicks, purchases, likes, shares etc.). The approach used in MLlib to deal with such data is taken from -[Collaborative Filtering for Implicit Feedback Datasets](http://research.yahoo.com/pub/2433). +[Collaborative Filtering for Implicit Feedback Datasets](http://www2.research.att.com/~yifanhu/PUB/cf.pdf). Essentially instead of trying to model the matrix of ratings directly, this approach treats the data as a combination of binary preferences and *confidence values*. The ratings are then related to the level of confidence in observed user preferences, rather than explicit ratings given to items. @@ -210,3 +163,259 @@ at each iteration. Available algorithms for gradient descent: * [GradientDescent](api/mllib/index.html#org.apache.spark.mllib.optimization.GradientDescent) + +# Using MLLib in Scala + +Following code snippets can be executed in `spark-shell`. + +## Binary Classification + +The following code snippet illustrates how to load a sample dataset, execute a +training algorithm on this training data using a static method in the algorithm +object, and make predictions with the resulting model to compute the training +error. + +{% highlight scala %} +import org.apache.spark.SparkContext +import org.apache.spark.mllib.classification.SVMWithSGD +import org.apache.spark.mllib.regression.LabeledPoint + +// Load and parse the data file +val data = sc.textFile("mllib/data/sample_svm_data.txt") +val parsedData = data.map { line => + val parts = line.split(' ') + LabeledPoint(parts(0).toDouble, parts.tail.map(x => x.toDouble).toArray) +} + +// Run training algorithm to build the model +val numIterations = 20 +val model = SVMWithSGD.train(parsedData, numIterations) + +// Evaluate model on training examples and compute training error +val labelAndPreds = parsedData.map { point => + val prediction = model.predict(point.features) + (point.label, prediction) +} +val trainErr = labelAndPreds.filter(r => r._1 != r._2).count.toDouble / parsedData.count +println("Training Error = " + trainErr) +{% endhighlight %} + + +The `SVMWithSGD.train()` method by default performs L2 regularization with the +regularization parameter set to 1.0. If we want to configure this algorithm, we +can customize `SVMWithSGD` further by creating a new object directly and +calling setter methods. All other MLlib algorithms support customization in +this way as well. For example, the following code produces an L1 regularized +variant of SVMs with regularization parameter set to 0.1, and runs the training +algorithm for 200 iterations. + +{% highlight scala %} +import org.apache.spark.mllib.optimization.L1Updater + +val svmAlg = new SVMWithSGD() +svmAlg.optimizer.setNumIterations(200) + .setRegParam(0.1) + .setUpdater(new L1Updater) +val modelL1 = svmAlg.run(parsedData) +{% endhighlight %} + +## Linear Regression +The following example demonstrate how to load training data, parse it as an RDD of LabeledPoint. The +example then uses LinearRegressionWithSGD to build a simple linear model to predict label values. We +compute the Mean Squared Error at the end to evaluate +[goodness of fit](http://en.wikipedia.org/wiki/Goodness_of_fit) + +{% highlight scala %} +import org.apache.spark.mllib.regression.LinearRegressionWithSGD +import org.apache.spark.mllib.regression.LabeledPoint + +// Load and parse the data +val data = sc.textFile("mllib/data/ridge-data/lpsa.data") +val parsedData = data.map { line => + val parts = line.split(',') + LabeledPoint(parts(0).toDouble, parts(1).split(' ').map(x => x.toDouble).toArray) +} + +// Building the model +val numIterations = 20 +val model = LinearRegressionWithSGD.train(parsedData, numIterations) + +// Evaluate model on training examples and compute training error +val valuesAndPreds = parsedData.map { point => + val prediction = model.predict(point.features) + (point.label, prediction) +} +val MSE = valuesAndPreds.map{ case(v, p) => math.pow((v - p), 2)}.reduce(_ + _)/valuesAndPreds.count +println("training Mean Squared Error = " + MSE) +{% endhighlight %} + + +Similarly you can use RidgeRegressionWithSGD and LassoWithSGD and compare training +[Mean Squared Errors](http://en.wikipedia.org/wiki/Mean_squared_error). + +## Clustering +In the following example after loading and parsing data, we use the KMeans object to cluster the data +into two clusters. The number of desired clusters is passed to the algorithm. We then compute Within +Set Sum of Squared Error (WSSSE). You can reduce this error measure by increasing *k*. In fact the +optimal *k* is usually one where there is an "elbow" in the WSSSE graph. + +{% highlight scala %} +import org.apache.spark.mllib.clustering.KMeans + +// Load and parse the data +val data = sc.textFile("kmeans_data.txt") +val parsedData = data.map( _.split(' ').map(_.toDouble)) + +// Cluster the data into two classes using KMeans +val numIterations = 20 +val numClusters = 2 +val clusters = KMeans.train(parsedData, numClusters, numIterations) + +// Evaluate clustering by computing Within Set Sum of Squared Errors +val WSSSE = clusters.computeCost(parsedData) +println("Within Set Sum of Squared Errors = " + WSSSE) +{% endhighlight %} + + +## Collaborative Filtering +In the following example we load rating data. Each row consists of a user, a product and a rating. +We use the default ALS.train() method which assumes ratings are explicit. We evaluate the recommendation +model by measuring the Mean Squared Error of rating prediction. + +{% highlight scala %} +import org.apache.spark.mllib.recommendation.ALS +import org.apache.spark.mllib.recommendation.Rating + +// Load and parse the data +val data = sc.textFile("mllib/data/als/test.data") +val ratings = data.map(_.split(',') match { + case Array(user, item, rate) => Rating(user.toInt, item.toInt, rate.toDouble) +}) + +// Build the recommendation model using ALS +val numIterations = 20 +val model = ALS.train(ratings, 1, 20, 0.01) + +// Evaluate the model on rating data +val ratesAndPreds = ratings.map{ case Rating(user, item, rate) => (rate, model.predict(user, item))} +val MSE = ratesAndPreds.map{ case(v, p) => math.pow((v - p), 2)}.reduce(_ + _)/ratesAndPreds.count +{% endhighlight %} + +If the rating matrix is derived from other source of information (i.e., it is inferred from +other signals), you can use the trainImplicit method to get better results. + +{% highlight scala %} +val model = ALS.trainImplicit(ratings, 1, 20, 0.01) +{% endhighlight %} + +# Using MLLib in Python +Following examples can be tested in the PySpark shell. + +## Binary Classification +The following example shows how to load a sample dataset, build Logistic Regression model, +and make predictions with the resulting model to compute the training error. + +{% highlight python %} +from pyspark.mllib.classification import LogisticRegressionWithSGD +from numpy import array + +# Load and parse the data +data = sc.textFile("mllib/data/sample_svm_data.txt") +parsedData = data.map(lambda line: array([float(x) for x in line.split(' ')])) +model = LogisticRegressionWithSGD.train(sc, parsedData) + +# Build the model +labelsAndPreds = parsedData.map(lambda point: (int(point.item(0)), + model.predict(point.take(range(1, point.size))))) + +# Evaluating the model on training data +trainErr = labelsAndPreds.filter(lambda (v, p): v != p).count() / float(parsedData.count()) +print("Training Error = " + str(trainErr)) +{% endhighlight %} + +## Linear Regression +The following example demonstrate how to load training data, parse it as an RDD of LabeledPoint. The +example then uses LinearRegressionWithSGD to build a simple linear model to predict label values. We +compute the Mean Squared Error at the end to evaluate +[goodness of fit](http://en.wikipedia.org/wiki/Goodness_of_fit) + +{% highlight python %} +from pyspark.mllib.regression import LinearRegressionWithSGD +from numpy import array + +# Load and parse the data +data = sc.textFile("mllib/data/ridge-data/lpsa.data") +parsedData = data.map(lambda line: array([float(x) for x in line.replace(',', ' ').split(' ')])) + +# Build the model +model = LinearRegressionWithSGD.train(sc, parsedData) + +# Evaluate the model on training data +valuesAndPreds = parsedData.map(lambda point: (point.item(0), + model.predict(point.take(range(1, point.size))))) +MSE = valuesAndPreds.map(lambda (v, p): (v - p)**2).reduce(lambda x, y: x + y)/valuesAndPreds.count() +print("Mean Squared Error = " + str(MSE)) +{% endhighlight %} + + +## Clustering +In the following example after loading and parsing data, we use the KMeans object to cluster the data +into two clusters. The number of desired clusters is passed to the algorithm. We then compute Within +Set Sum of Squared Error (WSSSE). You can reduce this error measure by increasing *k*. In fact the +optimal *k* is usually one where there is an "elbow" in the WSSSE graph. + +{% highlight python %} +from pyspark.mllib.clustering import KMeans +from numpy import array +from math import sqrt + +# Load and parse the data +data = sc.textFile("kmeans_data.txt") +parsedData = data.map(lambda line: array([float(x) for x in line.split(' ')])) + +# Build the model (cluster the data) +clusters = KMeans.train(sc, parsedData, 2, maxIterations=10, + runs=30, initialization_mode="random") + +# Evaluate clustering by computing Within Set Sum of Squared Errors +def error(point): + center = clusters.centers[clusters.predict(point)] + return sqrt(sum([x**2 for x in (point - center)])) + +WSSSE = parsedData.map(lambda point: error(point)).reduce(lambda x, y: x + y) +print("Within Set Sum of Squared Error = " + str(WSSSE)) +{% endhighlight %} + +Similarly you can use RidgeRegressionWithSGD and LassoWithSGD and compare training Mean Squared +Errors. + +## Collaborative Filtering +In the following example we load rating data. Each row consists of a user, a product and a rating. +We use the default ALS.train() method which assumes ratings are explicit. We evaluate the recommendation +model by measuring the Mean Squared Error of rating prediction. + +{% highlight python %} +from pyspark.mllib.recommendation import ALS +from numpy import array + +# Load and parse the data +data = sc.textFile("mllib/data/als/test.data") +ratings = data.map(lambda line: array([float(x) for x in line.split(',')])) + +# Build the recommendation model using Alternating Least Squares +model = ALS.train(sc, ratings, 1, 20) + +# Evaluate the model on training data +ratesAndPreds = ratings.map(lambda p: (p[2], model.predict(int(p[0]), int(p[1])))) +MSE = valuesAndPreds.map(lambda (v, p): (v - p)**2).reduce(lambda x, y: x + y)/valuesAndPreds.count() +print("Mean Squared Error = " + str(MSE)) + +{% endhighlight %} + +If the rating matrix is derived from other source of information (i.e., it is inferred from other +signals), you can use the trainImplicit method to get better results. + +{% highlight python %} +# Build the recommendation model using Alternating Least Squares based on implicit ratings +model = ALS.trainImplicit(sc, ratings, 1, 20) +{% endhighlight %} \ No newline at end of file -- cgit v1.2.3 From 81989e26647ede54e19ef8058846e1bd42c0bfb5 Mon Sep 17 00:00:00 2001 From: Hossein Falaki Date: Thu, 2 Jan 2014 16:22:13 -0800 Subject: Commented the last part of collaborative filtering examples that lead to errors --- docs/mllib-guide.md | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'docs') diff --git a/docs/mllib-guide.md b/docs/mllib-guide.md index e9d3785427..0bebc41137 100644 --- a/docs/mllib-guide.md +++ b/docs/mllib-guide.md @@ -297,8 +297,9 @@ val numIterations = 20 val model = ALS.train(ratings, 1, 20, 0.01) // Evaluate the model on rating data -val ratesAndPreds = ratings.map{ case Rating(user, item, rate) => (rate, model.predict(user, item))} -val MSE = ratesAndPreds.map{ case(v, p) => math.pow((v - p), 2)}.reduce(_ + _)/ratesAndPreds.count +//val ratesAndPreds = ratings.map{ case Rating(user, item, rate) => (rate, model.predict(user, item))} +//val MSE = ratesAndPreds.map{ case(v, p) => math.pow((v - p), 2)}.reduce(_ + _)/ratesAndPreds.count +//println("Mean Squared Error = " + MSE) {% endhighlight %} If the rating matrix is derived from other source of information (i.e., it is inferred from @@ -406,9 +407,9 @@ ratings = data.map(lambda line: array([float(x) for x in line.split(',')])) model = ALS.train(sc, ratings, 1, 20) # Evaluate the model on training data -ratesAndPreds = ratings.map(lambda p: (p[2], model.predict(int(p[0]), int(p[1])))) -MSE = valuesAndPreds.map(lambda (v, p): (v - p)**2).reduce(lambda x, y: x + y)/valuesAndPreds.count() -print("Mean Squared Error = " + str(MSE)) +#ratesAndPreds = ratings.map(lambda p: (p[2], model.predict(int(p[0]), int(p[1])))) +#MSE = valuesAndPreds.map(lambda (v, p): (v - p)**2).reduce(lambda x, y: x + y)/valuesAndPreds.count() +#print("Mean Squared Error = " + str(MSE)) {% endhighlight %} -- cgit v1.2.3 From be343d2a567b8ba3824c5e186e046f96ff77e334 Mon Sep 17 00:00:00 2001 From: Raymond Liu Date: Mon, 9 Dec 2013 09:47:52 +0800 Subject: Fix yarn/README.md and update docs/running-on-yarn.md --- docs/running-on-yarn.md | 2 +- yarn/README.md | 2 -- 2 files changed, 1 insertion(+), 3 deletions(-) (limited to 'docs') diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index aa75ca4324..5f995ce51b 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -116,7 +116,7 @@ For example: # Building Spark for Hadoop/YARN 2.2.x -Hadoop 2.2.x users must build Spark and publish it locally. The SBT build process handles Hadoop 2.2.x as a special case. This version of Hadoop has new YARN API changes and depends on a Protobuf version (2.5) that is not compatible with the Akka version (2.0.5) that Spark uses. Therefore, if the Hadoop version (e.g. set through ```SPARK_HADOOP_VERSION```) starts with 2.2.0 or higher then the build process will depend on Akka artifacts distributed by the Spark project compatible with Protobuf 2.5. Furthermore, the build process then uses the directory ```new-yarn``` (instead of ```yarn```), which supports the new YARN API. The build process should seamlessly work out of the box. +Hadoop 2.2.x users must build Spark and publish it locally. The SBT build process handles Hadoop 2.2.x as a special case. This version of Hadoop has new YARN API changes and depends on a Protobuf version (2.5) that is not compatible with the Akka version (2.0.5) that Spark uses. Therefore, if the Hadoop version (e.g. set through ```SPARK_HADOOP_VERSION```) starts with 2.2.0 or higher then the build process will depend on Akka artifacts distributed by the Spark project compatible with Protobuf 2.5. See [Building Spark with Maven](building-with-maven.html) for instructions on how to build Spark using the Maven process. diff --git a/yarn/README.md b/yarn/README.md index 3639f2c364..e2c3aea9a8 100644 --- a/yarn/README.md +++ b/yarn/README.md @@ -1,7 +1,5 @@ # YARN DIRECTORY LAYOUT -Hadoop Yarn related code are organized - Hadoop Yarn related codes are organized in separate modules with layered dependency. * common : Common codes that will be called into by other modules. -- cgit v1.2.3 From 7815a3ace97848b1d4bab4ef6823fc6d55cfb72e Mon Sep 17 00:00:00 2001 From: Raymond Liu Date: Fri, 6 Dec 2013 10:36:59 +0800 Subject: Update maven build documentation --- docs/building-with-maven.md | 10 +++------- docs/running-on-yarn.md | 2 +- 2 files changed, 4 insertions(+), 8 deletions(-) (limited to 'docs') diff --git a/docs/building-with-maven.md b/docs/building-with-maven.md index c709001632..699b10b188 100644 --- a/docs/building-with-maven.md +++ b/docs/building-with-maven.md @@ -37,7 +37,7 @@ For Apache Hadoop versions 1.x, Cloudera CDH MRv1, and other Hadoop versions wit # Cloudera CDH 4.2.0 with MapReduce v1 $ mvn -Dhadoop.version=2.0.0-mr1-cdh4.2.0 -DskipTests clean package -For Apache Hadoop 2.x, 0.23.x, Cloudera CDH MRv2, and other Hadoop versions with YARN, you should enable the "hadoop2-yarn" profile and set the "yarn.version" property: +For Apache Hadoop 2.x, 0.23.x, Cloudera CDH MRv2, and other Hadoop versions with YARN, you should enable the "hadoop2-yarn" or "hadoop2.2-yarn" profile and set the "hadoop.version", "yarn.version" property: # Apache Hadoop 2.0.5-alpha $ mvn -Phadoop2-yarn -Dhadoop.version=2.0.5-alpha -Dyarn.version=2.0.5-alpha -DskipTests clean package @@ -45,12 +45,8 @@ For Apache Hadoop 2.x, 0.23.x, Cloudera CDH MRv2, and other Hadoop versions with # Cloudera CDH 4.2.0 with MapReduce v2 $ mvn -Phadoop2-yarn -Dhadoop.version=2.0.0-cdh4.2.0 -Dyarn.version=2.0.0-chd4.2.0 -DskipTests clean package -Hadoop versions 2.2.x and newer can be built by setting the ```new-yarn``` and the ```yarn.version``` as follows: - - # Apache Hadoop 2.2.X and newer - $ mvn -Dyarn.version=2.2.0 -Dhadoop.version=2.2.0 -Pnew-yarn - -The build process handles Hadoop 2.2.x as a special case that uses the directory ```new-yarn```, which supports the new YARN API. Furthermore, for this version, the build depends on artifacts published by the spark-project to enable Akka 2.0.5 to work with protobuf 2.5. + # Apache Hadoop 2.2.X ( e.g. 2.2.0 as below ) and newer + $ mvn -Phadoop2.2-yarn -Dhadoop.version=2.2.0 -Dyarn.version=2.2.0 -DskipTests clean package ## Spark Tests in Maven ## diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index 5f995ce51b..f47098554e 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -116,7 +116,7 @@ For example: # Building Spark for Hadoop/YARN 2.2.x -Hadoop 2.2.x users must build Spark and publish it locally. The SBT build process handles Hadoop 2.2.x as a special case. This version of Hadoop has new YARN API changes and depends on a Protobuf version (2.5) that is not compatible with the Akka version (2.0.5) that Spark uses. Therefore, if the Hadoop version (e.g. set through ```SPARK_HADOOP_VERSION```) starts with 2.2.0 or higher then the build process will depend on Akka artifacts distributed by the Spark project compatible with Protobuf 2.5. +Hadoop 2.2.x users must build Spark and publish it locally. The SBT build process handles Hadoop 2.2.x as a special case. This version of Hadoop has new YARN API changes and depends on a Protobuf version (2.5). See [Building Spark with Maven](building-with-maven.html) for instructions on how to build Spark using the Maven process. -- cgit v1.2.3 From ebdfa6bb9766209bc5a3c4241fa47141c5e9c5cb Mon Sep 17 00:00:00 2001 From: Raymond Liu Date: Thu, 2 Jan 2014 11:23:33 +0800 Subject: Using name yarn-alpha/yarn instead of yarn-2.0/yarn-2.2 --- assembly/pom.xml | 8 +- docs/building-with-maven.md | 8 +- pom.xml | 4 +- project/SparkBuild.scala | 16 +- yarn/2.0/pom.xml | 112 ---- .../spark/deploy/yarn/ApplicationMaster.scala | 458 -------------- .../org/apache/spark/deploy/yarn/Client.scala | 505 --------------- .../apache/spark/deploy/yarn/WorkerLauncher.scala | 248 -------- .../apache/spark/deploy/yarn/WorkerRunnable.scala | 235 ------- .../spark/deploy/yarn/YarnAllocationHandler.scala | 680 -------------------- yarn/2.2/pom.xml | 112 ---- .../spark/deploy/yarn/ApplicationMaster.scala | 428 ------------- .../org/apache/spark/deploy/yarn/Client.scala | 523 ---------------- .../apache/spark/deploy/yarn/WorkerLauncher.scala | 227 ------- .../apache/spark/deploy/yarn/WorkerRunnable.scala | 209 ------- .../spark/deploy/yarn/YarnAllocationHandler.scala | 694 --------------------- yarn/README.md | 8 +- yarn/alpha/pom.xml | 112 ++++ .../spark/deploy/yarn/ApplicationMaster.scala | 458 ++++++++++++++ .../org/apache/spark/deploy/yarn/Client.scala | 505 +++++++++++++++ .../apache/spark/deploy/yarn/WorkerLauncher.scala | 248 ++++++++ .../apache/spark/deploy/yarn/WorkerRunnable.scala | 235 +++++++ .../spark/deploy/yarn/YarnAllocationHandler.scala | 680 ++++++++++++++++++++ yarn/pom.xml | 8 +- yarn/stable/pom.xml | 112 ++++ .../spark/deploy/yarn/ApplicationMaster.scala | 428 +++++++++++++ .../org/apache/spark/deploy/yarn/Client.scala | 523 ++++++++++++++++ .../apache/spark/deploy/yarn/WorkerLauncher.scala | 227 +++++++ .../apache/spark/deploy/yarn/WorkerRunnable.scala | 209 +++++++ .../spark/deploy/yarn/YarnAllocationHandler.scala | 694 +++++++++++++++++++++ 30 files changed, 4457 insertions(+), 4457 deletions(-) delete mode 100644 yarn/2.0/pom.xml delete mode 100644 yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala delete mode 100644 yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/Client.scala delete mode 100644 yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala delete mode 100644 yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala delete mode 100644 yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala delete mode 100644 yarn/2.2/pom.xml delete mode 100644 yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala delete mode 100644 yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/Client.scala delete mode 100644 yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala delete mode 100644 yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala delete mode 100644 yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala create mode 100644 yarn/alpha/pom.xml create mode 100644 yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala create mode 100644 yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/Client.scala create mode 100644 yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala create mode 100644 yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala create mode 100644 yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala create mode 100644 yarn/stable/pom.xml create mode 100644 yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala create mode 100644 yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/Client.scala create mode 100644 yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala create mode 100644 yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala create mode 100644 yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala (limited to 'docs') diff --git a/assembly/pom.xml b/assembly/pom.xml index bd94834415..079509bec6 100644 --- a/assembly/pom.xml +++ b/assembly/pom.xml @@ -124,21 +124,21 @@ - hadoop2-yarn + yarn-alpha org.apache.spark - spark-yarn-2.0_${scala.binary.version} + spark-yarn-alpha_${scala.binary.version} ${project.version} - hadoop2.2-yarn + yarn org.apache.spark - spark-yarn-2.2_${scala.binary.version} + spark-yarn_${scala.binary.version} ${project.version} diff --git a/docs/building-with-maven.md b/docs/building-with-maven.md index 699b10b188..b9ff0af76f 100644 --- a/docs/building-with-maven.md +++ b/docs/building-with-maven.md @@ -37,16 +37,16 @@ For Apache Hadoop versions 1.x, Cloudera CDH MRv1, and other Hadoop versions wit # Cloudera CDH 4.2.0 with MapReduce v1 $ mvn -Dhadoop.version=2.0.0-mr1-cdh4.2.0 -DskipTests clean package -For Apache Hadoop 2.x, 0.23.x, Cloudera CDH MRv2, and other Hadoop versions with YARN, you should enable the "hadoop2-yarn" or "hadoop2.2-yarn" profile and set the "hadoop.version", "yarn.version" property: +For Apache Hadoop 2.x, 0.23.x, Cloudera CDH MRv2, and other Hadoop versions with YARN, you should enable the "yarn-alpha" or "yarn" profile and set the "hadoop.version", "yarn.version" property: # Apache Hadoop 2.0.5-alpha - $ mvn -Phadoop2-yarn -Dhadoop.version=2.0.5-alpha -Dyarn.version=2.0.5-alpha -DskipTests clean package + $ mvn -Pyarn-alpha -Dhadoop.version=2.0.5-alpha -Dyarn.version=2.0.5-alpha -DskipTests clean package # Cloudera CDH 4.2.0 with MapReduce v2 - $ mvn -Phadoop2-yarn -Dhadoop.version=2.0.0-cdh4.2.0 -Dyarn.version=2.0.0-chd4.2.0 -DskipTests clean package + $ mvn -Pyarn-alpha -Dhadoop.version=2.0.0-cdh4.2.0 -Dyarn.version=2.0.0-chd4.2.0 -DskipTests clean package # Apache Hadoop 2.2.X ( e.g. 2.2.0 as below ) and newer - $ mvn -Phadoop2.2-yarn -Dhadoop.version=2.2.0 -Dyarn.version=2.2.0 -DskipTests clean package + $ mvn -Pyarn -Dhadoop.version=2.2.0 -Dyarn.version=2.2.0 -DskipTests clean package ## Spark Tests in Maven ## diff --git a/pom.xml b/pom.xml index a2c1f6d53d..aa2f076aac 100644 --- a/pom.xml +++ b/pom.xml @@ -722,7 +722,7 @@ - hadoop2-yarn + yarn-alpha 2 @@ -738,7 +738,7 @@ - hadoop2.2-yarn + yarn 2 2.2.0 diff --git a/project/SparkBuild.scala b/project/SparkBuild.scala index bc9c02d96f..b0749cc9c4 100644 --- a/project/SparkBuild.scala +++ b/project/SparkBuild.scala @@ -85,11 +85,11 @@ object SparkBuild extends Build { } // Conditionally include the yarn sub-project - lazy val yarn20 = Project("yarn2-alpha", file("yarn/2.0"), settings = yarn20Settings) dependsOn(core) - lazy val yarn22 = Project("yarn2-stable", file("yarn/2.2"), settings = yarn22Settings) dependsOn(core) + lazy val yarnAlpha = Project("yarn-alpha", file("yarn/alpha"), settings = yarnAlphaSettings) dependsOn(core) + lazy val yarn = Project("yarn", file("yarn/stable"), settings = yarnSettings) dependsOn(core) - lazy val maybeYarn = if (isYarnEnabled) Seq[ClasspathDependency](if (isNewHadoop) yarn22 else yarn20) else Seq[ClasspathDependency]() - lazy val maybeYarnRef = if (isYarnEnabled) Seq[ProjectReference](if (isNewHadoop) yarn22 else yarn20) else Seq[ProjectReference]() + lazy val maybeYarn = if (isYarnEnabled) Seq[ClasspathDependency](if (isNewHadoop) yarn else yarnAlpha) else Seq[ClasspathDependency]() + lazy val maybeYarnRef = if (isYarnEnabled) Seq[ProjectReference](if (isNewHadoop) yarn else yarnAlpha) else Seq[ProjectReference]() // Everything except assembly, tools and examples belong to packageProjects lazy val packageProjects = Seq[ProjectReference](core, repl, bagel, streaming, mllib) ++ maybeYarnRef @@ -334,12 +334,12 @@ object SparkBuild extends Build { ) ++ extraYarnSettings - def yarn20Settings = yarnCommonSettings ++ Seq( - name := "spark-yarn-2.0" + def yarnAlphaSettings = yarnCommonSettings ++ Seq( + name := "spark-yarn-alpha" ) - def yarn22Settings = yarnCommonSettings ++ Seq( - name := "spark-yarn-2.2" + def yarnSettings = yarnCommonSettings ++ Seq( + name := "spark-yarn" ) // Conditionally include the YARN dependencies because some tools look at all sub-projects and will complain diff --git a/yarn/2.0/pom.xml b/yarn/2.0/pom.xml deleted file mode 100644 index ec6738fa38..0000000000 --- a/yarn/2.0/pom.xml +++ /dev/null @@ -1,112 +0,0 @@ - - - - 4.0.0 - - org.apache.spark - spark-yarn-aggregator_2.10 - 0.9.0-incubating-SNAPSHOT - ../pom.xml - - - org.apache.spark - spark-yarn-2.0_2.10 - jar - Spark Project YARN 2.0 - http://spark.incubator.apache.org/ - - - - - org.codehaus.mojo - build-helper-maven-plugin - - - add-scala-sources - generate-sources - - add-source - - - - src/main/scala - ../common/src/main/scala - - - - - add-scala-test-sources - generate-test-sources - - add-test-source - - - - src/test/scala - ../common/src/test/scala - - - - - - - org.apache.maven.plugins - maven-antrun-plugin - - - test - - run - - - true - - - - - - - - - - - - - - - - - - - - org.scalatest - scalatest-maven-plugin - - - ${basedir}/../.. - 1 - ${spark.classpath} - - - - - - target/scala-${scala.binary.version}/classes - target/scala-${scala.binary.version}/test-classes - - diff --git a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala b/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala deleted file mode 100644 index 7cf120d3eb..0000000000 --- a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala +++ /dev/null @@ -1,458 +0,0 @@ -/* - * 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. - */ - -package org.apache.spark.deploy.yarn - -import java.io.IOException -import java.net.Socket -import java.util.concurrent.CopyOnWriteArrayList -import java.util.concurrent.atomic.{AtomicInteger, AtomicReference} - -import scala.collection.JavaConversions._ - -import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.fs.{FileSystem, Path} -import org.apache.hadoop.net.NetUtils -import org.apache.hadoop.security.UserGroupInformation -import org.apache.hadoop.util.ShutdownHookManager -import org.apache.hadoop.yarn.api._ -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.{ConverterUtils, Records} - -import org.apache.spark.{SparkConf, SparkContext, Logging} -import org.apache.spark.util.Utils - -class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) extends Logging { - - def this(args: ApplicationMasterArguments) = this(args, new Configuration()) - - private var rpc: YarnRPC = YarnRPC.create(conf) - private var resourceManager: AMRMProtocol = _ - private var appAttemptId: ApplicationAttemptId = _ - private var userThread: Thread = _ - private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) - private val fs = FileSystem.get(yarnConf) - - private var yarnAllocator: YarnAllocationHandler = _ - private var isFinished: Boolean = false - private var uiAddress: String = _ - private val maxAppAttempts: Int = conf.getInt(YarnConfiguration.RM_AM_MAX_RETRIES, - YarnConfiguration.DEFAULT_RM_AM_MAX_RETRIES) - private var isLastAMRetry: Boolean = true - - private val sparkConf = new SparkConf() - // Default to numWorkers * 2, with minimum of 3 - private val maxNumWorkerFailures = sparkConf.getInt("spark.yarn.max.worker.failures", - math.max(args.numWorkers * 2, 3)) - - def run() { - // Setup the directories so things go to yarn approved directories rather - // then user specified and /tmp. - System.setProperty("spark.local.dir", getLocalDirs()) - - // set the web ui port to be ephemeral for yarn so we don't conflict with - // other spark processes running on the same box - System.setProperty("spark.ui.port", "0") - - // Use priority 30 as its higher then HDFS. Its same priority as MapReduce is using. - ShutdownHookManager.get().addShutdownHook(new AppMasterShutdownHook(this), 30) - - appAttemptId = getApplicationAttemptId() - isLastAMRetry = appAttemptId.getAttemptId() >= maxAppAttempts - resourceManager = registerWithResourceManager() - - // Workaround until hadoop moves to something which has - // https://issues.apache.org/jira/browse/HADOOP-8406 - fixed in (2.0.2-alpha but no 0.23 line) - // ignore result. - // This does not, unfortunately, always work reliably ... but alleviates the bug a lot of times - // Hence args.workerCores = numCore disabled above. Any better option? - - // Compute number of threads for akka - //val minimumMemory = appMasterResponse.getMinimumResourceCapability().getMemory() - //if (minimumMemory > 0) { - // val mem = args.workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD - // val numCore = (mem / minimumMemory) + (if (0 != (mem % minimumMemory)) 1 else 0) - - // if (numCore > 0) { - // do not override - hits https://issues.apache.org/jira/browse/HADOOP-8406 - // TODO: Uncomment when hadoop is on a version which has this fixed. - // args.workerCores = numCore - // } - //} - // org.apache.hadoop.io.compress.CompressionCodecFactory.getCodecClasses(conf) - - ApplicationMaster.register(this) - // Start the user's JAR - userThread = startUserClass() - - // This a bit hacky, but we need to wait until the spark.driver.port property has - // been set by the Thread executing the user class. - waitForSparkContextInitialized() - - // Do this after spark master is up and SparkContext is created so that we can register UI Url - val appMasterResponse: RegisterApplicationMasterResponse = registerApplicationMaster() - - // Allocate all containers - allocateWorkers() - - // Wait for the user class to Finish - userThread.join() - - System.exit(0) - } - - /** Get the Yarn approved local directories. */ - private def getLocalDirs(): String = { - // Hadoop 0.23 and 2.x have different Environment variable names for the - // local dirs, so lets check both. We assume one of the 2 is set. - // LOCAL_DIRS => 2.X, YARN_LOCAL_DIRS => 0.23.X - val localDirs = Option(System.getenv("YARN_LOCAL_DIRS")) - .getOrElse(Option(System.getenv("LOCAL_DIRS")) - .getOrElse("")) - - if (localDirs.isEmpty()) { - throw new Exception("Yarn Local dirs can't be empty") - } - localDirs - } - - private def getApplicationAttemptId(): ApplicationAttemptId = { - val envs = System.getenv() - val containerIdString = envs.get(ApplicationConstants.AM_CONTAINER_ID_ENV) - val containerId = ConverterUtils.toContainerId(containerIdString) - val appAttemptId = containerId.getApplicationAttemptId() - logInfo("ApplicationAttemptId: " + appAttemptId) - appAttemptId - } - - private def registerWithResourceManager(): AMRMProtocol = { - val rmAddress = NetUtils.createSocketAddr(yarnConf.get( - YarnConfiguration.RM_SCHEDULER_ADDRESS, - YarnConfiguration.DEFAULT_RM_SCHEDULER_ADDRESS)) - logInfo("Connecting to ResourceManager at " + rmAddress) - rpc.getProxy(classOf[AMRMProtocol], rmAddress, conf).asInstanceOf[AMRMProtocol] - } - - private def registerApplicationMaster(): RegisterApplicationMasterResponse = { - logInfo("Registering the ApplicationMaster") - val appMasterRequest = Records.newRecord(classOf[RegisterApplicationMasterRequest]) - .asInstanceOf[RegisterApplicationMasterRequest] - appMasterRequest.setApplicationAttemptId(appAttemptId) - // Setting this to master host,port - so that the ApplicationReport at client has some - // sensible info. - // Users can then monitor stderr/stdout on that node if required. - appMasterRequest.setHost(Utils.localHostName()) - appMasterRequest.setRpcPort(0) - appMasterRequest.setTrackingUrl(uiAddress) - resourceManager.registerApplicationMaster(appMasterRequest) - } - - private def startUserClass(): Thread = { - logInfo("Starting the user JAR in a separate Thread") - val mainMethod = Class.forName( - args.userClass, - false /* initialize */, - Thread.currentThread.getContextClassLoader).getMethod("main", classOf[Array[String]]) - val t = new Thread { - override def run() { - 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 { - logDebug("finishing main") - isLastAMRetry = true - if (successed) { - ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) - } else { - ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.FAILED) - } - } - } - } - t.start() - t - } - - // this need to happen before allocateWorkers - private def waitForSparkContextInitialized() { - logInfo("Waiting for spark context initialization") - try { - var sparkContext: SparkContext = null - ApplicationMaster.sparkContextRef.synchronized { - var count = 0 - val waitTime = 10000L - val numTries = sparkConf.getInt("spark.yarn.ApplicationMaster.waitTries", 10) - while (ApplicationMaster.sparkContextRef.get() == null && count < numTries) { - logInfo("Waiting for spark context initialization ... " + count) - count = count + 1 - ApplicationMaster.sparkContextRef.wait(waitTime) - } - sparkContext = ApplicationMaster.sparkContextRef.get() - assert(sparkContext != null || count >= numTries) - - if (null != sparkContext) { - uiAddress = sparkContext.ui.appUIAddress - this.yarnAllocator = YarnAllocationHandler.newAllocator( - yarnConf, - resourceManager, - appAttemptId, - args, - sparkContext.preferredNodeLocationData, - sparkContext.getConf) - } else { - logWarning("Unable to retrieve sparkContext inspite of waiting for %d, numTries = %d". - format(count * waitTime, numTries)) - this.yarnAllocator = YarnAllocationHandler.newAllocator( - yarnConf, - resourceManager, - appAttemptId, - args, - sparkContext.getConf) - } - } - } finally { - // in case of exceptions, etc - ensure that count is atleast ALLOCATOR_LOOP_WAIT_COUNT : - // so that the loop (in ApplicationMaster.sparkContextInitialized) breaks - ApplicationMaster.incrementAllocatorLoop(ApplicationMaster.ALLOCATOR_LOOP_WAIT_COUNT) - } - } - - private def allocateWorkers() { - try { - logInfo("Allocating " + args.numWorkers + " workers.") - // Wait until all containers have finished - // TODO: This is a bit ugly. Can we make it nicer? - // TODO: Handle container failure - - // Exists the loop if the user thread exits. - while (yarnAllocator.getNumWorkersRunning < args.numWorkers && userThread.isAlive) { - if (yarnAllocator.getNumWorkersFailed >= maxNumWorkerFailures) { - finishApplicationMaster(FinalApplicationStatus.FAILED, - "max number of worker failures reached") - } - yarnAllocator.allocateContainers( - math.max(args.numWorkers - yarnAllocator.getNumWorkersRunning, 0)) - ApplicationMaster.incrementAllocatorLoop(1) - Thread.sleep(100) - } - } finally { - // In case of exceptions, etc - ensure that count is at least ALLOCATOR_LOOP_WAIT_COUNT, - // so that the loop in ApplicationMaster#sparkContextInitialized() breaks. - ApplicationMaster.incrementAllocatorLoop(ApplicationMaster.ALLOCATOR_LOOP_WAIT_COUNT) - } - logInfo("All workers have launched.") - - // Launch a progress reporter thread, else the app will get killed after expiration - // (def: 10mins) timeout. - // TODO(harvey): Verify the timeout - if (userThread.isAlive) { - // Ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapses. - val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) - - // we want to be reasonably responsive without causing too many requests to RM. - val schedulerInterval = - sparkConf.getLong("spark.yarn.scheduler.heartbeat.interval-ms", 5000) - - // must be <= timeoutInterval / 2. - val interval = math.min(timeoutInterval / 2, schedulerInterval) - - launchReporterThread(interval) - } - } - - private def launchReporterThread(_sleepTime: Long): Thread = { - val sleepTime = if (_sleepTime <= 0 ) 0 else _sleepTime - - val t = new Thread { - override def run() { - while (userThread.isAlive) { - if (yarnAllocator.getNumWorkersFailed >= maxNumWorkerFailures) { - finishApplicationMaster(FinalApplicationStatus.FAILED, - "max number of worker failures reached") - } - val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning - if (missingWorkerCount > 0) { - logInfo("Allocating %d containers to make up for (potentially) lost containers". - format(missingWorkerCount)) - yarnAllocator.allocateContainers(missingWorkerCount) - } - else sendProgress() - Thread.sleep(sleepTime) - } - } - } - // Setting to daemon status, though this is usually not a good idea. - t.setDaemon(true) - t.start() - logInfo("Started progress reporter thread - sleep time : " + sleepTime) - t - } - - private def sendProgress() { - logDebug("Sending progress") - // Simulated with an allocate request with no nodes requested ... - yarnAllocator.allocateContainers(0) - } - - /* - def printContainers(containers: List[Container]) = { - for (container <- containers) { - logInfo("Launching shell command on a new container." - + ", containerId=" + container.getId() - + ", containerNode=" + container.getNodeId().getHost() - + ":" + container.getNodeId().getPort() - + ", containerNodeURI=" + container.getNodeHttpAddress() - + ", containerState" + container.getState() - + ", containerResourceMemory" - + container.getResource().getMemory()) - } - } - */ - - def finishApplicationMaster(status: FinalApplicationStatus, diagnostics: String = "") { - synchronized { - if (isFinished) { - return - } - isFinished = true - } - - logInfo("finishApplicationMaster with " + status) - val finishReq = Records.newRecord(classOf[FinishApplicationMasterRequest]) - .asInstanceOf[FinishApplicationMasterRequest] - finishReq.setAppAttemptId(appAttemptId) - finishReq.setFinishApplicationStatus(status) - finishReq.setDiagnostics(diagnostics) - // Set tracking url to empty since we don't have a history server. - finishReq.setTrackingUrl("") - resourceManager.finishApplicationMaster(finishReq) - } - - /** - * Clean up the staging directory. - */ - private def cleanupStagingDir() { - var stagingDirPath: Path = null - try { - val preserveFiles = sparkConf.get("spark.yarn.preserve.staging.files", "false").toBoolean - if (!preserveFiles) { - stagingDirPath = new Path(System.getenv("SPARK_YARN_STAGING_DIR")) - if (stagingDirPath == null) { - logError("Staging directory is null") - return - } - logInfo("Deleting staging directory " + stagingDirPath) - fs.delete(stagingDirPath, true) - } - } catch { - case ioe: IOException => - logError("Failed to cleanup staging dir " + stagingDirPath, ioe) - } - } - - // The shutdown hook that runs when a signal is received AND during normal close of the JVM. - class AppMasterShutdownHook(appMaster: ApplicationMaster) extends Runnable { - - def run() { - logInfo("AppMaster received a signal.") - // we need to clean up staging dir before HDFS is shut down - // make sure we don't delete it until this is the last AM - if (appMaster.isLastAMRetry) appMaster.cleanupStagingDir() - } - } -} - -object ApplicationMaster { - // Number of times to wait for the allocator loop to complete. - // Each loop iteration waits for 100ms, so maximum of 3 seconds. - // This is to ensure that we have reasonable number of containers before we start - // TODO: Currently, task to container is computed once (TaskSetManager) - which need not be - // optimal as more containers are available. Might need to handle this better. - private val ALLOCATOR_LOOP_WAIT_COUNT = 30 - def incrementAllocatorLoop(by: Int) { - val count = yarnAllocatorLoop.getAndAdd(by) - if (count >= ALLOCATOR_LOOP_WAIT_COUNT) { - yarnAllocatorLoop.synchronized { - // to wake threads off wait ... - yarnAllocatorLoop.notifyAll() - } - } - } - - private val applicationMasters = new CopyOnWriteArrayList[ApplicationMaster]() - - def register(master: ApplicationMaster) { - applicationMasters.add(master) - } - - val sparkContextRef: AtomicReference[SparkContext] = - new AtomicReference[SparkContext](null /* initialValue */) - val yarnAllocatorLoop: AtomicInteger = new AtomicInteger(0) - - def sparkContextInitialized(sc: SparkContext): Boolean = { - var modified = false - sparkContextRef.synchronized { - modified = sparkContextRef.compareAndSet(null, sc) - sparkContextRef.notifyAll() - } - - // Add a shutdown hook - as a best case effort in case users do not call sc.stop or do - // System.exit. - // Should not really have to do this, but it helps YARN to evict resources earlier. - // Not to mention, prevent the Client from declaring failure even though we exited properly. - // Note that this will unfortunately not properly clean up the staging files because it gets - // called too late, after the filesystem is already shutdown. - if (modified) { - Runtime.getRuntime().addShutdownHook(new Thread with Logging { - // This is not only logs, but also ensures that log system is initialized for this instance - // when we are actually 'run'-ing. - logInfo("Adding shutdown hook for context " + sc) - override def run() { - logInfo("Invoking sc stop from shutdown hook") - sc.stop() - // Best case ... - 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) { - yarnAllocatorLoop.wait(1000L) - } - } - modified - } - - def main(argStrings: Array[String]) { - val args = new ApplicationMasterArguments(argStrings) - new ApplicationMaster(args).run() - } -} diff --git a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/Client.scala b/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/Client.scala deleted file mode 100644 index 2bd047c97a..0000000000 --- a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/Client.scala +++ /dev/null @@ -1,505 +0,0 @@ -/* - * 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. - */ - -package org.apache.spark.deploy.yarn - -import java.net.{InetAddress, UnknownHostException, URI} -import java.nio.ByteBuffer - -import scala.collection.JavaConversions._ -import scala.collection.mutable.HashMap -import scala.collection.mutable.Map - -import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.fs.{FileContext, FileStatus, FileSystem, Path, FileUtil} -import org.apache.hadoop.fs.permission.FsPermission; -import org.apache.hadoop.io.DataOutputBuffer -import org.apache.hadoop.mapred.Master -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.protocolrecords._ -import org.apache.hadoop.yarn.api.records._ -import org.apache.hadoop.yarn.client.YarnClientImpl -import org.apache.hadoop.yarn.conf.YarnConfiguration -import org.apache.hadoop.yarn.ipc.YarnRPC -import org.apache.hadoop.yarn.util.{Apps, Records} - -import org.apache.spark.{Logging, SparkConf} -import org.apache.spark.util.Utils -import org.apache.spark.deploy.SparkHadoopUtil - - -class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl with Logging { - - def this(args: ClientArguments) = this(new Configuration(), args) - - var rpc: YarnRPC = YarnRPC.create(conf) - val yarnConf: YarnConfiguration = new YarnConfiguration(conf) - val credentials = UserGroupInformation.getCurrentUser().getCredentials() - private val SPARK_STAGING: String = ".sparkStaging" - private val distCacheMgr = new ClientDistributedCacheManager() - private val sparkConf = new SparkConf - - // Staging directory is private! -> rwx-------- - val STAGING_DIR_PERMISSION: FsPermission = FsPermission.createImmutable(0700:Short) - - // App files are world-wide readable and owner writable -> rw-r--r-- - val APP_FILE_PERMISSION: FsPermission = FsPermission.createImmutable(0644:Short) - - // for client user who want to monitor app status by itself. - def runApp() = { - validateArgs() - - init(yarnConf) - start() - logClusterResourceDetails() - - val newApp = super.getNewApplication() - val appId = newApp.getApplicationId() - - verifyClusterResources(newApp) - val appContext = createApplicationSubmissionContext(appId) - val appStagingDir = getAppStagingDir(appId) - val localResources = prepareLocalResources(appStagingDir) - val env = setupLaunchEnv(localResources, appStagingDir) - val amContainer = createContainerLaunchContext(newApp, localResources, env) - - appContext.setQueue(args.amQueue) - appContext.setAMContainerSpec(amContainer) - appContext.setUser(UserGroupInformation.getCurrentUser().getShortUserName()) - - submitApp(appContext) - appId - } - - def run() { - val appId = runApp() - monitorApplication(appId) - System.exit(0) - } - - def validateArgs() = { - Map( - (System.getenv("SPARK_JAR") == null) -> "Error: You must set SPARK_JAR environment variable!", - (args.userJar == null) -> "Error: You must specify a user jar!", - (args.userClass == null) -> "Error: You must specify a user class!", - (args.numWorkers <= 0) -> "Error: You must specify atleast 1 worker!", - (args.amMemory <= YarnAllocationHandler.MEMORY_OVERHEAD) -> ("Error: AM memory size must be " + - "greater than: " + YarnAllocationHandler.MEMORY_OVERHEAD), - (args.workerMemory <= YarnAllocationHandler.MEMORY_OVERHEAD) -> ("Error: Worker memory size " + - "must be greater than: " + YarnAllocationHandler.MEMORY_OVERHEAD) - ).foreach { case(cond, errStr) => - if (cond) { - logError(errStr) - args.printUsageAndExit(1) - } - } - } - - def getAppStagingDir(appId: ApplicationId): String = { - SPARK_STAGING + Path.SEPARATOR + appId.toString() + Path.SEPARATOR - } - - def logClusterResourceDetails() { - val clusterMetrics: YarnClusterMetrics = super.getYarnClusterMetrics - logInfo("Got Cluster metric info from ASM, numNodeManagers = " + - clusterMetrics.getNumNodeManagers) - - val queueInfo: QueueInfo = super.getQueueInfo(args.amQueue) - logInfo("""Queue info ... queueName = %s, queueCurrentCapacity = %s, queueMaxCapacity = %s, - queueApplicationCount = %s, queueChildQueueCount = %s""".format( - queueInfo.getQueueName, - queueInfo.getCurrentCapacity, - queueInfo.getMaximumCapacity, - queueInfo.getApplications.size, - queueInfo.getChildQueues.size)) - } - - def verifyClusterResources(app: GetNewApplicationResponse) = { - val maxMem = app.getMaximumResourceCapability().getMemory() - logInfo("Max mem capabililty of a single resource in this cluster " + maxMem) - - // 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 = { - logInfo("Setting up application submission context for ASM") - val appContext = Records.newRecord(classOf[ApplicationSubmissionContext]) - appContext.setApplicationId(appId) - appContext.setApplicationName(args.appName) - return appContext - } - - /** See if two file systems are the same or not. */ - private def compareFs(srcFs: FileSystem, destFs: FileSystem): Boolean = { - val srcUri = srcFs.getUri() - val dstUri = destFs.getUri() - if (srcUri.getScheme() == null) { - return false - } - if (!srcUri.getScheme().equals(dstUri.getScheme())) { - return false - } - var srcHost = srcUri.getHost() - var dstHost = dstUri.getHost() - if ((srcHost != null) && (dstHost != null)) { - try { - srcHost = InetAddress.getByName(srcHost).getCanonicalHostName() - dstHost = InetAddress.getByName(dstHost).getCanonicalHostName() - } catch { - case e: UnknownHostException => - return false - } - if (!srcHost.equals(dstHost)) { - return false - } - } else if (srcHost == null && dstHost != null) { - return false - } else if (srcHost != null && dstHost == null) { - return false - } - //check for ports - if (srcUri.getPort() != dstUri.getPort()) { - return false - } - return true - } - - /** Copy the file into HDFS if needed. */ - private def copyRemoteFile( - dstDir: Path, - originalPath: Path, - replication: Short, - setPerms: Boolean = false): Path = { - val fs = FileSystem.get(conf) - val remoteFs = originalPath.getFileSystem(conf) - var newPath = originalPath - if (! compareFs(remoteFs, fs)) { - newPath = new Path(dstDir, originalPath.getName()) - logInfo("Uploading " + originalPath + " to " + newPath) - FileUtil.copy(remoteFs, originalPath, fs, newPath, false, conf) - fs.setReplication(newPath, replication) - if (setPerms) fs.setPermission(newPath, new FsPermission(APP_FILE_PERMISSION)) - } - // Resolve any symlinks in the URI path so using a "current" symlink to point to a specific - // version shows the specific version in the distributed cache configuration - val qualPath = fs.makeQualified(newPath) - val fc = FileContext.getFileContext(qualPath.toUri(), conf) - val destPath = fc.resolvePath(qualPath) - destPath - } - - def prepareLocalResources(appStagingDir: String): HashMap[String, LocalResource] = { - logInfo("Preparing Local resources") - // Upload Spark and the application JAR to the remote file system if necessary. 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) - } - } - val dst = new Path(fs.getHomeDirectory(), appStagingDir) - val replication = sparkConf.getInt("spark.yarn.submit.file.replication", 3).toShort - - if (UserGroupInformation.isSecurityEnabled()) { - val dstFs = dst.getFileSystem(conf) - dstFs.addDelegationTokens(delegTokenRenewer, credentials) - } - val localResources = HashMap[String, LocalResource]() - FileSystem.mkdirs(fs, dst, new FsPermission(STAGING_DIR_PERMISSION)) - - val statCache: Map[URI, FileStatus] = HashMap[URI, FileStatus]() - - Map(Client.SPARK_JAR -> System.getenv("SPARK_JAR"), Client.APP_JAR -> args.userJar, - Client.LOG4J_PROP -> System.getenv("SPARK_LOG4J_CONF")) - .foreach { case(destName, _localPath) => - val localPath: String = if (_localPath != null) _localPath.trim() else "" - if (! localPath.isEmpty()) { - var localURI = new URI(localPath) - // if not specified assume these are in the local filesystem to keep behavior like Hadoop - if (localURI.getScheme() == null) { - localURI = new URI(FileSystem.getLocal(conf).makeQualified(new Path(localPath)).toString) - } - val setPermissions = if (destName.equals(Client.APP_JAR)) true else false - val destPath = copyRemoteFile(dst, new Path(localURI), replication, setPermissions) - distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, - destName, statCache) - } - } - - // handle any add jars - if ((args.addJars != null) && (!args.addJars.isEmpty())){ - args.addJars.split(',').foreach { case file: String => - val localURI = new URI(file.trim()) - val localPath = new Path(localURI) - val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) - val destPath = copyRemoteFile(dst, localPath, replication) - distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, - linkname, statCache, true) - } - } - - // handle any distributed cache files - if ((args.files != null) && (!args.files.isEmpty())){ - args.files.split(',').foreach { case file: String => - val localURI = new URI(file.trim()) - val localPath = new Path(localURI) - val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) - val destPath = copyRemoteFile(dst, localPath, replication) - distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, - linkname, statCache) - } - } - - // handle any distributed cache archives - if ((args.archives != null) && (!args.archives.isEmpty())) { - args.archives.split(',').foreach { case file:String => - val localURI = new URI(file.trim()) - val localPath = new Path(localURI) - val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) - val destPath = copyRemoteFile(dst, localPath, replication) - distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.ARCHIVE, - linkname, statCache) - } - } - - UserGroupInformation.getCurrentUser().addCredentials(credentials) - return localResources - } - - def setupLaunchEnv( - localResources: HashMap[String, LocalResource], - stagingDir: String): HashMap[String, String] = { - logInfo("Setting up the launch environment") - val log4jConfLocalRes = localResources.getOrElse(Client.LOG4J_PROP, null) - - val env = new HashMap[String, String]() - - Client.populateClasspath(yarnConf, log4jConfLocalRes != null, env) - env("SPARK_YARN_MODE") = "true" - env("SPARK_YARN_STAGING_DIR") = stagingDir - - // Set the environment variables to be passed on to the Workers. - distCacheMgr.setDistFilesEnv(env) - distCacheMgr.setDistArchivesEnv(env) - - // Allow users to specify some environment variables. - Apps.setEnvFromInputString(env, System.getenv("SPARK_YARN_USER_ENV")) - - // Add each SPARK-* key to the environment. - System.getenv().filterKeys(_.startsWith("SPARK")).foreach { case (k,v) => env(k) = v } - env - } - - def userArgsToString(clientArgs: ClientArguments): String = { - val prefix = " --args " - val args = clientArgs.userArgs - val retval = new StringBuilder() - for (arg <- args){ - retval.append(prefix).append(" '").append(arg).append("' ") - } - retval.toString - } - - def createContainerLaunchContext( - newApp: GetNewApplicationResponse, - localResources: HashMap[String, LocalResource], - env: HashMap[String, String]): ContainerLaunchContext = { - logInfo("Setting up container launch context") - val amContainer = Records.newRecord(classOf[ContainerLaunchContext]) - amContainer.setLocalResources(localResources) - amContainer.setEnvironment(env) - - val minResMemory: Int = newApp.getMinimumResourceCapability().getMemory() - - // TODO(harvey): This can probably be a val. - var amMemory = ((args.amMemory / minResMemory) * minResMemory) + - ((if ((args.amMemory % minResMemory) == 0) 0 else minResMemory) - - YarnAllocationHandler.MEMORY_OVERHEAD) - - // Extra options for the JVM - var JAVA_OPTS = "" - - // Add Xmx for am memory - JAVA_OPTS += "-Xmx" + amMemory + "m " - - 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. - val useConcurrentAndIncrementalGC = env.isDefinedAt("SPARK_USE_CONC_INCR_GC") && - java.lang.Boolean.parseBoolean(env("SPARK_USE_CONC_INCR_GC")) - if (useConcurrentAndIncrementalGC) { - // In our expts, using (default) throughput collector has severe perf ramnifications in - // multi-tenant machines - JAVA_OPTS += " -XX:+UseConcMarkSweepGC " - JAVA_OPTS += " -XX:+CMSIncrementalMode " - JAVA_OPTS += " -XX:+CMSIncrementalPacing " - JAVA_OPTS += " -XX:CMSIncrementalDutyCycleMin=0 " - JAVA_OPTS += " -XX:CMSIncrementalDutyCycle=10 " - } - - if (env.isDefinedAt("SPARK_JAVA_OPTS")) { - JAVA_OPTS += env("SPARK_JAVA_OPTS") + " " - } - - // Command for the ApplicationMaster - 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 " + - JAVA_OPTS + - " " + args.amClass + - " --class " + args.userClass + - " --jar " + args.userJar + - userArgsToString(args) + - " --worker-memory " + args.workerMemory + - " --worker-cores " + args.workerCores + - " --num-workers " + args.numWorkers + - " 1> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stdout" + - " 2> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stderr") - logInfo("Command for the ApplicationMaster: " + commands(0)) - amContainer.setCommands(commands) - - val capability = Records.newRecord(classOf[Resource]).asInstanceOf[Resource] - // 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())) - - amContainer - } - - def submitApp(appContext: ApplicationSubmissionContext) = { - // Submit the application to the applications manager. - logInfo("Submitting application to ASM") - super.submitApplication(appContext) - } - - def monitorApplication(appId: ApplicationId): Boolean = { - val interval = sparkConf.getLong("spark.yarn.report.interval", 1000) - - while (true) { - Thread.sleep(interval) - val report = super.getApplicationReport(appId) - - logInfo("Application report from ASM: \n" + - "\t application identifier: " + appId.toString() + "\n" + - "\t appId: " + appId.getId() + "\n" + - "\t clientToken: " + report.getClientToken() + "\n" + - "\t appDiagnostics: " + report.getDiagnostics() + "\n" + - "\t appMasterHost: " + report.getHost() + "\n" + - "\t appQueue: " + report.getQueue() + "\n" + - "\t appMasterRpcPort: " + report.getRpcPort() + "\n" + - "\t appStartTime: " + report.getStartTime() + "\n" + - "\t yarnAppState: " + report.getYarnApplicationState() + "\n" + - "\t distributedFinalState: " + report.getFinalApplicationStatus() + "\n" + - "\t appTrackingUrl: " + report.getTrackingUrl() + "\n" + - "\t appUser: " + report.getUser() - ) - - val state = report.getYarnApplicationState() - val dsStatus = report.getFinalApplicationStatus() - if (state == YarnApplicationState.FINISHED || - state == YarnApplicationState.FAILED || - state == YarnApplicationState.KILLED) { - return true - } - } - true - } -} - -object Client { - val SPARK_JAR: String = "spark.jar" - val APP_JAR: String = "app.jar" - val LOG4J_PROP: String = "log4j.properties" - - def main(argStrings: Array[String]) { - // Set an env variable indicating we are running in YARN mode. - // Note that anything with SPARK prefix gets propagated to all (remote) processes - System.setProperty("SPARK_YARN_MODE", "true") - - val args = new ClientArguments(argStrings) - - new Client(args).run - } - - // Based on code from org.apache.hadoop.mapreduce.v2.util.MRApps - def populateHadoopClasspath(conf: Configuration, env: HashMap[String, String]) { - for (c <- conf.getStrings(YarnConfiguration.YARN_APPLICATION_CLASSPATH)) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, c.trim) - } - } - - def populateClasspath(conf: Configuration, addLog4j: Boolean, env: HashMap[String, String]) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$()) - // If log4j present, ensure ours overrides all others - if (addLog4j) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + LOG4J_PROP) - } - // Normally the users app.jar is last in case conflicts with spark jars - val userClasspathFirst = new SparkConf().get("spark.yarn.user.classpath.first", "false").toBoolean - if (userClasspathFirst) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + APP_JAR) - } - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + SPARK_JAR) - Client.populateHadoopClasspath(conf, env) - - if (!userClasspathFirst) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + APP_JAR) - } - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + "*") - } -} diff --git a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala b/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala deleted file mode 100644 index e64530702c..0000000000 --- a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala +++ /dev/null @@ -1,248 +0,0 @@ -/* - * 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. - */ - -package org.apache.spark.deploy.yarn - -import java.net.Socket -import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.net.NetUtils -import org.apache.hadoop.yarn.api._ -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.{ConverterUtils, Records} -import akka.actor._ -import akka.remote._ -import akka.actor.Terminated -import org.apache.spark.{SparkConf, SparkContext, Logging} -import org.apache.spark.util.{Utils, AkkaUtils} -import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend -import org.apache.spark.scheduler.SplitInfo - -class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration) extends Logging { - - def this(args: ApplicationMasterArguments) = this(args, new Configuration()) - - private val rpc: YarnRPC = YarnRPC.create(conf) - private var resourceManager: AMRMProtocol = _ - private var appAttemptId: ApplicationAttemptId = _ - private var reporterThread: Thread = _ - private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) - - private var yarnAllocator: YarnAllocationHandler = _ - private var driverClosed:Boolean = false - private val sparkConf = new SparkConf - - val actorSystem : ActorSystem = AkkaUtils.createActorSystem("sparkYarnAM", Utils.localHostName, 0, - conf = sparkConf)._1 - var actor: ActorRef = _ - - // This actor just working as a monitor to watch on Driver Actor. - class MonitorActor(driverUrl: String) extends Actor { - - var driver: ActorSelection = _ - - override def preStart() { - logInfo("Listen to driver: " + driverUrl) - driver = context.actorSelection(driverUrl) - // Send a hello message thus the connection is actually established, thus we can monitor Lifecycle Events. - driver ! "Hello" - context.system.eventStream.subscribe(self, classOf[RemotingLifecycleEvent]) - } - - override def receive = { - case x: DisassociatedEvent => - logInfo(s"Driver terminated or disconnected! Shutting down. $x") - driverClosed = true - } - } - - def run() { - - appAttemptId = getApplicationAttemptId() - resourceManager = registerWithResourceManager() - val appMasterResponse: RegisterApplicationMasterResponse = registerApplicationMaster() - - // Compute number of threads for akka - val minimumMemory = appMasterResponse.getMinimumResourceCapability().getMemory() - - if (minimumMemory > 0) { - val mem = args.workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD - val numCore = (mem / minimumMemory) + (if (0 != (mem % minimumMemory)) 1 else 0) - - if (numCore > 0) { - // do not override - hits https://issues.apache.org/jira/browse/HADOOP-8406 - // TODO: Uncomment when hadoop is on a version which has this fixed. - // args.workerCores = numCore - } - } - - waitForSparkMaster() - - // Allocate all containers - allocateWorkers() - - // Launch a progress reporter thread, else app will get killed after expiration (def: 10mins) timeout - // ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapse. - - val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) - // must be <= timeoutInterval/ 2. - // On other hand, also ensure that we are reasonably responsive without causing too many requests to RM. - // so atleast 1 minute or timeoutInterval / 10 - whichever is higher. - val interval = math.min(timeoutInterval / 2, math.max(timeoutInterval/ 10, 60000L)) - reporterThread = launchReporterThread(interval) - - // Wait for the reporter thread to Finish. - reporterThread.join() - - finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) - actorSystem.shutdown() - - logInfo("Exited") - System.exit(0) - } - - private def getApplicationAttemptId(): ApplicationAttemptId = { - val envs = System.getenv() - val containerIdString = envs.get(ApplicationConstants.AM_CONTAINER_ID_ENV) - val containerId = ConverterUtils.toContainerId(containerIdString) - val appAttemptId = containerId.getApplicationAttemptId() - logInfo("ApplicationAttemptId: " + appAttemptId) - return appAttemptId - } - - private def registerWithResourceManager(): AMRMProtocol = { - val rmAddress = NetUtils.createSocketAddr(yarnConf.get( - YarnConfiguration.RM_SCHEDULER_ADDRESS, - YarnConfiguration.DEFAULT_RM_SCHEDULER_ADDRESS)) - logInfo("Connecting to ResourceManager at " + rmAddress) - return rpc.getProxy(classOf[AMRMProtocol], rmAddress, conf).asInstanceOf[AMRMProtocol] - } - - private def registerApplicationMaster(): RegisterApplicationMasterResponse = { - logInfo("Registering the ApplicationMaster") - val appMasterRequest = Records.newRecord(classOf[RegisterApplicationMasterRequest]) - .asInstanceOf[RegisterApplicationMasterRequest] - appMasterRequest.setApplicationAttemptId(appAttemptId) - // Setting this to master host,port - so that the ApplicationReport at client has some sensible info. - // Users can then monitor stderr/stdout on that node if required. - appMasterRequest.setHost(Utils.localHostName()) - appMasterRequest.setRpcPort(0) - // What do we provide here ? Might make sense to expose something sensible later ? - appMasterRequest.setTrackingUrl("") - return resourceManager.registerApplicationMaster(appMasterRequest) - } - - private def waitForSparkMaster() { - logInfo("Waiting for spark driver to be reachable.") - var driverUp = false - val hostport = args.userArgs(0) - val (driverHost, driverPort) = Utils.parseHostPort(hostport) - while(!driverUp) { - try { - val socket = new Socket(driverHost, driverPort) - socket.close() - logInfo("Master now available: " + driverHost + ":" + driverPort) - driverUp = true - } catch { - case e: Exception => - logError("Failed to connect to driver at " + driverHost + ":" + driverPort) - Thread.sleep(100) - } - } - sparkConf.set("spark.driver.host", driverHost) - sparkConf.set("spark.driver.port", driverPort.toString) - - val driverUrl = "akka.tcp://spark@%s:%s/user/%s".format( - driverHost, driverPort.toString, CoarseGrainedSchedulerBackend.ACTOR_NAME) - - actor = actorSystem.actorOf(Props(new MonitorActor(driverUrl)), name = "YarnAM") - } - - - private def allocateWorkers() { - - // Fixme: should get preferredNodeLocationData from SparkContext, just fake a empty one for now. - val preferredNodeLocationData: scala.collection.Map[String, scala.collection.Set[SplitInfo]] = - scala.collection.immutable.Map() - - yarnAllocator = YarnAllocationHandler.newAllocator(yarnConf, resourceManager, appAttemptId, - args, preferredNodeLocationData, sparkConf) - - logInfo("Allocating " + args.numWorkers + " workers.") - // Wait until all containers have finished - // TODO: This is a bit ugly. Can we make it nicer? - // TODO: Handle container failure - while(yarnAllocator.getNumWorkersRunning < args.numWorkers) { - yarnAllocator.allocateContainers(math.max(args.numWorkers - yarnAllocator.getNumWorkersRunning, 0)) - Thread.sleep(100) - } - - logInfo("All workers have launched.") - - } - - // TODO: We might want to extend this to allocate more containers in case they die ! - private def launchReporterThread(_sleepTime: Long): Thread = { - val sleepTime = if (_sleepTime <= 0 ) 0 else _sleepTime - - val t = new Thread { - override def run() { - while (!driverClosed) { - val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning - if (missingWorkerCount > 0) { - logInfo("Allocating " + missingWorkerCount + " containers to make up for (potentially ?) lost containers") - yarnAllocator.allocateContainers(missingWorkerCount) - } - else sendProgress() - Thread.sleep(sleepTime) - } - } - } - // setting to daemon status, though this is usually not a good idea. - t.setDaemon(true) - t.start() - logInfo("Started progress reporter thread - sleep time : " + sleepTime) - return t - } - - private def sendProgress() { - logDebug("Sending progress") - // simulated with an allocate request with no nodes requested ... - yarnAllocator.allocateContainers(0) - } - - def finishApplicationMaster(status: FinalApplicationStatus) { - - logInfo("finish ApplicationMaster with " + status) - val finishReq = Records.newRecord(classOf[FinishApplicationMasterRequest]) - .asInstanceOf[FinishApplicationMasterRequest] - finishReq.setAppAttemptId(appAttemptId) - finishReq.setFinishApplicationStatus(status) - resourceManager.finishApplicationMaster(finishReq) - } - -} - - -object WorkerLauncher { - def main(argStrings: Array[String]) { - val args = new ApplicationMasterArguments(argStrings) - new WorkerLauncher(args).run() - } -} diff --git a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala b/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala deleted file mode 100644 index 4f34bd913e..0000000000 --- a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala +++ /dev/null @@ -1,235 +0,0 @@ -/* - * 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. - */ - -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.conf.YarnConfiguration -import org.apache.hadoop.yarn.ipc.YarnRPC -import org.apache.hadoop.yarn.util.{Apps, ConverterUtils, Records, ProtoUtils} - -import org.apache.spark.Logging - - -class WorkerRunnable( - container: Container, - conf: Configuration, - masterAddress: String, - slaveId: String, - hostname: String, - workerMemory: Int, - workerCores: Int) - extends Runnable with Logging { - - var rpc: YarnRPC = YarnRPC.create(conf) - var cm: ContainerManager = _ - val yarnConf: YarnConfiguration = new YarnConfiguration(conf) - - def run = { - logInfo("Starting Worker Container") - cm = connectToCM - startContainer - } - - def startContainer = { - logInfo("Setting up ContainerLaunchContext") - - val ctx = Records.newRecord(classOf[ContainerLaunchContext]) - .asInstanceOf[ContainerLaunchContext] - - ctx.setContainerId(container.getId()) - ctx.setResource(container.getResource()) - val localResources = prepareLocalResources - ctx.setLocalResources(localResources) - - val env = prepareEnvironment - ctx.setEnvironment(env) - - // 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 " - } -*/ - - ctx.setUser(UserGroupInformation.getCurrentUser().getShortUserName()) - - 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()) || 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") - logInfo("Setting up worker with commands: " + commands) - ctx.setCommands(commands) - - // Send the start request to the ContainerManager - val startReq = Records.newRecord(classOf[StartContainerRequest]) - .asInstanceOf[StartContainerRequest] - startReq.setContainerLaunchContext(ctx) - cm.startContainer(startReq) - } - - 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) - return localResources - } - - def prepareEnvironment: HashMap[String, String] = { - val env = new HashMap[String, String]() - - Client.populateClasspath(yarnConf, 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 } - return env - } - - def connectToCM: ContainerManager = { - val cmHostPortStr = container.getNodeId().getHost() + ":" + container.getNodeId().getPort() - val cmAddress = NetUtils.createSocketAddr(cmHostPortStr) - logInfo("Connecting to ContainerManager at " + cmHostPortStr) - - // 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] - } - }) - proxy - } - -} diff --git a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala b/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala deleted file mode 100644 index c8af653b3f..0000000000 --- a/yarn/2.0/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala +++ /dev/null @@ -1,680 +0,0 @@ -/* - * 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. - */ - -package org.apache.spark.deploy.yarn - -import java.lang.{Boolean => JBoolean} -import java.util.{Collections, Set => JSet} -import java.util.concurrent.{CopyOnWriteArrayList, ConcurrentHashMap} -import java.util.concurrent.atomic.AtomicInteger - -import scala.collection -import scala.collection.JavaConversions._ -import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} - -import org.apache.spark.{Logging, SparkConf} -import org.apache.spark.scheduler.{SplitInfo,TaskSchedulerImpl} -import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend -import org.apache.spark.util.Utils - -import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.yarn.api.AMRMProtocol -import org.apache.hadoop.yarn.api.records.{AMResponse, ApplicationAttemptId} -import org.apache.hadoop.yarn.api.records.{Container, ContainerId, ContainerStatus} -import org.apache.hadoop.yarn.api.records.{Priority, Resource, ResourceRequest} -import org.apache.hadoop.yarn.api.protocolrecords.{AllocateRequest, AllocateResponse} -import org.apache.hadoop.yarn.util.{RackResolver, Records} - - -object AllocationType extends Enumeration { - type AllocationType = Value - val HOST, RACK, ANY = Value -} - -// TODO: -// Too many params. -// Needs to be mt-safe -// Need to refactor this to make it 'cleaner' ... right now, all computation is reactive - should -// make it more proactive and decoupled. - -// Note that right now, we assume all node asks as uniform in terms of capabilities and priority -// Refer to http://developer.yahoo.com/blogs/hadoop/posts/2011/03/mapreduce-nextgen-scheduler/ for -// more info on how we are requesting for containers. -private[yarn] class YarnAllocationHandler( - val conf: Configuration, - val resourceManager: AMRMProtocol, - val appAttemptId: ApplicationAttemptId, - val maxWorkers: Int, - val workerMemory: Int, - val workerCores: Int, - val preferredHostToCount: Map[String, Int], - val preferredRackToCount: Map[String, Int], - val sparkConf: SparkConf) - extends Logging { - // These three are locked on allocatedHostToContainersMap. Complementary data structures - // allocatedHostToContainersMap : containers which are running : host, Set - // allocatedContainerToHostMap: container to host mapping. - private val allocatedHostToContainersMap = - new HashMap[String, collection.mutable.Set[ContainerId]]() - - private val allocatedContainerToHostMap = new HashMap[ContainerId, String]() - - // allocatedRackCount is populated ONLY if allocation happens (or decremented if this is an - // allocated node) - // As with the two data structures above, tightly coupled with them, and to be locked on - // allocatedHostToContainersMap - private val allocatedRackCount = new HashMap[String, Int]() - - // Containers which have been released. - private val releasedContainerList = new CopyOnWriteArrayList[ContainerId]() - // Containers to be released in next request to RM - private val pendingReleaseContainers = new ConcurrentHashMap[ContainerId, Boolean] - - private val numWorkersRunning = new AtomicInteger() - // Used to generate a unique id per worker - private val workerIdCounter = new AtomicInteger() - private val lastResponseId = new AtomicInteger() - private val numWorkersFailed = new AtomicInteger() - - def getNumWorkersRunning: Int = numWorkersRunning.intValue - - def getNumWorkersFailed: Int = numWorkersFailed.intValue - - def isResourceConstraintSatisfied(container: Container): Boolean = { - container.getResource.getMemory >= (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) - } - - def allocateContainers(workersToRequest: Int) { - // We need to send the request only once from what I understand ... but for now, not modifying - // this much. - - // Keep polling the Resource Manager for containers - val amResp = allocateWorkerResources(workersToRequest).getAMResponse - - val _allocatedContainers = amResp.getAllocatedContainers() - - if (_allocatedContainers.size > 0) { - logDebug(""" - Allocated containers: %d - Current worker count: %d - Containers released: %s - Containers to be released: %s - Cluster resources: %s - """.format( - _allocatedContainers.size, - numWorkersRunning.get(), - releasedContainerList, - pendingReleaseContainers, - amResp.getAvailableResources)) - - val hostToContainers = new HashMap[String, ArrayBuffer[Container]]() - - // Ignore if not satisfying constraints { - for (container <- _allocatedContainers) { - if (isResourceConstraintSatisfied(container)) { - // allocatedContainers += container - - val host = container.getNodeId.getHost - val containers = hostToContainers.getOrElseUpdate(host, new ArrayBuffer[Container]()) - - containers += container - } - // Add all ignored containers to released list - else releasedContainerList.add(container.getId()) - } - - // Find the appropriate containers to use. Slightly non trivial groupBy ... - val dataLocalContainers = new HashMap[String, ArrayBuffer[Container]]() - val rackLocalContainers = new HashMap[String, ArrayBuffer[Container]]() - val offRackContainers = new HashMap[String, ArrayBuffer[Container]]() - - for (candidateHost <- hostToContainers.keySet) - { - val maxExpectedHostCount = preferredHostToCount.getOrElse(candidateHost, 0) - val requiredHostCount = maxExpectedHostCount - allocatedContainersOnHost(candidateHost) - - var remainingContainers = hostToContainers.get(candidateHost).getOrElse(null) - assert(remainingContainers != null) - - if (requiredHostCount >= remainingContainers.size){ - // Since we got <= required containers, add all to dataLocalContainers - dataLocalContainers.put(candidateHost, remainingContainers) - // all consumed - remainingContainers = null - } - else if (requiredHostCount > 0) { - // Container list has more containers than we need for data locality. - // Split into two : data local container count of (remainingContainers.size - - // requiredHostCount) and rest as remainingContainer - val (dataLocal, remaining) = remainingContainers.splitAt( - remainingContainers.size - requiredHostCount) - dataLocalContainers.put(candidateHost, dataLocal) - // remainingContainers = remaining - - // yarn has nasty habit of allocating a tonne of containers on a host - discourage this : - // add remaining to release list. If we have insufficient containers, next allocation - // cycle will reallocate (but wont treat it as data local) - for (container <- remaining) releasedContainerList.add(container.getId()) - remainingContainers = null - } - - // Now rack local - if (remainingContainers != null){ - val rack = YarnAllocationHandler.lookupRack(conf, candidateHost) - - if (rack != null){ - val maxExpectedRackCount = preferredRackToCount.getOrElse(rack, 0) - val requiredRackCount = maxExpectedRackCount - allocatedContainersOnRack(rack) - - rackLocalContainers.get(rack).getOrElse(List()).size - - - if (requiredRackCount >= remainingContainers.size){ - // Add all to dataLocalContainers - dataLocalContainers.put(rack, remainingContainers) - // All consumed - remainingContainers = null - } - else if (requiredRackCount > 0) { - // container list has more containers than we need for data locality. - // Split into two : data local container count of (remainingContainers.size - - // requiredRackCount) and rest as remainingContainer - val (rackLocal, remaining) = remainingContainers.splitAt( - remainingContainers.size - requiredRackCount) - val existingRackLocal = rackLocalContainers.getOrElseUpdate(rack, - new ArrayBuffer[Container]()) - - existingRackLocal ++= rackLocal - remainingContainers = remaining - } - } - } - - // If still not consumed, then it is off rack host - add to that list. - if (remainingContainers != null){ - offRackContainers.put(candidateHost, remainingContainers) - } - } - - // Now that we have split the containers into various groups, go through them in order : - // first host local, then rack local and then off rack (everything else). - // Note that the list we create below tries to ensure that not all containers end up within a - // host if there are sufficiently large number of hosts/containers. - - val allocatedContainers = new ArrayBuffer[Container](_allocatedContainers.size) - allocatedContainers ++= TaskSchedulerImpl.prioritizeContainers(dataLocalContainers) - allocatedContainers ++= TaskSchedulerImpl.prioritizeContainers(rackLocalContainers) - allocatedContainers ++= TaskSchedulerImpl.prioritizeContainers(offRackContainers) - - // Run each of the allocated containers - for (container <- allocatedContainers) { - val numWorkersRunningNow = numWorkersRunning.incrementAndGet() - val workerHostname = container.getNodeId.getHost - val containerId = container.getId - - assert( - container.getResource.getMemory >= (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD)) - - if (numWorkersRunningNow > maxWorkers) { - logInfo("""Ignoring container %s at host %s, since we already have the required number of - containers for it.""".format(containerId, workerHostname)) - releasedContainerList.add(containerId) - // reset counter back to old value. - numWorkersRunning.decrementAndGet() - } - else { - // Deallocate + allocate can result in reusing id's wrongly - so use a different counter - // (workerIdCounter) - val workerId = workerIdCounter.incrementAndGet().toString - val driverUrl = "akka.tcp://spark@%s:%s/user/%s".format( - sparkConf.get("spark.driver.host"), sparkConf.get("spark.driver.port"), - CoarseGrainedSchedulerBackend.ACTOR_NAME) - - logInfo("launching container on " + containerId + " host " + workerHostname) - // Just to be safe, simply remove it from pendingReleaseContainers. - // Should not be there, but .. - pendingReleaseContainers.remove(containerId) - - val rack = YarnAllocationHandler.lookupRack(conf, workerHostname) - allocatedHostToContainersMap.synchronized { - val containerSet = allocatedHostToContainersMap.getOrElseUpdate(workerHostname, - new HashSet[ContainerId]()) - - containerSet += containerId - allocatedContainerToHostMap.put(containerId, workerHostname) - if (rack != null) { - allocatedRackCount.put(rack, allocatedRackCount.getOrElse(rack, 0) + 1) - } - } - - new Thread( - new WorkerRunnable(container, conf, driverUrl, workerId, - workerHostname, workerMemory, workerCores) - ).start() - } - } - logDebug(""" - Finished processing %d containers. - Current number of workers running: %d, - releasedContainerList: %s, - pendingReleaseContainers: %s - """.format( - allocatedContainers.size, - numWorkersRunning.get(), - releasedContainerList, - pendingReleaseContainers)) - } - - - val completedContainers = amResp.getCompletedContainersStatuses() - if (completedContainers.size > 0){ - logDebug("Completed %d containers, to-be-released: %s".format( - completedContainers.size, releasedContainerList)) - for (completedContainer <- completedContainers){ - val containerId = completedContainer.getContainerId - - // Was this released by us ? If yes, then simply remove from containerSet and move on. - if (pendingReleaseContainers.containsKey(containerId)) { - pendingReleaseContainers.remove(containerId) - } - else { - // Simply decrement count - next iteration of ReporterThread will take care of allocating. - numWorkersRunning.decrementAndGet() - logInfo("Completed container %s (state: %s, exit status: %s)".format( - containerId, - completedContainer.getState, - completedContainer.getExitStatus())) - // Hadoop 2.2.X added a ContainerExitStatus we should switch to use - // there are some exit status' we shouldn't necessarily count against us, but for - // now I think its ok as none of the containers are expected to exit - if (completedContainer.getExitStatus() != 0) { - logInfo("Container marked as failed: " + containerId) - numWorkersFailed.incrementAndGet() - } - } - - allocatedHostToContainersMap.synchronized { - if (allocatedContainerToHostMap.containsKey(containerId)) { - val host = allocatedContainerToHostMap.get(containerId).getOrElse(null) - assert (host != null) - - val containerSet = allocatedHostToContainersMap.get(host).getOrElse(null) - assert (containerSet != null) - - containerSet -= containerId - if (containerSet.isEmpty) allocatedHostToContainersMap.remove(host) - else allocatedHostToContainersMap.update(host, containerSet) - - allocatedContainerToHostMap -= containerId - - // Doing this within locked context, sigh ... move to outside ? - val rack = YarnAllocationHandler.lookupRack(conf, host) - if (rack != null) { - val rackCount = allocatedRackCount.getOrElse(rack, 0) - 1 - if (rackCount > 0) allocatedRackCount.put(rack, rackCount) - else allocatedRackCount.remove(rack) - } - } - } - } - logDebug(""" - Finished processing %d completed containers. - Current number of workers running: %d, - releasedContainerList: %s, - pendingReleaseContainers: %s - """.format( - completedContainers.size, - numWorkersRunning.get(), - releasedContainerList, - pendingReleaseContainers)) - } - } - - def createRackResourceRequests(hostContainers: List[ResourceRequest]): List[ResourceRequest] = { - // First generate modified racks and new set of hosts under it : then issue requests - val rackToCounts = new HashMap[String, Int]() - - // Within this lock - used to read/write to the rack related maps too. - for (container <- hostContainers) { - val candidateHost = container.getHostName - val candidateNumContainers = container.getNumContainers - assert(YarnAllocationHandler.ANY_HOST != candidateHost) - - val rack = YarnAllocationHandler.lookupRack(conf, candidateHost) - if (rack != null) { - var count = rackToCounts.getOrElse(rack, 0) - count += candidateNumContainers - rackToCounts.put(rack, count) - } - } - - val requestedContainers: ArrayBuffer[ResourceRequest] = - new ArrayBuffer[ResourceRequest](rackToCounts.size) - for ((rack, count) <- rackToCounts){ - requestedContainers += - createResourceRequest(AllocationType.RACK, rack, count, YarnAllocationHandler.PRIORITY) - } - - requestedContainers.toList - } - - def allocatedContainersOnHost(host: String): Int = { - var retval = 0 - allocatedHostToContainersMap.synchronized { - retval = allocatedHostToContainersMap.getOrElse(host, Set()).size - } - retval - } - - def allocatedContainersOnRack(rack: String): Int = { - var retval = 0 - allocatedHostToContainersMap.synchronized { - retval = allocatedRackCount.getOrElse(rack, 0) - } - retval - } - - private def allocateWorkerResources(numWorkers: Int): AllocateResponse = { - - var resourceRequests: List[ResourceRequest] = null - - // default. - if (numWorkers <= 0 || preferredHostToCount.isEmpty) { - logDebug("numWorkers: " + numWorkers + ", host preferences: " + preferredHostToCount.isEmpty) - resourceRequests = List( - createResourceRequest(AllocationType.ANY, null, numWorkers, YarnAllocationHandler.PRIORITY)) - } - else { - // request for all hosts in preferred nodes and for numWorkers - - // candidates.size, request by default allocation policy. - val hostContainerRequests: ArrayBuffer[ResourceRequest] = - new ArrayBuffer[ResourceRequest](preferredHostToCount.size) - for ((candidateHost, candidateCount) <- preferredHostToCount) { - val requiredCount = candidateCount - allocatedContainersOnHost(candidateHost) - - if (requiredCount > 0) { - hostContainerRequests += createResourceRequest( - AllocationType.HOST, - candidateHost, - requiredCount, - YarnAllocationHandler.PRIORITY) - } - } - val rackContainerRequests: List[ResourceRequest] = createRackResourceRequests( - hostContainerRequests.toList) - - val anyContainerRequests: ResourceRequest = createResourceRequest( - AllocationType.ANY, - resource = null, - numWorkers, - YarnAllocationHandler.PRIORITY) - - val containerRequests: ArrayBuffer[ResourceRequest] = new ArrayBuffer[ResourceRequest]( - hostContainerRequests.size + rackContainerRequests.size + 1) - - containerRequests ++= hostContainerRequests - containerRequests ++= rackContainerRequests - containerRequests += anyContainerRequests - - resourceRequests = containerRequests.toList - } - - val req = Records.newRecord(classOf[AllocateRequest]) - req.setResponseId(lastResponseId.incrementAndGet) - req.setApplicationAttemptId(appAttemptId) - - req.addAllAsks(resourceRequests) - - val releasedContainerList = createReleasedContainerList() - req.addAllReleases(releasedContainerList) - - if (numWorkers > 0) { - logInfo("Allocating %d worker containers with %d of memory each.".format(numWorkers, - workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD)) - } - else { - logDebug("Empty allocation req .. release : " + releasedContainerList) - } - - for (request <- resourceRequests) { - logInfo("ResourceRequest (host : %s, num containers: %d, priority = %s , capability : %s)". - format( - request.getHostName, - request.getNumContainers, - request.getPriority, - request.getCapability)) - } - resourceManager.allocate(req) - } - - - private def createResourceRequest( - requestType: AllocationType.AllocationType, - resource:String, - numWorkers: Int, - priority: Int): ResourceRequest = { - - // If hostname specified, we need atleast two requests - node local and rack local. - // There must be a third request - which is ANY : that will be specially handled. - requestType match { - case AllocationType.HOST => { - assert(YarnAllocationHandler.ANY_HOST != resource) - val hostname = resource - val nodeLocal = createResourceRequestImpl(hostname, numWorkers, priority) - - // Add to host->rack mapping - YarnAllocationHandler.populateRackInfo(conf, hostname) - - nodeLocal - } - case AllocationType.RACK => { - val rack = resource - createResourceRequestImpl(rack, numWorkers, priority) - } - case AllocationType.ANY => createResourceRequestImpl( - YarnAllocationHandler.ANY_HOST, numWorkers, priority) - case _ => throw new IllegalArgumentException( - "Unexpected/unsupported request type: " + requestType) - } - } - - private def createResourceRequestImpl( - hostname:String, - numWorkers: Int, - priority: Int): ResourceRequest = { - - val rsrcRequest = Records.newRecord(classOf[ResourceRequest]) - val memCapability = Records.newRecord(classOf[Resource]) - // There probably is some overhead here, let's reserve a bit more memory. - memCapability.setMemory(workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) - rsrcRequest.setCapability(memCapability) - - val pri = Records.newRecord(classOf[Priority]) - pri.setPriority(priority) - rsrcRequest.setPriority(pri) - - rsrcRequest.setHostName(hostname) - - rsrcRequest.setNumContainers(java.lang.Math.max(numWorkers, 0)) - rsrcRequest - } - - def createReleasedContainerList(): ArrayBuffer[ContainerId] = { - - val retval = new ArrayBuffer[ContainerId](1) - // Iterator on COW list ... - for (container <- releasedContainerList.iterator()){ - retval += container - } - // Remove from the original list. - if (! retval.isEmpty) { - releasedContainerList.removeAll(retval) - for (v <- retval) pendingReleaseContainers.put(v, true) - logInfo("Releasing " + retval.size + " containers. pendingReleaseContainers : " + - pendingReleaseContainers) - } - - retval - } -} - -object YarnAllocationHandler { - - val ANY_HOST = "*" - // All requests are issued with same priority : we do not (yet) have any distinction between - // request types (like map/reduce in hadoop for example) - val PRIORITY = 1 - - // Additional memory overhead - in mb - val MEMORY_OVERHEAD = 384 - - // Host to rack map - saved from allocation requests - // We are expecting this not to change. - // Note that it is possible for this to change : and RM will indicate that to us via update - // response to allocate. But we are punting on handling that for now. - private val hostToRack = new ConcurrentHashMap[String, String]() - private val rackToHostSet = new ConcurrentHashMap[String, JSet[String]]() - - - def newAllocator( - conf: Configuration, - resourceManager: AMRMProtocol, - appAttemptId: ApplicationAttemptId, - args: ApplicationMasterArguments, - sparkConf: SparkConf): YarnAllocationHandler = { - - new YarnAllocationHandler( - conf, - resourceManager, - appAttemptId, - args.numWorkers, - args.workerMemory, - args.workerCores, - Map[String, Int](), - Map[String, Int](), - sparkConf) - } - - def newAllocator( - conf: Configuration, - resourceManager: AMRMProtocol, - appAttemptId: ApplicationAttemptId, - args: ApplicationMasterArguments, - map: collection.Map[String, - collection.Set[SplitInfo]], - sparkConf: SparkConf): YarnAllocationHandler = { - - val (hostToCount, rackToCount) = generateNodeToWeight(conf, map) - new YarnAllocationHandler( - conf, - resourceManager, - appAttemptId, - args.numWorkers, - args.workerMemory, - args.workerCores, - hostToCount, - rackToCount, - sparkConf) - } - - def newAllocator( - conf: Configuration, - resourceManager: AMRMProtocol, - appAttemptId: ApplicationAttemptId, - maxWorkers: Int, - workerMemory: Int, - workerCores: Int, - map: collection.Map[String, collection.Set[SplitInfo]], - sparkConf: SparkConf): YarnAllocationHandler = { - - val (hostToCount, rackToCount) = generateNodeToWeight(conf, map) - - new YarnAllocationHandler( - conf, - resourceManager, - appAttemptId, - maxWorkers, - workerMemory, - workerCores, - hostToCount, - rackToCount, - sparkConf) - } - - // A simple method to copy the split info map. - private def generateNodeToWeight( - conf: Configuration, - input: collection.Map[String, collection.Set[SplitInfo]]) : - // host to count, rack to count - (Map[String, Int], Map[String, Int]) = { - - if (input == null) return (Map[String, Int](), Map[String, Int]()) - - val hostToCount = new HashMap[String, Int] - val rackToCount = new HashMap[String, Int] - - for ((host, splits) <- input) { - val hostCount = hostToCount.getOrElse(host, 0) - hostToCount.put(host, hostCount + splits.size) - - val rack = lookupRack(conf, host) - if (rack != null){ - val rackCount = rackToCount.getOrElse(host, 0) - rackToCount.put(host, rackCount + splits.size) - } - } - - (hostToCount.toMap, rackToCount.toMap) - } - - def lookupRack(conf: Configuration, host: String): String = { - if (!hostToRack.contains(host)) populateRackInfo(conf, host) - hostToRack.get(host) - } - - def fetchCachedHostsForRack(rack: String): Option[Set[String]] = { - val set = rackToHostSet.get(rack) - if (set == null) return None - - // No better way to get a Set[String] from JSet ? - val convertedSet: collection.mutable.Set[String] = set - Some(convertedSet.toSet) - } - - def populateRackInfo(conf: Configuration, hostname: String) { - Utils.checkHost(hostname) - - if (!hostToRack.containsKey(hostname)) { - // If there are repeated failures to resolve, all to an ignore list ? - val rackInfo = RackResolver.resolve(conf, hostname) - if (rackInfo != null && rackInfo.getNetworkLocation != null) { - val rack = rackInfo.getNetworkLocation - hostToRack.put(hostname, rack) - if (! rackToHostSet.containsKey(rack)) { - rackToHostSet.putIfAbsent(rack, - Collections.newSetFromMap(new ConcurrentHashMap[String, JBoolean]())) - } - rackToHostSet.get(rack).add(hostname) - - // TODO(harvey): Figure out this comment... - // Since RackResolver caches, we are disabling this for now ... - } /* else { - // right ? Else we will keep calling rack resolver in case we cant resolve rack info ... - hostToRack.put(hostname, null) - } */ - } - } -} diff --git a/yarn/2.2/pom.xml b/yarn/2.2/pom.xml deleted file mode 100644 index 345462a2c0..0000000000 --- a/yarn/2.2/pom.xml +++ /dev/null @@ -1,112 +0,0 @@ - - - - 4.0.0 - - org.apache.spark - spark-yarn-aggregator_2.10 - 0.9.0-incubating-SNAPSHOT - ../pom.xml - - - org.apache.spark - spark-yarn-2.2_2.10 - jar - Spark Project YARN 2.2 - http://spark.incubator.apache.org/ - - - - - org.codehaus.mojo - build-helper-maven-plugin - - - add-scala-sources - generate-sources - - add-source - - - - src/main/scala - ../common/src/main/scala - - - - - add-scala-test-sources - generate-test-sources - - add-test-source - - - - src/test/scala - ../common/src/test/scala - - - - - - - org.apache.maven.plugins - maven-antrun-plugin - - - test - - run - - - true - - - - - - - - - - - - - - - - - - - - org.scalatest - scalatest-maven-plugin - - - ${basedir}/../.. - 1 - ${spark.classpath} - - - - - - target/scala-${scala.binary.version}/classes - target/scala-${scala.binary.version}/test-classes - - diff --git a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala b/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala deleted file mode 100644 index 7c32e0ab9b..0000000000 --- a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala +++ /dev/null @@ -1,428 +0,0 @@ -/* - * 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. - */ - -package org.apache.spark.deploy.yarn - -import java.io.IOException -import java.net.Socket -import java.util.concurrent.CopyOnWriteArrayList -import java.util.concurrent.atomic.{AtomicInteger, AtomicReference} - -import scala.collection.JavaConversions._ - -import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.fs.{FileSystem, Path} -import org.apache.hadoop.net.NetUtils -import org.apache.hadoop.security.UserGroupInformation -import org.apache.hadoop.util.ShutdownHookManager -import org.apache.hadoop.yarn.api._ -import org.apache.hadoop.yarn.api.protocolrecords._ -import org.apache.hadoop.yarn.api.records._ -import org.apache.hadoop.yarn.client.api.AMRMClient -import org.apache.hadoop.yarn.client.api.AMRMClient.ContainerRequest -import org.apache.hadoop.yarn.conf.YarnConfiguration -import org.apache.hadoop.yarn.ipc.YarnRPC -import org.apache.hadoop.yarn.util.{ConverterUtils, Records} - -import org.apache.spark.{SparkConf, SparkContext, Logging} -import org.apache.spark.util.Utils - - -class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) extends Logging { - - def this(args: ApplicationMasterArguments) = this(args, new Configuration()) - - private var rpc: YarnRPC = YarnRPC.create(conf) - private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) - private var appAttemptId: ApplicationAttemptId = _ - private var userThread: Thread = _ - private val fs = FileSystem.get(yarnConf) - - private var yarnAllocator: YarnAllocationHandler = _ - private var isFinished: Boolean = false - private var uiAddress: String = _ - private val maxAppAttempts: Int = conf.getInt( - YarnConfiguration.RM_AM_MAX_ATTEMPTS, YarnConfiguration.DEFAULT_RM_AM_MAX_ATTEMPTS) - private var isLastAMRetry: Boolean = true - private var amClient: AMRMClient[ContainerRequest] = _ - - private val sparkConf = new SparkConf() - // Default to numWorkers * 2, with minimum of 3 - private val maxNumWorkerFailures = sparkConf.getInt("spark.yarn.max.worker.failures", - math.max(args.numWorkers * 2, 3)) - - def run() { - // Setup the directories so things go to YARN approved directories rather - // than user specified and /tmp. - System.setProperty("spark.local.dir", getLocalDirs()) - - // set the web ui port to be ephemeral for yarn so we don't conflict with - // other spark processes running on the same box - System.setProperty("spark.ui.port", "0") - - // Use priority 30 as it's higher then HDFS. It's same priority as MapReduce is using. - ShutdownHookManager.get().addShutdownHook(new AppMasterShutdownHook(this), 30) - - appAttemptId = getApplicationAttemptId() - isLastAMRetry = appAttemptId.getAttemptId() >= maxAppAttempts - amClient = AMRMClient.createAMRMClient() - amClient.init(yarnConf) - amClient.start() - - // Workaround until hadoop moves to something which has - // https://issues.apache.org/jira/browse/HADOOP-8406 - fixed in (2.0.2-alpha but no 0.23 line) - // org.apache.hadoop.io.compress.CompressionCodecFactory.getCodecClasses(conf) - - ApplicationMaster.register(this) - - // Start the user's JAR - userThread = startUserClass() - - // This a bit hacky, but we need to wait until the spark.driver.port property has - // been set by the Thread executing the user class. - waitForSparkContextInitialized() - - // Do this after Spark master is up and SparkContext is created so that we can register UI Url. - val appMasterResponse: RegisterApplicationMasterResponse = registerApplicationMaster() - - // Allocate all containers - allocateWorkers() - - // Wait for the user class to Finish - userThread.join() - - System.exit(0) - } - - /** Get the Yarn approved local directories. */ - private def getLocalDirs(): String = { - // Hadoop 0.23 and 2.x have different Environment variable names for the - // local dirs, so lets check both. We assume one of the 2 is set. - // LOCAL_DIRS => 2.X, YARN_LOCAL_DIRS => 0.23.X - val localDirs = Option(System.getenv("YARN_LOCAL_DIRS")) - .getOrElse(Option(System.getenv("LOCAL_DIRS")) - .getOrElse("")) - - if (localDirs.isEmpty()) { - throw new Exception("Yarn Local dirs can't be empty") - } - localDirs - } - - private def getApplicationAttemptId(): ApplicationAttemptId = { - val envs = System.getenv() - val containerIdString = envs.get(ApplicationConstants.Environment.CONTAINER_ID.name()) - val containerId = ConverterUtils.toContainerId(containerIdString) - val appAttemptId = containerId.getApplicationAttemptId() - logInfo("ApplicationAttemptId: " + appAttemptId) - appAttemptId - } - - private def registerApplicationMaster(): RegisterApplicationMasterResponse = { - logInfo("Registering the ApplicationMaster") - amClient.registerApplicationMaster(Utils.localHostName(), 0, uiAddress) - } - - private def startUserClass(): Thread = { - logInfo("Starting the user JAR in a separate Thread") - val mainMethod = Class.forName( - args.userClass, - false /* initialize */, - Thread.currentThread.getContextClassLoader).getMethod("main", classOf[Array[String]]) - val t = new Thread { - override def run() { - 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 { - logDebug("finishing main") - isLastAMRetry = true - if (successed) { - ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) - } else { - ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.FAILED) - } - } - } - } - t.start() - t - } - - // This need to happen before allocateWorkers() - private def waitForSparkContextInitialized() { - logInfo("Waiting for Spark context initialization") - try { - var sparkContext: SparkContext = null - ApplicationMaster.sparkContextRef.synchronized { - var numTries = 0 - val waitTime = 10000L - val maxNumTries = sparkConf.getInt("spark.yarn.applicationMaster.waitTries", 10) - while (ApplicationMaster.sparkContextRef.get() == null && numTries < maxNumTries) { - logInfo("Waiting for Spark context initialization ... " + numTries) - numTries = numTries + 1 - ApplicationMaster.sparkContextRef.wait(waitTime) - } - sparkContext = ApplicationMaster.sparkContextRef.get() - assert(sparkContext != null || numTries >= maxNumTries) - - if (sparkContext != null) { - uiAddress = sparkContext.ui.appUIAddress - this.yarnAllocator = YarnAllocationHandler.newAllocator( - yarnConf, - amClient, - appAttemptId, - args, - sparkContext.preferredNodeLocationData, - sparkContext.getConf) - } else { - logWarning("Unable to retrieve SparkContext inspite of waiting for %d, maxNumTries = %d". - format(numTries * waitTime, maxNumTries)) - this.yarnAllocator = YarnAllocationHandler.newAllocator( - yarnConf, - amClient, - appAttemptId, - args, - sparkContext.getConf) - } - } - } finally { - // In case of exceptions, etc - ensure that count is at least ALLOCATOR_LOOP_WAIT_COUNT : - // so that the loop (in ApplicationMaster.sparkContextInitialized) breaks. - ApplicationMaster.incrementAllocatorLoop(ApplicationMaster.ALLOCATOR_LOOP_WAIT_COUNT) - } - } - - private def allocateWorkers() { - try { - logInfo("Allocating " + args.numWorkers + " workers.") - // Wait until all containers have finished - // TODO: This is a bit ugly. Can we make it nicer? - // TODO: Handle container failure - yarnAllocator.addResourceRequests(args.numWorkers) - // Exits the loop if the user thread exits. - while (yarnAllocator.getNumWorkersRunning < args.numWorkers && userThread.isAlive) { - if (yarnAllocator.getNumWorkersFailed >= maxNumWorkerFailures) { - finishApplicationMaster(FinalApplicationStatus.FAILED, - "max number of worker failures reached") - } - yarnAllocator.allocateResources() - ApplicationMaster.incrementAllocatorLoop(1) - Thread.sleep(100) - } - } finally { - // In case of exceptions, etc - ensure that count is at least ALLOCATOR_LOOP_WAIT_COUNT, - // so that the loop in ApplicationMaster#sparkContextInitialized() breaks. - ApplicationMaster.incrementAllocatorLoop(ApplicationMaster.ALLOCATOR_LOOP_WAIT_COUNT) - } - logInfo("All workers have launched.") - - // Launch a progress reporter thread, else the app will get killed after expiration - // (def: 10mins) timeout. - if (userThread.isAlive) { - // Ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapses. - val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) - - // we want to be reasonably responsive without causing too many requests to RM. - val schedulerInterval = - sparkConf.getLong("spark.yarn.scheduler.heartbeat.interval-ms", 5000) - - - // must be <= timeoutInterval / 2. - val interval = math.min(timeoutInterval / 2, schedulerInterval) - - launchReporterThread(interval) - } - } - - private def launchReporterThread(_sleepTime: Long): Thread = { - val sleepTime = if (_sleepTime <= 0 ) 0 else _sleepTime - - val t = new Thread { - override def run() { - while (userThread.isAlive) { - if (yarnAllocator.getNumWorkersFailed >= maxNumWorkerFailures) { - finishApplicationMaster(FinalApplicationStatus.FAILED, - "max number of worker failures reached") - } - val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning - - yarnAllocator.getNumPendingAllocate - if (missingWorkerCount > 0) { - logInfo("Allocating %d containers to make up for (potentially) lost containers". - format(missingWorkerCount)) - yarnAllocator.addResourceRequests(missingWorkerCount) - } - sendProgress() - Thread.sleep(sleepTime) - } - } - } - // Setting to daemon status, though this is usually not a good idea. - t.setDaemon(true) - t.start() - logInfo("Started progress reporter thread - sleep time : " + sleepTime) - t - } - - private def sendProgress() { - logDebug("Sending progress") - // Simulated with an allocate request with no nodes requested. - yarnAllocator.allocateResources() - } - - /* - def printContainers(containers: List[Container]) = { - for (container <- containers) { - logInfo("Launching shell command on a new container." - + ", containerId=" + container.getId() - + ", containerNode=" + container.getNodeId().getHost() - + ":" + container.getNodeId().getPort() - + ", containerNodeURI=" + container.getNodeHttpAddress() - + ", containerState" + container.getState() - + ", containerResourceMemory" - + container.getResource().getMemory()) - } - } - */ - - def finishApplicationMaster(status: FinalApplicationStatus, diagnostics: String = "") { - synchronized { - if (isFinished) { - return - } - isFinished = true - } - - logInfo("finishApplicationMaster with " + status) - // Set tracking URL to empty since we don't have a history server. - amClient.unregisterApplicationMaster(status, "" /* appMessage */, "" /* appTrackingUrl */) - } - - /** - * Clean up the staging directory. - */ - private def cleanupStagingDir() { - var stagingDirPath: Path = null - try { - val preserveFiles = sparkConf.get("spark.yarn.preserve.staging.files", "false").toBoolean - if (!preserveFiles) { - stagingDirPath = new Path(System.getenv("SPARK_YARN_STAGING_DIR")) - if (stagingDirPath == null) { - logError("Staging directory is null") - return - } - logInfo("Deleting staging directory " + stagingDirPath) - fs.delete(stagingDirPath, true) - } - } catch { - case ioe: IOException => - logError("Failed to cleanup staging dir " + stagingDirPath, ioe) - } - } - - // The shutdown hook that runs when a signal is received AND during normal close of the JVM. - class AppMasterShutdownHook(appMaster: ApplicationMaster) extends Runnable { - - def run() { - logInfo("AppMaster received a signal.") - // we need to clean up staging dir before HDFS is shut down - // make sure we don't delete it until this is the last AM - if (appMaster.isLastAMRetry) appMaster.cleanupStagingDir() - } - } -} - -object ApplicationMaster { - // Number of times to wait for the allocator loop to complete. - // Each loop iteration waits for 100ms, so maximum of 3 seconds. - // This is to ensure that we have reasonable number of containers before we start - // TODO: Currently, task to container is computed once (TaskSetManager) - which need not be - // optimal as more containers are available. Might need to handle this better. - private val ALLOCATOR_LOOP_WAIT_COUNT = 30 - - private val applicationMasters = new CopyOnWriteArrayList[ApplicationMaster]() - - val sparkContextRef: AtomicReference[SparkContext] = - new AtomicReference[SparkContext](null /* initialValue */) - - val yarnAllocatorLoop: AtomicInteger = new AtomicInteger(0) - - def incrementAllocatorLoop(by: Int) { - val count = yarnAllocatorLoop.getAndAdd(by) - if (count >= ALLOCATOR_LOOP_WAIT_COUNT) { - yarnAllocatorLoop.synchronized { - // to wake threads off wait ... - yarnAllocatorLoop.notifyAll() - } - } - } - - def register(master: ApplicationMaster) { - applicationMasters.add(master) - } - - // TODO(harvey): See whether this should be discarded - it isn't used anywhere atm... - def sparkContextInitialized(sc: SparkContext): Boolean = { - var modified = false - sparkContextRef.synchronized { - modified = sparkContextRef.compareAndSet(null, sc) - sparkContextRef.notifyAll() - } - - // Add a shutdown hook - as a best case effort in case users do not call sc.stop or do - // System.exit. - // Should not really have to do this, but it helps YARN to evict resources earlier. - // Not to mention, prevent the Client from declaring failure even though we exited properly. - // Note that this will unfortunately not properly clean up the staging files because it gets - // called too late, after the filesystem is already shutdown. - if (modified) { - Runtime.getRuntime().addShutdownHook(new Thread with Logging { - // This is not only logs, but also ensures that log system is initialized for this instance - // when we are actually 'run'-ing. - logInfo("Adding shutdown hook for context " + sc) - override def run() { - logInfo("Invoking sc stop from shutdown hook") - sc.stop() - // Best case ... - 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) { - yarnAllocatorLoop.wait(1000L) - } - } - modified - } - - def main(argStrings: Array[String]) { - val args = new ApplicationMasterArguments(argStrings) - new ApplicationMaster(args).run() - } -} diff --git a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/Client.scala b/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/Client.scala deleted file mode 100644 index a75066888c..0000000000 --- a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/Client.scala +++ /dev/null @@ -1,523 +0,0 @@ -/* - * 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. - */ - -package org.apache.spark.deploy.yarn - -import java.net.{InetAddress, UnknownHostException, URI} -import java.nio.ByteBuffer - -import scala.collection.JavaConversions._ -import scala.collection.mutable.HashMap -import scala.collection.mutable.Map - -import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.fs.{FileContext, FileStatus, FileSystem, Path, FileUtil} -import org.apache.hadoop.fs.permission.FsPermission; -import org.apache.hadoop.io.DataOutputBuffer -import org.apache.hadoop.mapred.Master -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.protocolrecords._ -import org.apache.hadoop.yarn.api.records._ -import org.apache.hadoop.yarn.client.api.impl.YarnClientImpl -import org.apache.hadoop.yarn.conf.YarnConfiguration -import org.apache.hadoop.yarn.ipc.YarnRPC -import org.apache.hadoop.yarn.util.{Apps, Records} - -import org.apache.spark.{Logging, SparkConf} -import org.apache.spark.util.Utils -import org.apache.spark.deploy.SparkHadoopUtil - - -/** - * The entry point (starting in Client#main() and Client#run()) for launching Spark on YARN. The - * Client submits an application to the global ResourceManager to launch Spark's ApplicationMaster, - * which will launch a Spark master process and negotiate resources throughout its duration. - */ -class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl with Logging { - - var rpc: YarnRPC = YarnRPC.create(conf) - val yarnConf: YarnConfiguration = new YarnConfiguration(conf) - val credentials = UserGroupInformation.getCurrentUser().getCredentials() - private val SPARK_STAGING: String = ".sparkStaging" - private val distCacheMgr = new ClientDistributedCacheManager() - private val sparkConf = new SparkConf - - - // Staging directory is private! -> rwx-------- - val STAGING_DIR_PERMISSION: FsPermission = FsPermission.createImmutable(0700: Short) - // App files are world-wide readable and owner writable -> rw-r--r-- - val APP_FILE_PERMISSION: FsPermission = FsPermission.createImmutable(0644: Short) - - def this(args: ClientArguments) = this(new Configuration(), args) - - def runApp(): ApplicationId = { - validateArgs() - // Initialize and start the client service. - init(yarnConf) - start() - - // Log details about this YARN cluster (e.g, the number of slave machines/NodeManagers). - logClusterResourceDetails() - - // Prepare to submit a request to the ResourcManager (specifically its ApplicationsManager (ASM) - // interface). - - // Get a new client application. - val newApp = super.createApplication() - val newAppResponse = newApp.getNewApplicationResponse() - val appId = newAppResponse.getApplicationId() - - verifyClusterResources(newAppResponse) - - // Set up resource and environment variables. - val appStagingDir = getAppStagingDir(appId) - val localResources = prepareLocalResources(appStagingDir) - val launchEnv = setupLaunchEnv(localResources, appStagingDir) - val amContainer = createContainerLaunchContext(newAppResponse, localResources, launchEnv) - - // Set up an application submission context. - val appContext = newApp.getApplicationSubmissionContext() - appContext.setApplicationName(args.appName) - appContext.setQueue(args.amQueue) - appContext.setAMContainerSpec(amContainer) - - // Memory for the ApplicationMaster. - val memoryResource = Records.newRecord(classOf[Resource]).asInstanceOf[Resource] - memoryResource.setMemory(args.amMemory + YarnAllocationHandler.MEMORY_OVERHEAD) - appContext.setResource(memoryResource) - - // Finally, submit and monitor the application. - submitApp(appContext) - appId - } - - def run() { - val appId = runApp() - monitorApplication(appId) - System.exit(0) - } - - // TODO(harvey): This could just go in ClientArguments. - def validateArgs() = { - Map( - (System.getenv("SPARK_JAR") == null) -> "Error: You must set SPARK_JAR environment variable!", - (args.userJar == null) -> "Error: You must specify a user jar!", - (args.userClass == null) -> "Error: You must specify a user class!", - (args.numWorkers <= 0) -> "Error: You must specify atleast 1 worker!", - (args.amMemory <= YarnAllocationHandler.MEMORY_OVERHEAD) -> ("Error: AM memory size must be" + - "greater than: " + YarnAllocationHandler.MEMORY_OVERHEAD), - (args.workerMemory <= YarnAllocationHandler.MEMORY_OVERHEAD) -> ("Error: Worker memory size" + - "must be greater than: " + YarnAllocationHandler.MEMORY_OVERHEAD.toString) - ).foreach { case(cond, errStr) => - if (cond) { - logError(errStr) - args.printUsageAndExit(1) - } - } - } - - def getAppStagingDir(appId: ApplicationId): String = { - SPARK_STAGING + Path.SEPARATOR + appId.toString() + Path.SEPARATOR - } - - def logClusterResourceDetails() { - val clusterMetrics: YarnClusterMetrics = super.getYarnClusterMetrics - logInfo("Got Cluster metric info from ApplicationsManager (ASM), number of NodeManagers: " + - clusterMetrics.getNumNodeManagers) - - val queueInfo: QueueInfo = super.getQueueInfo(args.amQueue) - logInfo("""Queue info ... queueName: %s, queueCurrentCapacity: %s, queueMaxCapacity: %s, - queueApplicationCount = %s, queueChildQueueCount = %s""".format( - queueInfo.getQueueName, - queueInfo.getCurrentCapacity, - queueInfo.getMaximumCapacity, - queueInfo.getApplications.size, - queueInfo.getChildQueues.size)) - } - - def verifyClusterResources(app: GetNewApplicationResponse) = { - val maxMem = app.getMaximumResourceCapability().getMemory() - logInfo("Max mem capabililty of a single resource in this cluster " + maxMem) - - // If we have requested more then the clusters max for a single resource then exit. - if (args.workerMemory > maxMem) { - logError("Required worker memory (%d MB), is above the max threshold (%d MB) of this cluster.". - format(args.workerMemory, maxMem)) - System.exit(1) - } - val amMem = args.amMemory + YarnAllocationHandler.MEMORY_OVERHEAD - if (amMem > maxMem) { - logError("Required AM memory (%d) is above the max threshold (%d) of this cluster". - format(args.amMemory, maxMem)) - 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. - } - - /** See if two file systems are the same or not. */ - private def compareFs(srcFs: FileSystem, destFs: FileSystem): Boolean = { - val srcUri = srcFs.getUri() - val dstUri = destFs.getUri() - if (srcUri.getScheme() == null) { - return false - } - if (!srcUri.getScheme().equals(dstUri.getScheme())) { - return false - } - var srcHost = srcUri.getHost() - var dstHost = dstUri.getHost() - if ((srcHost != null) && (dstHost != null)) { - try { - srcHost = InetAddress.getByName(srcHost).getCanonicalHostName() - dstHost = InetAddress.getByName(dstHost).getCanonicalHostName() - } catch { - case e: UnknownHostException => - return false - } - if (!srcHost.equals(dstHost)) { - return false - } - } else if (srcHost == null && dstHost != null) { - return false - } else if (srcHost != null && dstHost == null) { - return false - } - //check for ports - if (srcUri.getPort() != dstUri.getPort()) { - return false - } - return true - } - - /** Copy the file into HDFS if needed. */ - private def copyRemoteFile( - dstDir: Path, - originalPath: Path, - replication: Short, - setPerms: Boolean = false): Path = { - val fs = FileSystem.get(conf) - val remoteFs = originalPath.getFileSystem(conf) - var newPath = originalPath - if (! compareFs(remoteFs, fs)) { - newPath = new Path(dstDir, originalPath.getName()) - logInfo("Uploading " + originalPath + " to " + newPath) - FileUtil.copy(remoteFs, originalPath, fs, newPath, false, conf) - fs.setReplication(newPath, replication) - if (setPerms) fs.setPermission(newPath, new FsPermission(APP_FILE_PERMISSION)) - } - // Resolve any symlinks in the URI path so using a "current" symlink to point to a specific - // version shows the specific version in the distributed cache configuration - val qualPath = fs.makeQualified(newPath) - val fc = FileContext.getFileContext(qualPath.toUri(), conf) - val destPath = fc.resolvePath(qualPath) - destPath - } - - def prepareLocalResources(appStagingDir: String): HashMap[String, LocalResource] = { - logInfo("Preparing Local resources") - // Upload Spark and the application JAR to the remote file system if necessary. Add them as - // local resources to the application master. - 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) - } - } - val dst = new Path(fs.getHomeDirectory(), appStagingDir) - val replication = sparkConf.getInt("spark.yarn.submit.file.replication", 3).toShort - - if (UserGroupInformation.isSecurityEnabled()) { - val dstFs = dst.getFileSystem(conf) - dstFs.addDelegationTokens(delegTokenRenewer, credentials) - } - - val localResources = HashMap[String, LocalResource]() - FileSystem.mkdirs(fs, dst, new FsPermission(STAGING_DIR_PERMISSION)) - - val statCache: Map[URI, FileStatus] = HashMap[URI, FileStatus]() - - Map( - Client.SPARK_JAR -> System.getenv("SPARK_JAR"), Client.APP_JAR -> args.userJar, - Client.LOG4J_PROP -> System.getenv("SPARK_LOG4J_CONF") - ).foreach { case(destName, _localPath) => - val localPath: String = if (_localPath != null) _localPath.trim() else "" - if (! localPath.isEmpty()) { - var localURI = new URI(localPath) - // If not specified assume these are in the local filesystem to keep behavior like Hadoop - if (localURI.getScheme() == null) { - localURI = new URI(FileSystem.getLocal(conf).makeQualified(new Path(localPath)).toString) - } - val setPermissions = if (destName.equals(Client.APP_JAR)) true else false - val destPath = copyRemoteFile(dst, new Path(localURI), replication, setPermissions) - distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, - destName, statCache) - } - } - - // Handle jars local to the ApplicationMaster. - if ((args.addJars != null) && (!args.addJars.isEmpty())){ - args.addJars.split(',').foreach { case file: String => - val localURI = new URI(file.trim()) - val localPath = new Path(localURI) - val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) - val destPath = copyRemoteFile(dst, localPath, replication) - // Only add the resource to the Spark ApplicationMaster. - val appMasterOnly = true - distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, - linkname, statCache, appMasterOnly) - } - } - - // Handle any distributed cache files - if ((args.files != null) && (!args.files.isEmpty())){ - args.files.split(',').foreach { case file: String => - val localURI = new URI(file.trim()) - val localPath = new Path(localURI) - val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) - val destPath = copyRemoteFile(dst, localPath, replication) - distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, - linkname, statCache) - } - } - - // Handle any distributed cache archives - if ((args.archives != null) && (!args.archives.isEmpty())) { - args.archives.split(',').foreach { case file:String => - val localURI = new URI(file.trim()) - val localPath = new Path(localURI) - val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) - val destPath = copyRemoteFile(dst, localPath, replication) - distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.ARCHIVE, - linkname, statCache) - } - } - - UserGroupInformation.getCurrentUser().addCredentials(credentials) - localResources - } - - def setupLaunchEnv( - localResources: HashMap[String, LocalResource], - stagingDir: String): HashMap[String, String] = { - logInfo("Setting up the launch environment") - val log4jConfLocalRes = localResources.getOrElse(Client.LOG4J_PROP, null) - - val env = new HashMap[String, String]() - - Client.populateClasspath(yarnConf, log4jConfLocalRes != null, env) - env("SPARK_YARN_MODE") = "true" - env("SPARK_YARN_STAGING_DIR") = stagingDir - - // Set the environment variables to be passed on to the Workers. - distCacheMgr.setDistFilesEnv(env) - distCacheMgr.setDistArchivesEnv(env) - - // Allow users to specify some environment variables. - Apps.setEnvFromInputString(env, System.getenv("SPARK_YARN_USER_ENV")) - - // Add each SPARK_* key to the environment. - System.getenv().filterKeys(_.startsWith("SPARK")).foreach { case (k,v) => env(k) = v } - - env - } - - def userArgsToString(clientArgs: ClientArguments): String = { - val prefix = " --args " - val args = clientArgs.userArgs - val retval = new StringBuilder() - for (arg <- args){ - retval.append(prefix).append(" '").append(arg).append("' ") - } - retval.toString - } - - def createContainerLaunchContext( - newApp: GetNewApplicationResponse, - localResources: HashMap[String, LocalResource], - env: HashMap[String, String]): ContainerLaunchContext = { - logInfo("Setting up container launch context") - val amContainer = Records.newRecord(classOf[ContainerLaunchContext]) - amContainer.setLocalResources(localResources) - amContainer.setEnvironment(env) - - // TODO: Need a replacement for the following code to fix -Xmx? - // val minResMemory: Int = newApp.getMinimumResourceCapability().getMemory() - // var amMemory = ((args.amMemory / minResMemory) * minResMemory) + - // ((if ((args.amMemory % minResMemory) == 0) 0 else minResMemory) - - // YarnAllocationHandler.MEMORY_OVERHEAD) - - // Extra options for the JVM - var JAVA_OPTS = "" - - // Add Xmx for AM memory - JAVA_OPTS += "-Xmx" + args.amMemory + "m" - - val tmpDir = new Path(Environment.PWD.$(), YarnConfiguration.DEFAULT_CONTAINER_TEMP_DIR) - JAVA_OPTS += " -Djava.io.tmpdir=" + tmpDir - - // TODO: Remove once cpuset version is pushed out. - // The context is, default gc for server class machines ends up using all cores to do gc - - // hence if there are multiple containers in same node, Spark GC affects all other containers' - // performance (which can be that of other Spark containers) - // Instead of using this, rely on cpusets by YARN to enforce "proper" Spark behavior in - // multi-tenant environments. Not sure how default Java GC behaves if it is limited to subset - // of cores on a node. - val useConcurrentAndIncrementalGC = env.isDefinedAt("SPARK_USE_CONC_INCR_GC") && - java.lang.Boolean.parseBoolean(env("SPARK_USE_CONC_INCR_GC")) - if (useConcurrentAndIncrementalGC) { - // In our expts, using (default) throughput collector has severe perf ramifications in - // multi-tenant machines - JAVA_OPTS += " -XX:+UseConcMarkSweepGC " - JAVA_OPTS += " -XX:+CMSIncrementalMode " - JAVA_OPTS += " -XX:+CMSIncrementalPacing " - JAVA_OPTS += " -XX:CMSIncrementalDutyCycleMin=0 " - JAVA_OPTS += " -XX:CMSIncrementalDutyCycle=10 " - } - - if (env.isDefinedAt("SPARK_JAVA_OPTS")) { - JAVA_OPTS += " " + env("SPARK_JAVA_OPTS") - } - - // Command for the ApplicationMaster - 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 " + - JAVA_OPTS + - " " + args.amClass + - " --class " + args.userClass + - " --jar " + args.userJar + - userArgsToString(args) + - " --worker-memory " + args.workerMemory + - " --worker-cores " + args.workerCores + - " --num-workers " + args.numWorkers + - " 1> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stdout" + - " 2> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stderr") - - logInfo("Command for starting the Spark ApplicationMaster: " + commands(0)) - amContainer.setCommands(commands) - - // Setup security tokens. - val dob = new DataOutputBuffer() - credentials.writeTokenStorageToStream(dob) - amContainer.setTokens(ByteBuffer.wrap(dob.getData())) - - amContainer - } - - def submitApp(appContext: ApplicationSubmissionContext) = { - // Submit the application to the applications manager. - logInfo("Submitting application to ASM") - super.submitApplication(appContext) - } - - def monitorApplication(appId: ApplicationId): Boolean = { - val interval = sparkConf.getLong("spark.yarn.report.interval", 1000) - - while (true) { - Thread.sleep(interval) - val report = super.getApplicationReport(appId) - - logInfo("Application report from ASM: \n" + - "\t application identifier: " + appId.toString() + "\n" + - "\t appId: " + appId.getId() + "\n" + - "\t clientToAMToken: " + report.getClientToAMToken() + "\n" + - "\t appDiagnostics: " + report.getDiagnostics() + "\n" + - "\t appMasterHost: " + report.getHost() + "\n" + - "\t appQueue: " + report.getQueue() + "\n" + - "\t appMasterRpcPort: " + report.getRpcPort() + "\n" + - "\t appStartTime: " + report.getStartTime() + "\n" + - "\t yarnAppState: " + report.getYarnApplicationState() + "\n" + - "\t distributedFinalState: " + report.getFinalApplicationStatus() + "\n" + - "\t appTrackingUrl: " + report.getTrackingUrl() + "\n" + - "\t appUser: " + report.getUser() - ) - - val state = report.getYarnApplicationState() - val dsStatus = report.getFinalApplicationStatus() - if (state == YarnApplicationState.FINISHED || - state == YarnApplicationState.FAILED || - state == YarnApplicationState.KILLED) { - return true - } - } - true - } -} - -object Client { - val SPARK_JAR: String = "spark.jar" - val APP_JAR: String = "app.jar" - val LOG4J_PROP: String = "log4j.properties" - - def main(argStrings: Array[String]) { - // Set an env variable indicating we are running in YARN mode. - // Note: anything env variable with SPARK_ prefix gets propagated to all (remote) processes - - // see Client#setupLaunchEnv(). - System.setProperty("SPARK_YARN_MODE", "true") - - val args = new ClientArguments(argStrings) - - (new Client(args)).run() - } - - // Based on code from org.apache.hadoop.mapreduce.v2.util.MRApps - def populateHadoopClasspath(conf: Configuration, env: HashMap[String, String]) { - for (c <- conf.getStrings(YarnConfiguration.YARN_APPLICATION_CLASSPATH)) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, c.trim) - } - } - - def populateClasspath(conf: Configuration, addLog4j: Boolean, env: HashMap[String, String]) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$()) - // If log4j present, ensure ours overrides all others - if (addLog4j) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + LOG4J_PROP) - } - // Normally the users app.jar is last in case conflicts with spark jars - val userClasspathFirst = new SparkConf().get("spark.yarn.user.classpath.first", "false") - .toBoolean - if (userClasspathFirst) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + APP_JAR) - } - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + SPARK_JAR) - Client.populateHadoopClasspath(conf, env) - - if (!userClasspathFirst) { - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + APP_JAR) - } - Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + - Path.SEPARATOR + "*") - } -} diff --git a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala b/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala deleted file mode 100644 index 4d9cca0766..0000000000 --- a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala +++ /dev/null @@ -1,227 +0,0 @@ -/* - * 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. - */ - -package org.apache.spark.deploy.yarn - -import java.net.Socket -import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.net.NetUtils -import org.apache.hadoop.yarn.api._ -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.util.{ConverterUtils, Records} -import akka.actor._ -import akka.remote._ -import akka.actor.Terminated -import org.apache.spark.{SparkConf, SparkContext, Logging} -import org.apache.spark.util.{Utils, AkkaUtils} -import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend -import org.apache.spark.scheduler.SplitInfo -import org.apache.hadoop.yarn.client.api.AMRMClient -import org.apache.hadoop.yarn.client.api.AMRMClient.ContainerRequest - -class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration) extends Logging { - - def this(args: ApplicationMasterArguments) = this(args, new Configuration()) - - private var appAttemptId: ApplicationAttemptId = _ - private var reporterThread: Thread = _ - private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) - - private var yarnAllocator: YarnAllocationHandler = _ - private var driverClosed:Boolean = false - - private var amClient: AMRMClient[ContainerRequest] = _ - private val sparkConf = new SparkConf - - val actorSystem : ActorSystem = AkkaUtils.createActorSystem("sparkYarnAM", Utils.localHostName, 0, - conf = sparkConf)._1 - var actor: ActorRef = _ - - // This actor just working as a monitor to watch on Driver Actor. - class MonitorActor(driverUrl: String) extends Actor { - - var driver: ActorSelection = _ - - override def preStart() { - logInfo("Listen to driver: " + driverUrl) - driver = context.actorSelection(driverUrl) - // Send a hello message thus the connection is actually established, thus we can monitor Lifecycle Events. - driver ! "Hello" - context.system.eventStream.subscribe(self, classOf[RemotingLifecycleEvent]) - } - - override def receive = { - case x: DisassociatedEvent => - logInfo("Driver terminated or disconnected! Shutting down. $x") - driverClosed = true - } - } - - def run() { - - amClient = AMRMClient.createAMRMClient() - amClient.init(yarnConf) - amClient.start() - - appAttemptId = getApplicationAttemptId() - registerApplicationMaster() - - waitForSparkMaster() - - // Allocate all containers - allocateWorkers() - - // Launch a progress reporter thread, else app will get killed after expiration (def: 10mins) timeout - // ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapse. - - val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) - // must be <= timeoutInterval/ 2. - // On other hand, also ensure that we are reasonably responsive without causing too many requests to RM. - // so atleast 1 minute or timeoutInterval / 10 - whichever is higher. - val interval = math.min(timeoutInterval / 2, math.max(timeoutInterval/ 10, 60000L)) - reporterThread = launchReporterThread(interval) - - // Wait for the reporter thread to Finish. - reporterThread.join() - - finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) - actorSystem.shutdown() - - logInfo("Exited") - System.exit(0) - } - - private def getApplicationAttemptId(): ApplicationAttemptId = { - val envs = System.getenv() - val containerIdString = envs.get(ApplicationConstants.Environment.CONTAINER_ID.name()) - val containerId = ConverterUtils.toContainerId(containerIdString) - val appAttemptId = containerId.getApplicationAttemptId() - logInfo("ApplicationAttemptId: " + appAttemptId) - appAttemptId - } - - private def registerApplicationMaster(): RegisterApplicationMasterResponse = { - logInfo("Registering the ApplicationMaster") - // TODO:(Raymond) Find out Spark UI address and fill in here? - amClient.registerApplicationMaster(Utils.localHostName(), 0, "") - } - - private def waitForSparkMaster() { - logInfo("Waiting for Spark driver to be reachable.") - var driverUp = false - val hostport = args.userArgs(0) - val (driverHost, driverPort) = Utils.parseHostPort(hostport) - while(!driverUp) { - try { - val socket = new Socket(driverHost, driverPort) - socket.close() - logInfo("Driver now available: %s:%s".format(driverHost, driverPort)) - driverUp = true - } catch { - case e: Exception => - logError("Failed to connect to driver at %s:%s, retrying ...". - format(driverHost, driverPort)) - Thread.sleep(100) - } - } - sparkConf.set("spark.driver.host", driverHost) - sparkConf.set("spark.driver.port", driverPort.toString) - - val driverUrl = "akka.tcp://spark@%s:%s/user/%s".format( - driverHost, driverPort.toString, CoarseGrainedSchedulerBackend.ACTOR_NAME) - - actor = actorSystem.actorOf(Props(new MonitorActor(driverUrl)), name = "YarnAM") - } - - - private def allocateWorkers() { - - // Fixme: should get preferredNodeLocationData from SparkContext, just fake a empty one for now. - val preferredNodeLocationData: scala.collection.Map[String, scala.collection.Set[SplitInfo]] = - scala.collection.immutable.Map() - - yarnAllocator = YarnAllocationHandler.newAllocator( - yarnConf, - amClient, - appAttemptId, - args, - preferredNodeLocationData, - sparkConf) - - logInfo("Allocating " + args.numWorkers + " workers.") - // Wait until all containers have finished - // TODO: This is a bit ugly. Can we make it nicer? - // TODO: Handle container failure - - yarnAllocator.addResourceRequests(args.numWorkers) - while(yarnAllocator.getNumWorkersRunning < args.numWorkers) { - yarnAllocator.allocateResources() - Thread.sleep(100) - } - - logInfo("All workers have launched.") - - } - - // TODO: We might want to extend this to allocate more containers in case they die ! - private def launchReporterThread(_sleepTime: Long): Thread = { - val sleepTime = if (_sleepTime <= 0 ) 0 else _sleepTime - - val t = new Thread { - override def run() { - while (!driverClosed) { - val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning - - yarnAllocator.getNumPendingAllocate - if (missingWorkerCount > 0) { - logInfo("Allocating %d containers to make up for (potentially) lost containers". - format(missingWorkerCount)) - yarnAllocator.addResourceRequests(missingWorkerCount) - } - sendProgress() - Thread.sleep(sleepTime) - } - } - } - // setting to daemon status, though this is usually not a good idea. - t.setDaemon(true) - t.start() - logInfo("Started progress reporter thread - sleep time : " + sleepTime) - t - } - - private def sendProgress() { - logDebug("Sending progress") - // simulated with an allocate request with no nodes requested ... - yarnAllocator.allocateResources() - } - - def finishApplicationMaster(status: FinalApplicationStatus) { - logInfo("finish ApplicationMaster with " + status) - amClient.unregisterApplicationMaster(status, "" /* appMessage */, "" /* appTrackingUrl */) - } - -} - - -object WorkerLauncher { - def main(argStrings: Array[String]) { - val args = new ApplicationMasterArguments(argStrings) - new WorkerLauncher(args).run() - } -} diff --git a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala b/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala deleted file mode 100644 index 9f5523c4b9..0000000000 --- a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala +++ /dev/null @@ -1,209 +0,0 @@ -/* - * 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. - */ - -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.records.impl.pb.ProtoUtils -import org.apache.hadoop.yarn.api.protocolrecords._ -import org.apache.hadoop.yarn.client.api.NMClient -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.spark.Logging - - -class WorkerRunnable( - container: Container, - conf: Configuration, - masterAddress: String, - slaveId: String, - hostname: String, - workerMemory: Int, - workerCores: Int) - extends Runnable with Logging { - - var rpc: YarnRPC = YarnRPC.create(conf) - var nmClient: NMClient = _ - val yarnConf: YarnConfiguration = new YarnConfiguration(conf) - - def run = { - logInfo("Starting Worker Container") - nmClient = NMClient.createNMClient() - nmClient.init(yarnConf) - nmClient.start() - startContainer - } - - def startContainer = { - logInfo("Setting up ContainerLaunchContext") - - val ctx = Records.newRecord(classOf[ContainerLaunchContext]) - .asInstanceOf[ContainerLaunchContext] - - val localResources = prepareLocalResources - ctx.setLocalResources(localResources) - - val env = prepareEnvironment - ctx.setEnvironment(env) - - // 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 " - } -*/ - - val credentials = UserGroupInformation.getCurrentUser().getCredentials() - val dob = new DataOutputBuffer() - credentials.writeTokenStorageToStream(dob) - ctx.setTokens(ByteBuffer.wrap(dob.getData())) - - 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") - logInfo("Setting up worker with commands: " + commands) - ctx.setCommands(commands) - - // Send the start request to the ContainerManager - nmClient.startContainer(container, ctx) - } - - 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]() - - Client.populateClasspath(yarnConf, 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 - } - -} diff --git a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala b/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala deleted file mode 100644 index 8a9a73f5b4..0000000000 --- a/yarn/2.2/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala +++ /dev/null @@ -1,694 +0,0 @@ -/* - * 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. - */ - -package org.apache.spark.deploy.yarn - -import java.lang.{Boolean => JBoolean} -import java.util.{Collections, Set => JSet} -import java.util.concurrent.{CopyOnWriteArrayList, ConcurrentHashMap} -import java.util.concurrent.atomic.AtomicInteger - -import scala.collection -import scala.collection.JavaConversions._ -import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} - -import org.apache.spark.{Logging, SparkConf} -import org.apache.spark.scheduler.{SplitInfo,TaskSchedulerImpl} -import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend -import org.apache.spark.util.Utils - -import org.apache.hadoop.conf.Configuration -import org.apache.hadoop.yarn.api.ApplicationMasterProtocol -import org.apache.hadoop.yarn.api.records.ApplicationAttemptId -import org.apache.hadoop.yarn.api.records.{Container, ContainerId, ContainerStatus} -import org.apache.hadoop.yarn.api.records.{Priority, Resource, ResourceRequest} -import org.apache.hadoop.yarn.api.protocolrecords.{AllocateRequest, AllocateResponse} -import org.apache.hadoop.yarn.client.api.AMRMClient -import org.apache.hadoop.yarn.client.api.AMRMClient.ContainerRequest -import org.apache.hadoop.yarn.util.{RackResolver, Records} - - -object AllocationType extends Enumeration { - type AllocationType = Value - val HOST, RACK, ANY = Value -} - -// TODO: -// Too many params. -// Needs to be mt-safe -// Need to refactor this to make it 'cleaner' ... right now, all computation is reactive - should -// make it more proactive and decoupled. - -// Note that right now, we assume all node asks as uniform in terms of capabilities and priority -// Refer to http://developer.yahoo.com/blogs/hadoop/posts/2011/03/mapreduce-nextgen-scheduler/ for -// more info on how we are requesting for containers. -private[yarn] class YarnAllocationHandler( - val conf: Configuration, - val amClient: AMRMClient[ContainerRequest], - val appAttemptId: ApplicationAttemptId, - val maxWorkers: Int, - val workerMemory: Int, - val workerCores: Int, - val preferredHostToCount: Map[String, Int], - val preferredRackToCount: Map[String, Int], - val sparkConf: SparkConf) - extends Logging { - // These three are locked on allocatedHostToContainersMap. Complementary data structures - // allocatedHostToContainersMap : containers which are running : host, Set - // allocatedContainerToHostMap: container to host mapping. - private val allocatedHostToContainersMap = - new HashMap[String, collection.mutable.Set[ContainerId]]() - - private val allocatedContainerToHostMap = new HashMap[ContainerId, String]() - - // allocatedRackCount is populated ONLY if allocation happens (or decremented if this is an - // allocated node) - // As with the two data structures above, tightly coupled with them, and to be locked on - // allocatedHostToContainersMap - private val allocatedRackCount = new HashMap[String, Int]() - - // Containers which have been released. - private val releasedContainerList = new CopyOnWriteArrayList[ContainerId]() - // Containers to be released in next request to RM - private val pendingReleaseContainers = new ConcurrentHashMap[ContainerId, Boolean] - - // Number of container requests that have been sent to, but not yet allocated by the - // ApplicationMaster. - private val numPendingAllocate = new AtomicInteger() - private val numWorkersRunning = new AtomicInteger() - // Used to generate a unique id per worker - private val workerIdCounter = new AtomicInteger() - private val lastResponseId = new AtomicInteger() - private val numWorkersFailed = new AtomicInteger() - - def getNumPendingAllocate: Int = numPendingAllocate.intValue - - def getNumWorkersRunning: Int = numWorkersRunning.intValue - - def getNumWorkersFailed: Int = numWorkersFailed.intValue - - def isResourceConstraintSatisfied(container: Container): Boolean = { - container.getResource.getMemory >= (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) - } - - def releaseContainer(container: Container) { - val containerId = container.getId - pendingReleaseContainers.put(containerId, true) - amClient.releaseAssignedContainer(containerId) - } - - def allocateResources() { - // We have already set the container request. Poll the ResourceManager for a response. - // This doubles as a heartbeat if there are no pending container requests. - val progressIndicator = 0.1f - val allocateResponse = amClient.allocate(progressIndicator) - - val allocatedContainers = allocateResponse.getAllocatedContainers() - if (allocatedContainers.size > 0) { - var numPendingAllocateNow = numPendingAllocate.addAndGet(-1 * allocatedContainers.size) - - if (numPendingAllocateNow < 0) { - numPendingAllocateNow = numPendingAllocate.addAndGet(-1 * numPendingAllocateNow) - } - - logDebug(""" - Allocated containers: %d - Current worker count: %d - Containers released: %s - Containers to-be-released: %s - Cluster resources: %s - """.format( - allocatedContainers.size, - numWorkersRunning.get(), - releasedContainerList, - pendingReleaseContainers, - allocateResponse.getAvailableResources)) - - val hostToContainers = new HashMap[String, ArrayBuffer[Container]]() - - for (container <- allocatedContainers) { - if (isResourceConstraintSatisfied(container)) { - // Add the accepted `container` to the host's list of already accepted, - // allocated containers - val host = container.getNodeId.getHost - val containersForHost = hostToContainers.getOrElseUpdate(host, - new ArrayBuffer[Container]()) - containersForHost += container - } else { - // Release container, since it doesn't satisfy resource constraints. - releaseContainer(container) - } - } - - // Find the appropriate containers to use. - // TODO: Cleanup this group-by... - val dataLocalContainers = new HashMap[String, ArrayBuffer[Container]]() - val rackLocalContainers = new HashMap[String, ArrayBuffer[Container]]() - val offRackContainers = new HashMap[String, ArrayBuffer[Container]]() - - for (candidateHost <- hostToContainers.keySet) { - val maxExpectedHostCount = preferredHostToCount.getOrElse(candidateHost, 0) - val requiredHostCount = maxExpectedHostCount - allocatedContainersOnHost(candidateHost) - - val remainingContainersOpt = hostToContainers.get(candidateHost) - assert(remainingContainersOpt.isDefined) - var remainingContainers = remainingContainersOpt.get - - if (requiredHostCount >= remainingContainers.size) { - // Since we have <= required containers, add all remaining containers to - // `dataLocalContainers`. - dataLocalContainers.put(candidateHost, remainingContainers) - // There are no more free containers remaining. - remainingContainers = null - } else if (requiredHostCount > 0) { - // Container list has more containers than we need for data locality. - // Split the list into two: one based on the data local container count, - // (`remainingContainers.size` - `requiredHostCount`), and the other to hold remaining - // containers. - val (dataLocal, remaining) = remainingContainers.splitAt( - remainingContainers.size - requiredHostCount) - dataLocalContainers.put(candidateHost, dataLocal) - - // Invariant: remainingContainers == remaining - - // YARN has a nasty habit of allocating a ton of containers on a host - discourage this. - // Add each container in `remaining` to list of containers to release. If we have an - // insufficient number of containers, then the next allocation cycle will reallocate - // (but won't treat it as data local). - // TODO(harvey): Rephrase this comment some more. - for (container <- remaining) releaseContainer(container) - remainingContainers = null - } - - // For rack local containers - if (remainingContainers != null) { - val rack = YarnAllocationHandler.lookupRack(conf, candidateHost) - if (rack != null) { - val maxExpectedRackCount = preferredRackToCount.getOrElse(rack, 0) - val requiredRackCount = maxExpectedRackCount - allocatedContainersOnRack(rack) - - rackLocalContainers.getOrElse(rack, List()).size - - if (requiredRackCount >= remainingContainers.size) { - // Add all remaining containers to to `dataLocalContainers`. - dataLocalContainers.put(rack, remainingContainers) - remainingContainers = null - } else if (requiredRackCount > 0) { - // Container list has more containers that we need for data locality. - // Split the list into two: one based on the data local container count, - // (`remainingContainers.size` - `requiredHostCount`), and the other to hold remaining - // containers. - val (rackLocal, remaining) = remainingContainers.splitAt( - remainingContainers.size - requiredRackCount) - val existingRackLocal = rackLocalContainers.getOrElseUpdate(rack, - new ArrayBuffer[Container]()) - - existingRackLocal ++= rackLocal - - remainingContainers = remaining - } - } - } - - if (remainingContainers != null) { - // Not all containers have been consumed - add them to the list of off-rack containers. - offRackContainers.put(candidateHost, remainingContainers) - } - } - - // Now that we have split the containers into various groups, go through them in order: - // first host-local, then rack-local, and finally off-rack. - // Note that the list we create below tries to ensure that not all containers end up within - // a host if there is a sufficiently large number of hosts/containers. - val allocatedContainersToProcess = new ArrayBuffer[Container](allocatedContainers.size) - allocatedContainersToProcess ++= TaskSchedulerImpl.prioritizeContainers(dataLocalContainers) - allocatedContainersToProcess ++= TaskSchedulerImpl.prioritizeContainers(rackLocalContainers) - allocatedContainersToProcess ++= TaskSchedulerImpl.prioritizeContainers(offRackContainers) - - // Run each of the allocated containers. - for (container <- allocatedContainersToProcess) { - val numWorkersRunningNow = numWorkersRunning.incrementAndGet() - val workerHostname = container.getNodeId.getHost - val containerId = container.getId - - val workerMemoryOverhead = (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) - assert(container.getResource.getMemory >= workerMemoryOverhead) - - if (numWorkersRunningNow > maxWorkers) { - logInfo("""Ignoring container %s at host %s, since we already have the required number of - containers for it.""".format(containerId, workerHostname)) - releaseContainer(container) - numWorkersRunning.decrementAndGet() - } else { - val workerId = workerIdCounter.incrementAndGet().toString - val driverUrl = "akka.tcp://spark@%s:%s/user/%s".format( - sparkConf.get("spark.driver.host"), - sparkConf.get("spark.driver.port"), - CoarseGrainedSchedulerBackend.ACTOR_NAME) - - logInfo("Launching container %s for on host %s".format(containerId, workerHostname)) - - // To be safe, remove the container from `pendingReleaseContainers`. - pendingReleaseContainers.remove(containerId) - - val rack = YarnAllocationHandler.lookupRack(conf, workerHostname) - allocatedHostToContainersMap.synchronized { - val containerSet = allocatedHostToContainersMap.getOrElseUpdate(workerHostname, - new HashSet[ContainerId]()) - - containerSet += containerId - allocatedContainerToHostMap.put(containerId, workerHostname) - - if (rack != null) { - allocatedRackCount.put(rack, allocatedRackCount.getOrElse(rack, 0) + 1) - } - } - logInfo("Launching WorkerRunnable. driverUrl: %s, workerHostname: %s".format(driverUrl, workerHostname)) - val workerRunnable = new WorkerRunnable( - container, - conf, - driverUrl, - workerId, - workerHostname, - workerMemory, - workerCores) - new Thread(workerRunnable).start() - } - } - logDebug(""" - Finished allocating %s containers (from %s originally). - Current number of workers running: %d, - releasedContainerList: %s, - pendingReleaseContainers: %s - """.format( - allocatedContainersToProcess, - allocatedContainers, - numWorkersRunning.get(), - releasedContainerList, - pendingReleaseContainers)) - } - - val completedContainers = allocateResponse.getCompletedContainersStatuses() - if (completedContainers.size > 0) { - logDebug("Completed %d containers".format(completedContainers.size)) - - for (completedContainer <- completedContainers) { - val containerId = completedContainer.getContainerId - - if (pendingReleaseContainers.containsKey(containerId)) { - // YarnAllocationHandler already marked the container for release, so remove it from - // `pendingReleaseContainers`. - pendingReleaseContainers.remove(containerId) - } else { - // Decrement the number of workers running. The next iteration of the ApplicationMaster's - // reporting thread will take care of allocating. - numWorkersRunning.decrementAndGet() - logInfo("Completed container %s (state: %s, exit status: %s)".format( - containerId, - completedContainer.getState, - completedContainer.getExitStatus())) - // Hadoop 2.2.X added a ContainerExitStatus we should switch to use - // there are some exit status' we shouldn't necessarily count against us, but for - // now I think its ok as none of the containers are expected to exit - if (completedContainer.getExitStatus() != 0) { - logInfo("Container marked as failed: " + containerId) - numWorkersFailed.incrementAndGet() - } - } - - allocatedHostToContainersMap.synchronized { - if (allocatedContainerToHostMap.containsKey(containerId)) { - val hostOpt = allocatedContainerToHostMap.get(containerId) - assert(hostOpt.isDefined) - val host = hostOpt.get - - val containerSetOpt = allocatedHostToContainersMap.get(host) - assert(containerSetOpt.isDefined) - val containerSet = containerSetOpt.get - - containerSet.remove(containerId) - if (containerSet.isEmpty) { - allocatedHostToContainersMap.remove(host) - } else { - allocatedHostToContainersMap.update(host, containerSet) - } - - allocatedContainerToHostMap.remove(containerId) - - // TODO: Move this part outside the synchronized block? - val rack = YarnAllocationHandler.lookupRack(conf, host) - if (rack != null) { - val rackCount = allocatedRackCount.getOrElse(rack, 0) - 1 - if (rackCount > 0) { - allocatedRackCount.put(rack, rackCount) - } else { - allocatedRackCount.remove(rack) - } - } - } - } - } - logDebug(""" - Finished processing %d completed containers. - Current number of workers running: %d, - releasedContainerList: %s, - pendingReleaseContainers: %s - """.format( - completedContainers.size, - numWorkersRunning.get(), - releasedContainerList, - pendingReleaseContainers)) - } - } - - def createRackResourceRequests( - hostContainers: ArrayBuffer[ContainerRequest] - ): ArrayBuffer[ContainerRequest] = { - // Generate modified racks and new set of hosts under it before issuing requests. - val rackToCounts = new HashMap[String, Int]() - - for (container <- hostContainers) { - val candidateHost = container.getNodes.last - assert(YarnAllocationHandler.ANY_HOST != candidateHost) - - val rack = YarnAllocationHandler.lookupRack(conf, candidateHost) - if (rack != null) { - var count = rackToCounts.getOrElse(rack, 0) - count += 1 - rackToCounts.put(rack, count) - } - } - - val requestedContainers = new ArrayBuffer[ContainerRequest](rackToCounts.size) - for ((rack, count) <- rackToCounts) { - requestedContainers ++= createResourceRequests( - AllocationType.RACK, - rack, - count, - YarnAllocationHandler.PRIORITY) - } - - requestedContainers - } - - def allocatedContainersOnHost(host: String): Int = { - var retval = 0 - allocatedHostToContainersMap.synchronized { - retval = allocatedHostToContainersMap.getOrElse(host, Set()).size - } - retval - } - - def allocatedContainersOnRack(rack: String): Int = { - var retval = 0 - allocatedHostToContainersMap.synchronized { - retval = allocatedRackCount.getOrElse(rack, 0) - } - retval - } - - def addResourceRequests(numWorkers: Int) { - val containerRequests: List[ContainerRequest] = - if (numWorkers <= 0 || preferredHostToCount.isEmpty) { - logDebug("numWorkers: " + numWorkers + ", host preferences: " + - preferredHostToCount.isEmpty) - createResourceRequests( - AllocationType.ANY, - resource = null, - numWorkers, - YarnAllocationHandler.PRIORITY).toList - } else { - // Request for all hosts in preferred nodes and for numWorkers - - // candidates.size, request by default allocation policy. - val hostContainerRequests = new ArrayBuffer[ContainerRequest](preferredHostToCount.size) - for ((candidateHost, candidateCount) <- preferredHostToCount) { - val requiredCount = candidateCount - allocatedContainersOnHost(candidateHost) - - if (requiredCount > 0) { - hostContainerRequests ++= createResourceRequests( - AllocationType.HOST, - candidateHost, - requiredCount, - YarnAllocationHandler.PRIORITY) - } - } - val rackContainerRequests: List[ContainerRequest] = createRackResourceRequests( - hostContainerRequests).toList - - val anyContainerRequests = createResourceRequests( - AllocationType.ANY, - resource = null, - numWorkers, - YarnAllocationHandler.PRIORITY) - - val containerRequestBuffer = new ArrayBuffer[ContainerRequest]( - hostContainerRequests.size + rackContainerRequests.size() + anyContainerRequests.size) - - containerRequestBuffer ++= hostContainerRequests - containerRequestBuffer ++= rackContainerRequests - containerRequestBuffer ++= anyContainerRequests - containerRequestBuffer.toList - } - - for (request <- containerRequests) { - amClient.addContainerRequest(request) - } - - if (numWorkers > 0) { - numPendingAllocate.addAndGet(numWorkers) - logInfo("Will Allocate %d worker containers, each with %d memory".format( - numWorkers, - (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD))) - } else { - logDebug("Empty allocation request ...") - } - - for (request <- containerRequests) { - val nodes = request.getNodes - var hostStr = if (nodes == null || nodes.isEmpty) { - "Any" - } else { - nodes.last - } - logInfo("Container request (host: %s, priority: %s, capability: %s".format( - hostStr, - request.getPriority().getPriority, - request.getCapability)) - } - } - - private def createResourceRequests( - requestType: AllocationType.AllocationType, - resource: String, - numWorkers: Int, - priority: Int - ): ArrayBuffer[ContainerRequest] = { - - // If hostname is specified, then we need at least two requests - node local and rack local. - // There must be a third request, which is ANY. That will be specially handled. - requestType match { - case AllocationType.HOST => { - assert(YarnAllocationHandler.ANY_HOST != resource) - val hostname = resource - val nodeLocal = constructContainerRequests( - Array(hostname), - racks = null, - numWorkers, - priority) - - // Add `hostname` to the global (singleton) host->rack mapping in YarnAllocationHandler. - YarnAllocationHandler.populateRackInfo(conf, hostname) - nodeLocal - } - case AllocationType.RACK => { - val rack = resource - constructContainerRequests(hosts = null, Array(rack), numWorkers, priority) - } - case AllocationType.ANY => constructContainerRequests( - hosts = null, racks = null, numWorkers, priority) - case _ => throw new IllegalArgumentException( - "Unexpected/unsupported request type: " + requestType) - } - } - - private def constructContainerRequests( - hosts: Array[String], - racks: Array[String], - numWorkers: Int, - priority: Int - ): ArrayBuffer[ContainerRequest] = { - - val memoryResource = Records.newRecord(classOf[Resource]) - memoryResource.setMemory(workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) - - val prioritySetting = Records.newRecord(classOf[Priority]) - prioritySetting.setPriority(priority) - - val requests = new ArrayBuffer[ContainerRequest]() - for (i <- 0 until numWorkers) { - requests += new ContainerRequest(memoryResource, hosts, racks, prioritySetting) - } - requests - } -} - -object YarnAllocationHandler { - - val ANY_HOST = "*" - // All requests are issued with same priority : we do not (yet) have any distinction between - // request types (like map/reduce in hadoop for example) - val PRIORITY = 1 - - // Additional memory overhead - in mb. - val MEMORY_OVERHEAD = 384 - - // Host to rack map - saved from allocation requests. We are expecting this not to change. - // Note that it is possible for this to change : and ResurceManager will indicate that to us via - // update response to allocate. But we are punting on handling that for now. - private val hostToRack = new ConcurrentHashMap[String, String]() - private val rackToHostSet = new ConcurrentHashMap[String, JSet[String]]() - - - def newAllocator( - conf: Configuration, - amClient: AMRMClient[ContainerRequest], - appAttemptId: ApplicationAttemptId, - args: ApplicationMasterArguments, - sparkConf: SparkConf - ): YarnAllocationHandler = { - new YarnAllocationHandler( - conf, - amClient, - appAttemptId, - args.numWorkers, - args.workerMemory, - args.workerCores, - Map[String, Int](), - Map[String, Int](), - sparkConf) - } - - def newAllocator( - conf: Configuration, - amClient: AMRMClient[ContainerRequest], - appAttemptId: ApplicationAttemptId, - args: ApplicationMasterArguments, - map: collection.Map[String, - collection.Set[SplitInfo]], - sparkConf: SparkConf - ): YarnAllocationHandler = { - val (hostToSplitCount, rackToSplitCount) = generateNodeToWeight(conf, map) - new YarnAllocationHandler( - conf, - amClient, - appAttemptId, - args.numWorkers, - args.workerMemory, - args.workerCores, - hostToSplitCount, - rackToSplitCount, - sparkConf) - } - - def newAllocator( - conf: Configuration, - amClient: AMRMClient[ContainerRequest], - appAttemptId: ApplicationAttemptId, - maxWorkers: Int, - workerMemory: Int, - workerCores: Int, - map: collection.Map[String, collection.Set[SplitInfo]], - sparkConf: SparkConf - ): YarnAllocationHandler = { - val (hostToCount, rackToCount) = generateNodeToWeight(conf, map) - new YarnAllocationHandler( - conf, - amClient, - appAttemptId, - maxWorkers, - workerMemory, - workerCores, - hostToCount, - rackToCount, - sparkConf) - } - - // A simple method to copy the split info map. - private def generateNodeToWeight( - conf: Configuration, - input: collection.Map[String, collection.Set[SplitInfo]] - ): (Map[String, Int], Map[String, Int]) = { - - if (input == null) { - return (Map[String, Int](), Map[String, Int]()) - } - - val hostToCount = new HashMap[String, Int] - val rackToCount = new HashMap[String, Int] - - for ((host, splits) <- input) { - val hostCount = hostToCount.getOrElse(host, 0) - hostToCount.put(host, hostCount + splits.size) - - val rack = lookupRack(conf, host) - if (rack != null){ - val rackCount = rackToCount.getOrElse(host, 0) - rackToCount.put(host, rackCount + splits.size) - } - } - - (hostToCount.toMap, rackToCount.toMap) - } - - def lookupRack(conf: Configuration, host: String): String = { - if (!hostToRack.contains(host)) { - populateRackInfo(conf, host) - } - hostToRack.get(host) - } - - def fetchCachedHostsForRack(rack: String): Option[Set[String]] = { - Option(rackToHostSet.get(rack)).map { set => - val convertedSet: collection.mutable.Set[String] = set - // TODO: Better way to get a Set[String] from JSet. - convertedSet.toSet - } - } - - def populateRackInfo(conf: Configuration, hostname: String) { - Utils.checkHost(hostname) - - if (!hostToRack.containsKey(hostname)) { - // If there are repeated failures to resolve, all to an ignore list. - val rackInfo = RackResolver.resolve(conf, hostname) - if (rackInfo != null && rackInfo.getNetworkLocation != null) { - val rack = rackInfo.getNetworkLocation - hostToRack.put(hostname, rack) - if (! rackToHostSet.containsKey(rack)) { - rackToHostSet.putIfAbsent(rack, - Collections.newSetFromMap(new ConcurrentHashMap[String, JBoolean]())) - } - rackToHostSet.get(rack).add(hostname) - - // TODO(harvey): Figure out what this comment means... - // Since RackResolver caches, we are disabling this for now ... - } /* else { - // right ? Else we will keep calling rack resolver in case we cant resolve rack info ... - hostToRack.put(hostname, null) - } */ - } - } -} diff --git a/yarn/README.md b/yarn/README.md index 0d63bd9352..9a7a1dd838 100644 --- a/yarn/README.md +++ b/yarn/README.md @@ -4,9 +4,9 @@ Hadoop Yarn related codes are organized in separate directories for easy managem * common : Common codes that do not depending on specific version of Hadoop. - * 2.0 / 2.2 : Codes that involve specific version of Hadoop YARN API. + * alpha / stable : Codes that involve specific version of Hadoop YARN API. - 2.0 actually represents 0.23 and 2.0 - 2.2 actually represents 2.2 and later, until the API is break again. + alpha represents 0.23 and 2.0.x + stable represents 2.2 and later, until the API is break again. -2.0 / 2.2 will build together with common dir into a single jar +alpha / stable will build together with common dir into a single jar diff --git a/yarn/alpha/pom.xml b/yarn/alpha/pom.xml new file mode 100644 index 0000000000..94812f8c45 --- /dev/null +++ b/yarn/alpha/pom.xml @@ -0,0 +1,112 @@ + + + + 4.0.0 + + org.apache.spark + spark-yarn-aggregator_2.10 + 0.9.0-incubating-SNAPSHOT + ../pom.xml + + + org.apache.spark + spark-yarn-alpha_2.10 + jar + Spark Project YARN Alpha API + http://spark.incubator.apache.org/ + + + + + org.codehaus.mojo + build-helper-maven-plugin + + + add-scala-sources + generate-sources + + add-source + + + + src/main/scala + ../common/src/main/scala + + + + + add-scala-test-sources + generate-test-sources + + add-test-source + + + + src/test/scala + ../common/src/test/scala + + + + + + + org.apache.maven.plugins + maven-antrun-plugin + + + test + + run + + + true + + + + + + + + + + + + + + + + + + + + org.scalatest + scalatest-maven-plugin + + + ${basedir}/../.. + 1 + ${spark.classpath} + + + + + + target/scala-${scala.binary.version}/classes + target/scala-${scala.binary.version}/test-classes + + diff --git a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala new file mode 100644 index 0000000000..7cf120d3eb --- /dev/null +++ b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala @@ -0,0 +1,458 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.io.IOException +import java.net.Socket +import java.util.concurrent.CopyOnWriteArrayList +import java.util.concurrent.atomic.{AtomicInteger, AtomicReference} + +import scala.collection.JavaConversions._ + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.fs.{FileSystem, Path} +import org.apache.hadoop.net.NetUtils +import org.apache.hadoop.security.UserGroupInformation +import org.apache.hadoop.util.ShutdownHookManager +import org.apache.hadoop.yarn.api._ +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.{ConverterUtils, Records} + +import org.apache.spark.{SparkConf, SparkContext, Logging} +import org.apache.spark.util.Utils + +class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) extends Logging { + + def this(args: ApplicationMasterArguments) = this(args, new Configuration()) + + private var rpc: YarnRPC = YarnRPC.create(conf) + private var resourceManager: AMRMProtocol = _ + private var appAttemptId: ApplicationAttemptId = _ + private var userThread: Thread = _ + private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + private val fs = FileSystem.get(yarnConf) + + private var yarnAllocator: YarnAllocationHandler = _ + private var isFinished: Boolean = false + private var uiAddress: String = _ + private val maxAppAttempts: Int = conf.getInt(YarnConfiguration.RM_AM_MAX_RETRIES, + YarnConfiguration.DEFAULT_RM_AM_MAX_RETRIES) + private var isLastAMRetry: Boolean = true + + private val sparkConf = new SparkConf() + // Default to numWorkers * 2, with minimum of 3 + private val maxNumWorkerFailures = sparkConf.getInt("spark.yarn.max.worker.failures", + math.max(args.numWorkers * 2, 3)) + + def run() { + // Setup the directories so things go to yarn approved directories rather + // then user specified and /tmp. + System.setProperty("spark.local.dir", getLocalDirs()) + + // set the web ui port to be ephemeral for yarn so we don't conflict with + // other spark processes running on the same box + System.setProperty("spark.ui.port", "0") + + // Use priority 30 as its higher then HDFS. Its same priority as MapReduce is using. + ShutdownHookManager.get().addShutdownHook(new AppMasterShutdownHook(this), 30) + + appAttemptId = getApplicationAttemptId() + isLastAMRetry = appAttemptId.getAttemptId() >= maxAppAttempts + resourceManager = registerWithResourceManager() + + // Workaround until hadoop moves to something which has + // https://issues.apache.org/jira/browse/HADOOP-8406 - fixed in (2.0.2-alpha but no 0.23 line) + // ignore result. + // This does not, unfortunately, always work reliably ... but alleviates the bug a lot of times + // Hence args.workerCores = numCore disabled above. Any better option? + + // Compute number of threads for akka + //val minimumMemory = appMasterResponse.getMinimumResourceCapability().getMemory() + //if (minimumMemory > 0) { + // val mem = args.workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD + // val numCore = (mem / minimumMemory) + (if (0 != (mem % minimumMemory)) 1 else 0) + + // if (numCore > 0) { + // do not override - hits https://issues.apache.org/jira/browse/HADOOP-8406 + // TODO: Uncomment when hadoop is on a version which has this fixed. + // args.workerCores = numCore + // } + //} + // org.apache.hadoop.io.compress.CompressionCodecFactory.getCodecClasses(conf) + + ApplicationMaster.register(this) + // Start the user's JAR + userThread = startUserClass() + + // This a bit hacky, but we need to wait until the spark.driver.port property has + // been set by the Thread executing the user class. + waitForSparkContextInitialized() + + // Do this after spark master is up and SparkContext is created so that we can register UI Url + val appMasterResponse: RegisterApplicationMasterResponse = registerApplicationMaster() + + // Allocate all containers + allocateWorkers() + + // Wait for the user class to Finish + userThread.join() + + System.exit(0) + } + + /** Get the Yarn approved local directories. */ + private def getLocalDirs(): String = { + // Hadoop 0.23 and 2.x have different Environment variable names for the + // local dirs, so lets check both. We assume one of the 2 is set. + // LOCAL_DIRS => 2.X, YARN_LOCAL_DIRS => 0.23.X + val localDirs = Option(System.getenv("YARN_LOCAL_DIRS")) + .getOrElse(Option(System.getenv("LOCAL_DIRS")) + .getOrElse("")) + + if (localDirs.isEmpty()) { + throw new Exception("Yarn Local dirs can't be empty") + } + localDirs + } + + private def getApplicationAttemptId(): ApplicationAttemptId = { + val envs = System.getenv() + val containerIdString = envs.get(ApplicationConstants.AM_CONTAINER_ID_ENV) + val containerId = ConverterUtils.toContainerId(containerIdString) + val appAttemptId = containerId.getApplicationAttemptId() + logInfo("ApplicationAttemptId: " + appAttemptId) + appAttemptId + } + + private def registerWithResourceManager(): AMRMProtocol = { + val rmAddress = NetUtils.createSocketAddr(yarnConf.get( + YarnConfiguration.RM_SCHEDULER_ADDRESS, + YarnConfiguration.DEFAULT_RM_SCHEDULER_ADDRESS)) + logInfo("Connecting to ResourceManager at " + rmAddress) + rpc.getProxy(classOf[AMRMProtocol], rmAddress, conf).asInstanceOf[AMRMProtocol] + } + + private def registerApplicationMaster(): RegisterApplicationMasterResponse = { + logInfo("Registering the ApplicationMaster") + val appMasterRequest = Records.newRecord(classOf[RegisterApplicationMasterRequest]) + .asInstanceOf[RegisterApplicationMasterRequest] + appMasterRequest.setApplicationAttemptId(appAttemptId) + // Setting this to master host,port - so that the ApplicationReport at client has some + // sensible info. + // Users can then monitor stderr/stdout on that node if required. + appMasterRequest.setHost(Utils.localHostName()) + appMasterRequest.setRpcPort(0) + appMasterRequest.setTrackingUrl(uiAddress) + resourceManager.registerApplicationMaster(appMasterRequest) + } + + private def startUserClass(): Thread = { + logInfo("Starting the user JAR in a separate Thread") + val mainMethod = Class.forName( + args.userClass, + false /* initialize */, + Thread.currentThread.getContextClassLoader).getMethod("main", classOf[Array[String]]) + val t = new Thread { + override def run() { + 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 { + logDebug("finishing main") + isLastAMRetry = true + if (successed) { + ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) + } else { + ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.FAILED) + } + } + } + } + t.start() + t + } + + // this need to happen before allocateWorkers + private def waitForSparkContextInitialized() { + logInfo("Waiting for spark context initialization") + try { + var sparkContext: SparkContext = null + ApplicationMaster.sparkContextRef.synchronized { + var count = 0 + val waitTime = 10000L + val numTries = sparkConf.getInt("spark.yarn.ApplicationMaster.waitTries", 10) + while (ApplicationMaster.sparkContextRef.get() == null && count < numTries) { + logInfo("Waiting for spark context initialization ... " + count) + count = count + 1 + ApplicationMaster.sparkContextRef.wait(waitTime) + } + sparkContext = ApplicationMaster.sparkContextRef.get() + assert(sparkContext != null || count >= numTries) + + if (null != sparkContext) { + uiAddress = sparkContext.ui.appUIAddress + this.yarnAllocator = YarnAllocationHandler.newAllocator( + yarnConf, + resourceManager, + appAttemptId, + args, + sparkContext.preferredNodeLocationData, + sparkContext.getConf) + } else { + logWarning("Unable to retrieve sparkContext inspite of waiting for %d, numTries = %d". + format(count * waitTime, numTries)) + this.yarnAllocator = YarnAllocationHandler.newAllocator( + yarnConf, + resourceManager, + appAttemptId, + args, + sparkContext.getConf) + } + } + } finally { + // in case of exceptions, etc - ensure that count is atleast ALLOCATOR_LOOP_WAIT_COUNT : + // so that the loop (in ApplicationMaster.sparkContextInitialized) breaks + ApplicationMaster.incrementAllocatorLoop(ApplicationMaster.ALLOCATOR_LOOP_WAIT_COUNT) + } + } + + private def allocateWorkers() { + try { + logInfo("Allocating " + args.numWorkers + " workers.") + // Wait until all containers have finished + // TODO: This is a bit ugly. Can we make it nicer? + // TODO: Handle container failure + + // Exists the loop if the user thread exits. + while (yarnAllocator.getNumWorkersRunning < args.numWorkers && userThread.isAlive) { + if (yarnAllocator.getNumWorkersFailed >= maxNumWorkerFailures) { + finishApplicationMaster(FinalApplicationStatus.FAILED, + "max number of worker failures reached") + } + yarnAllocator.allocateContainers( + math.max(args.numWorkers - yarnAllocator.getNumWorkersRunning, 0)) + ApplicationMaster.incrementAllocatorLoop(1) + Thread.sleep(100) + } + } finally { + // In case of exceptions, etc - ensure that count is at least ALLOCATOR_LOOP_WAIT_COUNT, + // so that the loop in ApplicationMaster#sparkContextInitialized() breaks. + ApplicationMaster.incrementAllocatorLoop(ApplicationMaster.ALLOCATOR_LOOP_WAIT_COUNT) + } + logInfo("All workers have launched.") + + // Launch a progress reporter thread, else the app will get killed after expiration + // (def: 10mins) timeout. + // TODO(harvey): Verify the timeout + if (userThread.isAlive) { + // Ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapses. + val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) + + // we want to be reasonably responsive without causing too many requests to RM. + val schedulerInterval = + sparkConf.getLong("spark.yarn.scheduler.heartbeat.interval-ms", 5000) + + // must be <= timeoutInterval / 2. + val interval = math.min(timeoutInterval / 2, schedulerInterval) + + launchReporterThread(interval) + } + } + + private def launchReporterThread(_sleepTime: Long): Thread = { + val sleepTime = if (_sleepTime <= 0 ) 0 else _sleepTime + + val t = new Thread { + override def run() { + while (userThread.isAlive) { + if (yarnAllocator.getNumWorkersFailed >= maxNumWorkerFailures) { + finishApplicationMaster(FinalApplicationStatus.FAILED, + "max number of worker failures reached") + } + val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning + if (missingWorkerCount > 0) { + logInfo("Allocating %d containers to make up for (potentially) lost containers". + format(missingWorkerCount)) + yarnAllocator.allocateContainers(missingWorkerCount) + } + else sendProgress() + Thread.sleep(sleepTime) + } + } + } + // Setting to daemon status, though this is usually not a good idea. + t.setDaemon(true) + t.start() + logInfo("Started progress reporter thread - sleep time : " + sleepTime) + t + } + + private def sendProgress() { + logDebug("Sending progress") + // Simulated with an allocate request with no nodes requested ... + yarnAllocator.allocateContainers(0) + } + + /* + def printContainers(containers: List[Container]) = { + for (container <- containers) { + logInfo("Launching shell command on a new container." + + ", containerId=" + container.getId() + + ", containerNode=" + container.getNodeId().getHost() + + ":" + container.getNodeId().getPort() + + ", containerNodeURI=" + container.getNodeHttpAddress() + + ", containerState" + container.getState() + + ", containerResourceMemory" + + container.getResource().getMemory()) + } + } + */ + + def finishApplicationMaster(status: FinalApplicationStatus, diagnostics: String = "") { + synchronized { + if (isFinished) { + return + } + isFinished = true + } + + logInfo("finishApplicationMaster with " + status) + val finishReq = Records.newRecord(classOf[FinishApplicationMasterRequest]) + .asInstanceOf[FinishApplicationMasterRequest] + finishReq.setAppAttemptId(appAttemptId) + finishReq.setFinishApplicationStatus(status) + finishReq.setDiagnostics(diagnostics) + // Set tracking url to empty since we don't have a history server. + finishReq.setTrackingUrl("") + resourceManager.finishApplicationMaster(finishReq) + } + + /** + * Clean up the staging directory. + */ + private def cleanupStagingDir() { + var stagingDirPath: Path = null + try { + val preserveFiles = sparkConf.get("spark.yarn.preserve.staging.files", "false").toBoolean + if (!preserveFiles) { + stagingDirPath = new Path(System.getenv("SPARK_YARN_STAGING_DIR")) + if (stagingDirPath == null) { + logError("Staging directory is null") + return + } + logInfo("Deleting staging directory " + stagingDirPath) + fs.delete(stagingDirPath, true) + } + } catch { + case ioe: IOException => + logError("Failed to cleanup staging dir " + stagingDirPath, ioe) + } + } + + // The shutdown hook that runs when a signal is received AND during normal close of the JVM. + class AppMasterShutdownHook(appMaster: ApplicationMaster) extends Runnable { + + def run() { + logInfo("AppMaster received a signal.") + // we need to clean up staging dir before HDFS is shut down + // make sure we don't delete it until this is the last AM + if (appMaster.isLastAMRetry) appMaster.cleanupStagingDir() + } + } +} + +object ApplicationMaster { + // Number of times to wait for the allocator loop to complete. + // Each loop iteration waits for 100ms, so maximum of 3 seconds. + // This is to ensure that we have reasonable number of containers before we start + // TODO: Currently, task to container is computed once (TaskSetManager) - which need not be + // optimal as more containers are available. Might need to handle this better. + private val ALLOCATOR_LOOP_WAIT_COUNT = 30 + def incrementAllocatorLoop(by: Int) { + val count = yarnAllocatorLoop.getAndAdd(by) + if (count >= ALLOCATOR_LOOP_WAIT_COUNT) { + yarnAllocatorLoop.synchronized { + // to wake threads off wait ... + yarnAllocatorLoop.notifyAll() + } + } + } + + private val applicationMasters = new CopyOnWriteArrayList[ApplicationMaster]() + + def register(master: ApplicationMaster) { + applicationMasters.add(master) + } + + val sparkContextRef: AtomicReference[SparkContext] = + new AtomicReference[SparkContext](null /* initialValue */) + val yarnAllocatorLoop: AtomicInteger = new AtomicInteger(0) + + def sparkContextInitialized(sc: SparkContext): Boolean = { + var modified = false + sparkContextRef.synchronized { + modified = sparkContextRef.compareAndSet(null, sc) + sparkContextRef.notifyAll() + } + + // Add a shutdown hook - as a best case effort in case users do not call sc.stop or do + // System.exit. + // Should not really have to do this, but it helps YARN to evict resources earlier. + // Not to mention, prevent the Client from declaring failure even though we exited properly. + // Note that this will unfortunately not properly clean up the staging files because it gets + // called too late, after the filesystem is already shutdown. + if (modified) { + Runtime.getRuntime().addShutdownHook(new Thread with Logging { + // This is not only logs, but also ensures that log system is initialized for this instance + // when we are actually 'run'-ing. + logInfo("Adding shutdown hook for context " + sc) + override def run() { + logInfo("Invoking sc stop from shutdown hook") + sc.stop() + // Best case ... + 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) { + yarnAllocatorLoop.wait(1000L) + } + } + modified + } + + def main(argStrings: Array[String]) { + val args = new ApplicationMasterArguments(argStrings) + new ApplicationMaster(args).run() + } +} diff --git a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/Client.scala b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/Client.scala new file mode 100644 index 0000000000..2bd047c97a --- /dev/null +++ b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/Client.scala @@ -0,0 +1,505 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.{InetAddress, UnknownHostException, URI} +import java.nio.ByteBuffer + +import scala.collection.JavaConversions._ +import scala.collection.mutable.HashMap +import scala.collection.mutable.Map + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.fs.{FileContext, FileStatus, FileSystem, Path, FileUtil} +import org.apache.hadoop.fs.permission.FsPermission; +import org.apache.hadoop.io.DataOutputBuffer +import org.apache.hadoop.mapred.Master +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.protocolrecords._ +import org.apache.hadoop.yarn.api.records._ +import org.apache.hadoop.yarn.client.YarnClientImpl +import org.apache.hadoop.yarn.conf.YarnConfiguration +import org.apache.hadoop.yarn.ipc.YarnRPC +import org.apache.hadoop.yarn.util.{Apps, Records} + +import org.apache.spark.{Logging, SparkConf} +import org.apache.spark.util.Utils +import org.apache.spark.deploy.SparkHadoopUtil + + +class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl with Logging { + + def this(args: ClientArguments) = this(new Configuration(), args) + + var rpc: YarnRPC = YarnRPC.create(conf) + val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + val credentials = UserGroupInformation.getCurrentUser().getCredentials() + private val SPARK_STAGING: String = ".sparkStaging" + private val distCacheMgr = new ClientDistributedCacheManager() + private val sparkConf = new SparkConf + + // Staging directory is private! -> rwx-------- + val STAGING_DIR_PERMISSION: FsPermission = FsPermission.createImmutable(0700:Short) + + // App files are world-wide readable and owner writable -> rw-r--r-- + val APP_FILE_PERMISSION: FsPermission = FsPermission.createImmutable(0644:Short) + + // for client user who want to monitor app status by itself. + def runApp() = { + validateArgs() + + init(yarnConf) + start() + logClusterResourceDetails() + + val newApp = super.getNewApplication() + val appId = newApp.getApplicationId() + + verifyClusterResources(newApp) + val appContext = createApplicationSubmissionContext(appId) + val appStagingDir = getAppStagingDir(appId) + val localResources = prepareLocalResources(appStagingDir) + val env = setupLaunchEnv(localResources, appStagingDir) + val amContainer = createContainerLaunchContext(newApp, localResources, env) + + appContext.setQueue(args.amQueue) + appContext.setAMContainerSpec(amContainer) + appContext.setUser(UserGroupInformation.getCurrentUser().getShortUserName()) + + submitApp(appContext) + appId + } + + def run() { + val appId = runApp() + monitorApplication(appId) + System.exit(0) + } + + def validateArgs() = { + Map( + (System.getenv("SPARK_JAR") == null) -> "Error: You must set SPARK_JAR environment variable!", + (args.userJar == null) -> "Error: You must specify a user jar!", + (args.userClass == null) -> "Error: You must specify a user class!", + (args.numWorkers <= 0) -> "Error: You must specify atleast 1 worker!", + (args.amMemory <= YarnAllocationHandler.MEMORY_OVERHEAD) -> ("Error: AM memory size must be " + + "greater than: " + YarnAllocationHandler.MEMORY_OVERHEAD), + (args.workerMemory <= YarnAllocationHandler.MEMORY_OVERHEAD) -> ("Error: Worker memory size " + + "must be greater than: " + YarnAllocationHandler.MEMORY_OVERHEAD) + ).foreach { case(cond, errStr) => + if (cond) { + logError(errStr) + args.printUsageAndExit(1) + } + } + } + + def getAppStagingDir(appId: ApplicationId): String = { + SPARK_STAGING + Path.SEPARATOR + appId.toString() + Path.SEPARATOR + } + + def logClusterResourceDetails() { + val clusterMetrics: YarnClusterMetrics = super.getYarnClusterMetrics + logInfo("Got Cluster metric info from ASM, numNodeManagers = " + + clusterMetrics.getNumNodeManagers) + + val queueInfo: QueueInfo = super.getQueueInfo(args.amQueue) + logInfo("""Queue info ... queueName = %s, queueCurrentCapacity = %s, queueMaxCapacity = %s, + queueApplicationCount = %s, queueChildQueueCount = %s""".format( + queueInfo.getQueueName, + queueInfo.getCurrentCapacity, + queueInfo.getMaximumCapacity, + queueInfo.getApplications.size, + queueInfo.getChildQueues.size)) + } + + def verifyClusterResources(app: GetNewApplicationResponse) = { + val maxMem = app.getMaximumResourceCapability().getMemory() + logInfo("Max mem capabililty of a single resource in this cluster " + maxMem) + + // 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 = { + logInfo("Setting up application submission context for ASM") + val appContext = Records.newRecord(classOf[ApplicationSubmissionContext]) + appContext.setApplicationId(appId) + appContext.setApplicationName(args.appName) + return appContext + } + + /** See if two file systems are the same or not. */ + private def compareFs(srcFs: FileSystem, destFs: FileSystem): Boolean = { + val srcUri = srcFs.getUri() + val dstUri = destFs.getUri() + if (srcUri.getScheme() == null) { + return false + } + if (!srcUri.getScheme().equals(dstUri.getScheme())) { + return false + } + var srcHost = srcUri.getHost() + var dstHost = dstUri.getHost() + if ((srcHost != null) && (dstHost != null)) { + try { + srcHost = InetAddress.getByName(srcHost).getCanonicalHostName() + dstHost = InetAddress.getByName(dstHost).getCanonicalHostName() + } catch { + case e: UnknownHostException => + return false + } + if (!srcHost.equals(dstHost)) { + return false + } + } else if (srcHost == null && dstHost != null) { + return false + } else if (srcHost != null && dstHost == null) { + return false + } + //check for ports + if (srcUri.getPort() != dstUri.getPort()) { + return false + } + return true + } + + /** Copy the file into HDFS if needed. */ + private def copyRemoteFile( + dstDir: Path, + originalPath: Path, + replication: Short, + setPerms: Boolean = false): Path = { + val fs = FileSystem.get(conf) + val remoteFs = originalPath.getFileSystem(conf) + var newPath = originalPath + if (! compareFs(remoteFs, fs)) { + newPath = new Path(dstDir, originalPath.getName()) + logInfo("Uploading " + originalPath + " to " + newPath) + FileUtil.copy(remoteFs, originalPath, fs, newPath, false, conf) + fs.setReplication(newPath, replication) + if (setPerms) fs.setPermission(newPath, new FsPermission(APP_FILE_PERMISSION)) + } + // Resolve any symlinks in the URI path so using a "current" symlink to point to a specific + // version shows the specific version in the distributed cache configuration + val qualPath = fs.makeQualified(newPath) + val fc = FileContext.getFileContext(qualPath.toUri(), conf) + val destPath = fc.resolvePath(qualPath) + destPath + } + + def prepareLocalResources(appStagingDir: String): HashMap[String, LocalResource] = { + logInfo("Preparing Local resources") + // Upload Spark and the application JAR to the remote file system if necessary. 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) + } + } + val dst = new Path(fs.getHomeDirectory(), appStagingDir) + val replication = sparkConf.getInt("spark.yarn.submit.file.replication", 3).toShort + + if (UserGroupInformation.isSecurityEnabled()) { + val dstFs = dst.getFileSystem(conf) + dstFs.addDelegationTokens(delegTokenRenewer, credentials) + } + val localResources = HashMap[String, LocalResource]() + FileSystem.mkdirs(fs, dst, new FsPermission(STAGING_DIR_PERMISSION)) + + val statCache: Map[URI, FileStatus] = HashMap[URI, FileStatus]() + + Map(Client.SPARK_JAR -> System.getenv("SPARK_JAR"), Client.APP_JAR -> args.userJar, + Client.LOG4J_PROP -> System.getenv("SPARK_LOG4J_CONF")) + .foreach { case(destName, _localPath) => + val localPath: String = if (_localPath != null) _localPath.trim() else "" + if (! localPath.isEmpty()) { + var localURI = new URI(localPath) + // if not specified assume these are in the local filesystem to keep behavior like Hadoop + if (localURI.getScheme() == null) { + localURI = new URI(FileSystem.getLocal(conf).makeQualified(new Path(localPath)).toString) + } + val setPermissions = if (destName.equals(Client.APP_JAR)) true else false + val destPath = copyRemoteFile(dst, new Path(localURI), replication, setPermissions) + distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, + destName, statCache) + } + } + + // handle any add jars + if ((args.addJars != null) && (!args.addJars.isEmpty())){ + args.addJars.split(',').foreach { case file: String => + val localURI = new URI(file.trim()) + val localPath = new Path(localURI) + val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) + val destPath = copyRemoteFile(dst, localPath, replication) + distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, + linkname, statCache, true) + } + } + + // handle any distributed cache files + if ((args.files != null) && (!args.files.isEmpty())){ + args.files.split(',').foreach { case file: String => + val localURI = new URI(file.trim()) + val localPath = new Path(localURI) + val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) + val destPath = copyRemoteFile(dst, localPath, replication) + distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, + linkname, statCache) + } + } + + // handle any distributed cache archives + if ((args.archives != null) && (!args.archives.isEmpty())) { + args.archives.split(',').foreach { case file:String => + val localURI = new URI(file.trim()) + val localPath = new Path(localURI) + val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) + val destPath = copyRemoteFile(dst, localPath, replication) + distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.ARCHIVE, + linkname, statCache) + } + } + + UserGroupInformation.getCurrentUser().addCredentials(credentials) + return localResources + } + + def setupLaunchEnv( + localResources: HashMap[String, LocalResource], + stagingDir: String): HashMap[String, String] = { + logInfo("Setting up the launch environment") + val log4jConfLocalRes = localResources.getOrElse(Client.LOG4J_PROP, null) + + val env = new HashMap[String, String]() + + Client.populateClasspath(yarnConf, log4jConfLocalRes != null, env) + env("SPARK_YARN_MODE") = "true" + env("SPARK_YARN_STAGING_DIR") = stagingDir + + // Set the environment variables to be passed on to the Workers. + distCacheMgr.setDistFilesEnv(env) + distCacheMgr.setDistArchivesEnv(env) + + // Allow users to specify some environment variables. + Apps.setEnvFromInputString(env, System.getenv("SPARK_YARN_USER_ENV")) + + // Add each SPARK-* key to the environment. + System.getenv().filterKeys(_.startsWith("SPARK")).foreach { case (k,v) => env(k) = v } + env + } + + def userArgsToString(clientArgs: ClientArguments): String = { + val prefix = " --args " + val args = clientArgs.userArgs + val retval = new StringBuilder() + for (arg <- args){ + retval.append(prefix).append(" '").append(arg).append("' ") + } + retval.toString + } + + def createContainerLaunchContext( + newApp: GetNewApplicationResponse, + localResources: HashMap[String, LocalResource], + env: HashMap[String, String]): ContainerLaunchContext = { + logInfo("Setting up container launch context") + val amContainer = Records.newRecord(classOf[ContainerLaunchContext]) + amContainer.setLocalResources(localResources) + amContainer.setEnvironment(env) + + val minResMemory: Int = newApp.getMinimumResourceCapability().getMemory() + + // TODO(harvey): This can probably be a val. + var amMemory = ((args.amMemory / minResMemory) * minResMemory) + + ((if ((args.amMemory % minResMemory) == 0) 0 else minResMemory) - + YarnAllocationHandler.MEMORY_OVERHEAD) + + // Extra options for the JVM + var JAVA_OPTS = "" + + // Add Xmx for am memory + JAVA_OPTS += "-Xmx" + amMemory + "m " + + 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. + val useConcurrentAndIncrementalGC = env.isDefinedAt("SPARK_USE_CONC_INCR_GC") && + java.lang.Boolean.parseBoolean(env("SPARK_USE_CONC_INCR_GC")) + if (useConcurrentAndIncrementalGC) { + // In our expts, using (default) throughput collector has severe perf ramnifications in + // multi-tenant machines + JAVA_OPTS += " -XX:+UseConcMarkSweepGC " + JAVA_OPTS += " -XX:+CMSIncrementalMode " + JAVA_OPTS += " -XX:+CMSIncrementalPacing " + JAVA_OPTS += " -XX:CMSIncrementalDutyCycleMin=0 " + JAVA_OPTS += " -XX:CMSIncrementalDutyCycle=10 " + } + + if (env.isDefinedAt("SPARK_JAVA_OPTS")) { + JAVA_OPTS += env("SPARK_JAVA_OPTS") + " " + } + + // Command for the ApplicationMaster + 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 " + + JAVA_OPTS + + " " + args.amClass + + " --class " + args.userClass + + " --jar " + args.userJar + + userArgsToString(args) + + " --worker-memory " + args.workerMemory + + " --worker-cores " + args.workerCores + + " --num-workers " + args.numWorkers + + " 1> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stdout" + + " 2> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stderr") + logInfo("Command for the ApplicationMaster: " + commands(0)) + amContainer.setCommands(commands) + + val capability = Records.newRecord(classOf[Resource]).asInstanceOf[Resource] + // 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())) + + amContainer + } + + def submitApp(appContext: ApplicationSubmissionContext) = { + // Submit the application to the applications manager. + logInfo("Submitting application to ASM") + super.submitApplication(appContext) + } + + def monitorApplication(appId: ApplicationId): Boolean = { + val interval = sparkConf.getLong("spark.yarn.report.interval", 1000) + + while (true) { + Thread.sleep(interval) + val report = super.getApplicationReport(appId) + + logInfo("Application report from ASM: \n" + + "\t application identifier: " + appId.toString() + "\n" + + "\t appId: " + appId.getId() + "\n" + + "\t clientToken: " + report.getClientToken() + "\n" + + "\t appDiagnostics: " + report.getDiagnostics() + "\n" + + "\t appMasterHost: " + report.getHost() + "\n" + + "\t appQueue: " + report.getQueue() + "\n" + + "\t appMasterRpcPort: " + report.getRpcPort() + "\n" + + "\t appStartTime: " + report.getStartTime() + "\n" + + "\t yarnAppState: " + report.getYarnApplicationState() + "\n" + + "\t distributedFinalState: " + report.getFinalApplicationStatus() + "\n" + + "\t appTrackingUrl: " + report.getTrackingUrl() + "\n" + + "\t appUser: " + report.getUser() + ) + + val state = report.getYarnApplicationState() + val dsStatus = report.getFinalApplicationStatus() + if (state == YarnApplicationState.FINISHED || + state == YarnApplicationState.FAILED || + state == YarnApplicationState.KILLED) { + return true + } + } + true + } +} + +object Client { + val SPARK_JAR: String = "spark.jar" + val APP_JAR: String = "app.jar" + val LOG4J_PROP: String = "log4j.properties" + + def main(argStrings: Array[String]) { + // Set an env variable indicating we are running in YARN mode. + // Note that anything with SPARK prefix gets propagated to all (remote) processes + System.setProperty("SPARK_YARN_MODE", "true") + + val args = new ClientArguments(argStrings) + + new Client(args).run + } + + // Based on code from org.apache.hadoop.mapreduce.v2.util.MRApps + def populateHadoopClasspath(conf: Configuration, env: HashMap[String, String]) { + for (c <- conf.getStrings(YarnConfiguration.YARN_APPLICATION_CLASSPATH)) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, c.trim) + } + } + + def populateClasspath(conf: Configuration, addLog4j: Boolean, env: HashMap[String, String]) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$()) + // If log4j present, ensure ours overrides all others + if (addLog4j) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + LOG4J_PROP) + } + // Normally the users app.jar is last in case conflicts with spark jars + val userClasspathFirst = new SparkConf().get("spark.yarn.user.classpath.first", "false").toBoolean + if (userClasspathFirst) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + APP_JAR) + } + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + SPARK_JAR) + Client.populateHadoopClasspath(conf, env) + + if (!userClasspathFirst) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + APP_JAR) + } + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + "*") + } +} diff --git a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala new file mode 100644 index 0000000000..e64530702c --- /dev/null +++ b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala @@ -0,0 +1,248 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.Socket +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.net.NetUtils +import org.apache.hadoop.yarn.api._ +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.{ConverterUtils, Records} +import akka.actor._ +import akka.remote._ +import akka.actor.Terminated +import org.apache.spark.{SparkConf, SparkContext, Logging} +import org.apache.spark.util.{Utils, AkkaUtils} +import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend +import org.apache.spark.scheduler.SplitInfo + +class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration) extends Logging { + + def this(args: ApplicationMasterArguments) = this(args, new Configuration()) + + private val rpc: YarnRPC = YarnRPC.create(conf) + private var resourceManager: AMRMProtocol = _ + private var appAttemptId: ApplicationAttemptId = _ + private var reporterThread: Thread = _ + private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + + private var yarnAllocator: YarnAllocationHandler = _ + private var driverClosed:Boolean = false + private val sparkConf = new SparkConf + + val actorSystem : ActorSystem = AkkaUtils.createActorSystem("sparkYarnAM", Utils.localHostName, 0, + conf = sparkConf)._1 + var actor: ActorRef = _ + + // This actor just working as a monitor to watch on Driver Actor. + class MonitorActor(driverUrl: String) extends Actor { + + var driver: ActorSelection = _ + + override def preStart() { + logInfo("Listen to driver: " + driverUrl) + driver = context.actorSelection(driverUrl) + // Send a hello message thus the connection is actually established, thus we can monitor Lifecycle Events. + driver ! "Hello" + context.system.eventStream.subscribe(self, classOf[RemotingLifecycleEvent]) + } + + override def receive = { + case x: DisassociatedEvent => + logInfo(s"Driver terminated or disconnected! Shutting down. $x") + driverClosed = true + } + } + + def run() { + + appAttemptId = getApplicationAttemptId() + resourceManager = registerWithResourceManager() + val appMasterResponse: RegisterApplicationMasterResponse = registerApplicationMaster() + + // Compute number of threads for akka + val minimumMemory = appMasterResponse.getMinimumResourceCapability().getMemory() + + if (minimumMemory > 0) { + val mem = args.workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD + val numCore = (mem / minimumMemory) + (if (0 != (mem % minimumMemory)) 1 else 0) + + if (numCore > 0) { + // do not override - hits https://issues.apache.org/jira/browse/HADOOP-8406 + // TODO: Uncomment when hadoop is on a version which has this fixed. + // args.workerCores = numCore + } + } + + waitForSparkMaster() + + // Allocate all containers + allocateWorkers() + + // Launch a progress reporter thread, else app will get killed after expiration (def: 10mins) timeout + // ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapse. + + val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) + // must be <= timeoutInterval/ 2. + // On other hand, also ensure that we are reasonably responsive without causing too many requests to RM. + // so atleast 1 minute or timeoutInterval / 10 - whichever is higher. + val interval = math.min(timeoutInterval / 2, math.max(timeoutInterval/ 10, 60000L)) + reporterThread = launchReporterThread(interval) + + // Wait for the reporter thread to Finish. + reporterThread.join() + + finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) + actorSystem.shutdown() + + logInfo("Exited") + System.exit(0) + } + + private def getApplicationAttemptId(): ApplicationAttemptId = { + val envs = System.getenv() + val containerIdString = envs.get(ApplicationConstants.AM_CONTAINER_ID_ENV) + val containerId = ConverterUtils.toContainerId(containerIdString) + val appAttemptId = containerId.getApplicationAttemptId() + logInfo("ApplicationAttemptId: " + appAttemptId) + return appAttemptId + } + + private def registerWithResourceManager(): AMRMProtocol = { + val rmAddress = NetUtils.createSocketAddr(yarnConf.get( + YarnConfiguration.RM_SCHEDULER_ADDRESS, + YarnConfiguration.DEFAULT_RM_SCHEDULER_ADDRESS)) + logInfo("Connecting to ResourceManager at " + rmAddress) + return rpc.getProxy(classOf[AMRMProtocol], rmAddress, conf).asInstanceOf[AMRMProtocol] + } + + private def registerApplicationMaster(): RegisterApplicationMasterResponse = { + logInfo("Registering the ApplicationMaster") + val appMasterRequest = Records.newRecord(classOf[RegisterApplicationMasterRequest]) + .asInstanceOf[RegisterApplicationMasterRequest] + appMasterRequest.setApplicationAttemptId(appAttemptId) + // Setting this to master host,port - so that the ApplicationReport at client has some sensible info. + // Users can then monitor stderr/stdout on that node if required. + appMasterRequest.setHost(Utils.localHostName()) + appMasterRequest.setRpcPort(0) + // What do we provide here ? Might make sense to expose something sensible later ? + appMasterRequest.setTrackingUrl("") + return resourceManager.registerApplicationMaster(appMasterRequest) + } + + private def waitForSparkMaster() { + logInfo("Waiting for spark driver to be reachable.") + var driverUp = false + val hostport = args.userArgs(0) + val (driverHost, driverPort) = Utils.parseHostPort(hostport) + while(!driverUp) { + try { + val socket = new Socket(driverHost, driverPort) + socket.close() + logInfo("Master now available: " + driverHost + ":" + driverPort) + driverUp = true + } catch { + case e: Exception => + logError("Failed to connect to driver at " + driverHost + ":" + driverPort) + Thread.sleep(100) + } + } + sparkConf.set("spark.driver.host", driverHost) + sparkConf.set("spark.driver.port", driverPort.toString) + + val driverUrl = "akka.tcp://spark@%s:%s/user/%s".format( + driverHost, driverPort.toString, CoarseGrainedSchedulerBackend.ACTOR_NAME) + + actor = actorSystem.actorOf(Props(new MonitorActor(driverUrl)), name = "YarnAM") + } + + + private def allocateWorkers() { + + // Fixme: should get preferredNodeLocationData from SparkContext, just fake a empty one for now. + val preferredNodeLocationData: scala.collection.Map[String, scala.collection.Set[SplitInfo]] = + scala.collection.immutable.Map() + + yarnAllocator = YarnAllocationHandler.newAllocator(yarnConf, resourceManager, appAttemptId, + args, preferredNodeLocationData, sparkConf) + + logInfo("Allocating " + args.numWorkers + " workers.") + // Wait until all containers have finished + // TODO: This is a bit ugly. Can we make it nicer? + // TODO: Handle container failure + while(yarnAllocator.getNumWorkersRunning < args.numWorkers) { + yarnAllocator.allocateContainers(math.max(args.numWorkers - yarnAllocator.getNumWorkersRunning, 0)) + Thread.sleep(100) + } + + logInfo("All workers have launched.") + + } + + // TODO: We might want to extend this to allocate more containers in case they die ! + private def launchReporterThread(_sleepTime: Long): Thread = { + val sleepTime = if (_sleepTime <= 0 ) 0 else _sleepTime + + val t = new Thread { + override def run() { + while (!driverClosed) { + val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning + if (missingWorkerCount > 0) { + logInfo("Allocating " + missingWorkerCount + " containers to make up for (potentially ?) lost containers") + yarnAllocator.allocateContainers(missingWorkerCount) + } + else sendProgress() + Thread.sleep(sleepTime) + } + } + } + // setting to daemon status, though this is usually not a good idea. + t.setDaemon(true) + t.start() + logInfo("Started progress reporter thread - sleep time : " + sleepTime) + return t + } + + private def sendProgress() { + logDebug("Sending progress") + // simulated with an allocate request with no nodes requested ... + yarnAllocator.allocateContainers(0) + } + + def finishApplicationMaster(status: FinalApplicationStatus) { + + logInfo("finish ApplicationMaster with " + status) + val finishReq = Records.newRecord(classOf[FinishApplicationMasterRequest]) + .asInstanceOf[FinishApplicationMasterRequest] + finishReq.setAppAttemptId(appAttemptId) + finishReq.setFinishApplicationStatus(status) + resourceManager.finishApplicationMaster(finishReq) + } + +} + + +object WorkerLauncher { + def main(argStrings: Array[String]) { + val args = new ApplicationMasterArguments(argStrings) + new WorkerLauncher(args).run() + } +} diff --git a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala new file mode 100644 index 0000000000..4f34bd913e --- /dev/null +++ b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala @@ -0,0 +1,235 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.conf.YarnConfiguration +import org.apache.hadoop.yarn.ipc.YarnRPC +import org.apache.hadoop.yarn.util.{Apps, ConverterUtils, Records, ProtoUtils} + +import org.apache.spark.Logging + + +class WorkerRunnable( + container: Container, + conf: Configuration, + masterAddress: String, + slaveId: String, + hostname: String, + workerMemory: Int, + workerCores: Int) + extends Runnable with Logging { + + var rpc: YarnRPC = YarnRPC.create(conf) + var cm: ContainerManager = _ + val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + + def run = { + logInfo("Starting Worker Container") + cm = connectToCM + startContainer + } + + def startContainer = { + logInfo("Setting up ContainerLaunchContext") + + val ctx = Records.newRecord(classOf[ContainerLaunchContext]) + .asInstanceOf[ContainerLaunchContext] + + ctx.setContainerId(container.getId()) + ctx.setResource(container.getResource()) + val localResources = prepareLocalResources + ctx.setLocalResources(localResources) + + val env = prepareEnvironment + ctx.setEnvironment(env) + + // 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 " + } +*/ + + ctx.setUser(UserGroupInformation.getCurrentUser().getShortUserName()) + + 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()) || 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") + logInfo("Setting up worker with commands: " + commands) + ctx.setCommands(commands) + + // Send the start request to the ContainerManager + val startReq = Records.newRecord(classOf[StartContainerRequest]) + .asInstanceOf[StartContainerRequest] + startReq.setContainerLaunchContext(ctx) + cm.startContainer(startReq) + } + + 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) + return localResources + } + + def prepareEnvironment: HashMap[String, String] = { + val env = new HashMap[String, String]() + + Client.populateClasspath(yarnConf, 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 } + return env + } + + def connectToCM: ContainerManager = { + val cmHostPortStr = container.getNodeId().getHost() + ":" + container.getNodeId().getPort() + val cmAddress = NetUtils.createSocketAddr(cmHostPortStr) + logInfo("Connecting to ContainerManager at " + cmHostPortStr) + + // 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] + } + }) + proxy + } + +} diff --git a/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala new file mode 100644 index 0000000000..c8af653b3f --- /dev/null +++ b/yarn/alpha/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala @@ -0,0 +1,680 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.lang.{Boolean => JBoolean} +import java.util.{Collections, Set => JSet} +import java.util.concurrent.{CopyOnWriteArrayList, ConcurrentHashMap} +import java.util.concurrent.atomic.AtomicInteger + +import scala.collection +import scala.collection.JavaConversions._ +import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} + +import org.apache.spark.{Logging, SparkConf} +import org.apache.spark.scheduler.{SplitInfo,TaskSchedulerImpl} +import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend +import org.apache.spark.util.Utils + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.yarn.api.AMRMProtocol +import org.apache.hadoop.yarn.api.records.{AMResponse, ApplicationAttemptId} +import org.apache.hadoop.yarn.api.records.{Container, ContainerId, ContainerStatus} +import org.apache.hadoop.yarn.api.records.{Priority, Resource, ResourceRequest} +import org.apache.hadoop.yarn.api.protocolrecords.{AllocateRequest, AllocateResponse} +import org.apache.hadoop.yarn.util.{RackResolver, Records} + + +object AllocationType extends Enumeration { + type AllocationType = Value + val HOST, RACK, ANY = Value +} + +// TODO: +// Too many params. +// Needs to be mt-safe +// Need to refactor this to make it 'cleaner' ... right now, all computation is reactive - should +// make it more proactive and decoupled. + +// Note that right now, we assume all node asks as uniform in terms of capabilities and priority +// Refer to http://developer.yahoo.com/blogs/hadoop/posts/2011/03/mapreduce-nextgen-scheduler/ for +// more info on how we are requesting for containers. +private[yarn] class YarnAllocationHandler( + val conf: Configuration, + val resourceManager: AMRMProtocol, + val appAttemptId: ApplicationAttemptId, + val maxWorkers: Int, + val workerMemory: Int, + val workerCores: Int, + val preferredHostToCount: Map[String, Int], + val preferredRackToCount: Map[String, Int], + val sparkConf: SparkConf) + extends Logging { + // These three are locked on allocatedHostToContainersMap. Complementary data structures + // allocatedHostToContainersMap : containers which are running : host, Set + // allocatedContainerToHostMap: container to host mapping. + private val allocatedHostToContainersMap = + new HashMap[String, collection.mutable.Set[ContainerId]]() + + private val allocatedContainerToHostMap = new HashMap[ContainerId, String]() + + // allocatedRackCount is populated ONLY if allocation happens (or decremented if this is an + // allocated node) + // As with the two data structures above, tightly coupled with them, and to be locked on + // allocatedHostToContainersMap + private val allocatedRackCount = new HashMap[String, Int]() + + // Containers which have been released. + private val releasedContainerList = new CopyOnWriteArrayList[ContainerId]() + // Containers to be released in next request to RM + private val pendingReleaseContainers = new ConcurrentHashMap[ContainerId, Boolean] + + private val numWorkersRunning = new AtomicInteger() + // Used to generate a unique id per worker + private val workerIdCounter = new AtomicInteger() + private val lastResponseId = new AtomicInteger() + private val numWorkersFailed = new AtomicInteger() + + def getNumWorkersRunning: Int = numWorkersRunning.intValue + + def getNumWorkersFailed: Int = numWorkersFailed.intValue + + def isResourceConstraintSatisfied(container: Container): Boolean = { + container.getResource.getMemory >= (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) + } + + def allocateContainers(workersToRequest: Int) { + // We need to send the request only once from what I understand ... but for now, not modifying + // this much. + + // Keep polling the Resource Manager for containers + val amResp = allocateWorkerResources(workersToRequest).getAMResponse + + val _allocatedContainers = amResp.getAllocatedContainers() + + if (_allocatedContainers.size > 0) { + logDebug(""" + Allocated containers: %d + Current worker count: %d + Containers released: %s + Containers to be released: %s + Cluster resources: %s + """.format( + _allocatedContainers.size, + numWorkersRunning.get(), + releasedContainerList, + pendingReleaseContainers, + amResp.getAvailableResources)) + + val hostToContainers = new HashMap[String, ArrayBuffer[Container]]() + + // Ignore if not satisfying constraints { + for (container <- _allocatedContainers) { + if (isResourceConstraintSatisfied(container)) { + // allocatedContainers += container + + val host = container.getNodeId.getHost + val containers = hostToContainers.getOrElseUpdate(host, new ArrayBuffer[Container]()) + + containers += container + } + // Add all ignored containers to released list + else releasedContainerList.add(container.getId()) + } + + // Find the appropriate containers to use. Slightly non trivial groupBy ... + val dataLocalContainers = new HashMap[String, ArrayBuffer[Container]]() + val rackLocalContainers = new HashMap[String, ArrayBuffer[Container]]() + val offRackContainers = new HashMap[String, ArrayBuffer[Container]]() + + for (candidateHost <- hostToContainers.keySet) + { + val maxExpectedHostCount = preferredHostToCount.getOrElse(candidateHost, 0) + val requiredHostCount = maxExpectedHostCount - allocatedContainersOnHost(candidateHost) + + var remainingContainers = hostToContainers.get(candidateHost).getOrElse(null) + assert(remainingContainers != null) + + if (requiredHostCount >= remainingContainers.size){ + // Since we got <= required containers, add all to dataLocalContainers + dataLocalContainers.put(candidateHost, remainingContainers) + // all consumed + remainingContainers = null + } + else if (requiredHostCount > 0) { + // Container list has more containers than we need for data locality. + // Split into two : data local container count of (remainingContainers.size - + // requiredHostCount) and rest as remainingContainer + val (dataLocal, remaining) = remainingContainers.splitAt( + remainingContainers.size - requiredHostCount) + dataLocalContainers.put(candidateHost, dataLocal) + // remainingContainers = remaining + + // yarn has nasty habit of allocating a tonne of containers on a host - discourage this : + // add remaining to release list. If we have insufficient containers, next allocation + // cycle will reallocate (but wont treat it as data local) + for (container <- remaining) releasedContainerList.add(container.getId()) + remainingContainers = null + } + + // Now rack local + if (remainingContainers != null){ + val rack = YarnAllocationHandler.lookupRack(conf, candidateHost) + + if (rack != null){ + val maxExpectedRackCount = preferredRackToCount.getOrElse(rack, 0) + val requiredRackCount = maxExpectedRackCount - allocatedContainersOnRack(rack) - + rackLocalContainers.get(rack).getOrElse(List()).size + + + if (requiredRackCount >= remainingContainers.size){ + // Add all to dataLocalContainers + dataLocalContainers.put(rack, remainingContainers) + // All consumed + remainingContainers = null + } + else if (requiredRackCount > 0) { + // container list has more containers than we need for data locality. + // Split into two : data local container count of (remainingContainers.size - + // requiredRackCount) and rest as remainingContainer + val (rackLocal, remaining) = remainingContainers.splitAt( + remainingContainers.size - requiredRackCount) + val existingRackLocal = rackLocalContainers.getOrElseUpdate(rack, + new ArrayBuffer[Container]()) + + existingRackLocal ++= rackLocal + remainingContainers = remaining + } + } + } + + // If still not consumed, then it is off rack host - add to that list. + if (remainingContainers != null){ + offRackContainers.put(candidateHost, remainingContainers) + } + } + + // Now that we have split the containers into various groups, go through them in order : + // first host local, then rack local and then off rack (everything else). + // Note that the list we create below tries to ensure that not all containers end up within a + // host if there are sufficiently large number of hosts/containers. + + val allocatedContainers = new ArrayBuffer[Container](_allocatedContainers.size) + allocatedContainers ++= TaskSchedulerImpl.prioritizeContainers(dataLocalContainers) + allocatedContainers ++= TaskSchedulerImpl.prioritizeContainers(rackLocalContainers) + allocatedContainers ++= TaskSchedulerImpl.prioritizeContainers(offRackContainers) + + // Run each of the allocated containers + for (container <- allocatedContainers) { + val numWorkersRunningNow = numWorkersRunning.incrementAndGet() + val workerHostname = container.getNodeId.getHost + val containerId = container.getId + + assert( + container.getResource.getMemory >= (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD)) + + if (numWorkersRunningNow > maxWorkers) { + logInfo("""Ignoring container %s at host %s, since we already have the required number of + containers for it.""".format(containerId, workerHostname)) + releasedContainerList.add(containerId) + // reset counter back to old value. + numWorkersRunning.decrementAndGet() + } + else { + // Deallocate + allocate can result in reusing id's wrongly - so use a different counter + // (workerIdCounter) + val workerId = workerIdCounter.incrementAndGet().toString + val driverUrl = "akka.tcp://spark@%s:%s/user/%s".format( + sparkConf.get("spark.driver.host"), sparkConf.get("spark.driver.port"), + CoarseGrainedSchedulerBackend.ACTOR_NAME) + + logInfo("launching container on " + containerId + " host " + workerHostname) + // Just to be safe, simply remove it from pendingReleaseContainers. + // Should not be there, but .. + pendingReleaseContainers.remove(containerId) + + val rack = YarnAllocationHandler.lookupRack(conf, workerHostname) + allocatedHostToContainersMap.synchronized { + val containerSet = allocatedHostToContainersMap.getOrElseUpdate(workerHostname, + new HashSet[ContainerId]()) + + containerSet += containerId + allocatedContainerToHostMap.put(containerId, workerHostname) + if (rack != null) { + allocatedRackCount.put(rack, allocatedRackCount.getOrElse(rack, 0) + 1) + } + } + + new Thread( + new WorkerRunnable(container, conf, driverUrl, workerId, + workerHostname, workerMemory, workerCores) + ).start() + } + } + logDebug(""" + Finished processing %d containers. + Current number of workers running: %d, + releasedContainerList: %s, + pendingReleaseContainers: %s + """.format( + allocatedContainers.size, + numWorkersRunning.get(), + releasedContainerList, + pendingReleaseContainers)) + } + + + val completedContainers = amResp.getCompletedContainersStatuses() + if (completedContainers.size > 0){ + logDebug("Completed %d containers, to-be-released: %s".format( + completedContainers.size, releasedContainerList)) + for (completedContainer <- completedContainers){ + val containerId = completedContainer.getContainerId + + // Was this released by us ? If yes, then simply remove from containerSet and move on. + if (pendingReleaseContainers.containsKey(containerId)) { + pendingReleaseContainers.remove(containerId) + } + else { + // Simply decrement count - next iteration of ReporterThread will take care of allocating. + numWorkersRunning.decrementAndGet() + logInfo("Completed container %s (state: %s, exit status: %s)".format( + containerId, + completedContainer.getState, + completedContainer.getExitStatus())) + // Hadoop 2.2.X added a ContainerExitStatus we should switch to use + // there are some exit status' we shouldn't necessarily count against us, but for + // now I think its ok as none of the containers are expected to exit + if (completedContainer.getExitStatus() != 0) { + logInfo("Container marked as failed: " + containerId) + numWorkersFailed.incrementAndGet() + } + } + + allocatedHostToContainersMap.synchronized { + if (allocatedContainerToHostMap.containsKey(containerId)) { + val host = allocatedContainerToHostMap.get(containerId).getOrElse(null) + assert (host != null) + + val containerSet = allocatedHostToContainersMap.get(host).getOrElse(null) + assert (containerSet != null) + + containerSet -= containerId + if (containerSet.isEmpty) allocatedHostToContainersMap.remove(host) + else allocatedHostToContainersMap.update(host, containerSet) + + allocatedContainerToHostMap -= containerId + + // Doing this within locked context, sigh ... move to outside ? + val rack = YarnAllocationHandler.lookupRack(conf, host) + if (rack != null) { + val rackCount = allocatedRackCount.getOrElse(rack, 0) - 1 + if (rackCount > 0) allocatedRackCount.put(rack, rackCount) + else allocatedRackCount.remove(rack) + } + } + } + } + logDebug(""" + Finished processing %d completed containers. + Current number of workers running: %d, + releasedContainerList: %s, + pendingReleaseContainers: %s + """.format( + completedContainers.size, + numWorkersRunning.get(), + releasedContainerList, + pendingReleaseContainers)) + } + } + + def createRackResourceRequests(hostContainers: List[ResourceRequest]): List[ResourceRequest] = { + // First generate modified racks and new set of hosts under it : then issue requests + val rackToCounts = new HashMap[String, Int]() + + // Within this lock - used to read/write to the rack related maps too. + for (container <- hostContainers) { + val candidateHost = container.getHostName + val candidateNumContainers = container.getNumContainers + assert(YarnAllocationHandler.ANY_HOST != candidateHost) + + val rack = YarnAllocationHandler.lookupRack(conf, candidateHost) + if (rack != null) { + var count = rackToCounts.getOrElse(rack, 0) + count += candidateNumContainers + rackToCounts.put(rack, count) + } + } + + val requestedContainers: ArrayBuffer[ResourceRequest] = + new ArrayBuffer[ResourceRequest](rackToCounts.size) + for ((rack, count) <- rackToCounts){ + requestedContainers += + createResourceRequest(AllocationType.RACK, rack, count, YarnAllocationHandler.PRIORITY) + } + + requestedContainers.toList + } + + def allocatedContainersOnHost(host: String): Int = { + var retval = 0 + allocatedHostToContainersMap.synchronized { + retval = allocatedHostToContainersMap.getOrElse(host, Set()).size + } + retval + } + + def allocatedContainersOnRack(rack: String): Int = { + var retval = 0 + allocatedHostToContainersMap.synchronized { + retval = allocatedRackCount.getOrElse(rack, 0) + } + retval + } + + private def allocateWorkerResources(numWorkers: Int): AllocateResponse = { + + var resourceRequests: List[ResourceRequest] = null + + // default. + if (numWorkers <= 0 || preferredHostToCount.isEmpty) { + logDebug("numWorkers: " + numWorkers + ", host preferences: " + preferredHostToCount.isEmpty) + resourceRequests = List( + createResourceRequest(AllocationType.ANY, null, numWorkers, YarnAllocationHandler.PRIORITY)) + } + else { + // request for all hosts in preferred nodes and for numWorkers - + // candidates.size, request by default allocation policy. + val hostContainerRequests: ArrayBuffer[ResourceRequest] = + new ArrayBuffer[ResourceRequest](preferredHostToCount.size) + for ((candidateHost, candidateCount) <- preferredHostToCount) { + val requiredCount = candidateCount - allocatedContainersOnHost(candidateHost) + + if (requiredCount > 0) { + hostContainerRequests += createResourceRequest( + AllocationType.HOST, + candidateHost, + requiredCount, + YarnAllocationHandler.PRIORITY) + } + } + val rackContainerRequests: List[ResourceRequest] = createRackResourceRequests( + hostContainerRequests.toList) + + val anyContainerRequests: ResourceRequest = createResourceRequest( + AllocationType.ANY, + resource = null, + numWorkers, + YarnAllocationHandler.PRIORITY) + + val containerRequests: ArrayBuffer[ResourceRequest] = new ArrayBuffer[ResourceRequest]( + hostContainerRequests.size + rackContainerRequests.size + 1) + + containerRequests ++= hostContainerRequests + containerRequests ++= rackContainerRequests + containerRequests += anyContainerRequests + + resourceRequests = containerRequests.toList + } + + val req = Records.newRecord(classOf[AllocateRequest]) + req.setResponseId(lastResponseId.incrementAndGet) + req.setApplicationAttemptId(appAttemptId) + + req.addAllAsks(resourceRequests) + + val releasedContainerList = createReleasedContainerList() + req.addAllReleases(releasedContainerList) + + if (numWorkers > 0) { + logInfo("Allocating %d worker containers with %d of memory each.".format(numWorkers, + workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD)) + } + else { + logDebug("Empty allocation req .. release : " + releasedContainerList) + } + + for (request <- resourceRequests) { + logInfo("ResourceRequest (host : %s, num containers: %d, priority = %s , capability : %s)". + format( + request.getHostName, + request.getNumContainers, + request.getPriority, + request.getCapability)) + } + resourceManager.allocate(req) + } + + + private def createResourceRequest( + requestType: AllocationType.AllocationType, + resource:String, + numWorkers: Int, + priority: Int): ResourceRequest = { + + // If hostname specified, we need atleast two requests - node local and rack local. + // There must be a third request - which is ANY : that will be specially handled. + requestType match { + case AllocationType.HOST => { + assert(YarnAllocationHandler.ANY_HOST != resource) + val hostname = resource + val nodeLocal = createResourceRequestImpl(hostname, numWorkers, priority) + + // Add to host->rack mapping + YarnAllocationHandler.populateRackInfo(conf, hostname) + + nodeLocal + } + case AllocationType.RACK => { + val rack = resource + createResourceRequestImpl(rack, numWorkers, priority) + } + case AllocationType.ANY => createResourceRequestImpl( + YarnAllocationHandler.ANY_HOST, numWorkers, priority) + case _ => throw new IllegalArgumentException( + "Unexpected/unsupported request type: " + requestType) + } + } + + private def createResourceRequestImpl( + hostname:String, + numWorkers: Int, + priority: Int): ResourceRequest = { + + val rsrcRequest = Records.newRecord(classOf[ResourceRequest]) + val memCapability = Records.newRecord(classOf[Resource]) + // There probably is some overhead here, let's reserve a bit more memory. + memCapability.setMemory(workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) + rsrcRequest.setCapability(memCapability) + + val pri = Records.newRecord(classOf[Priority]) + pri.setPriority(priority) + rsrcRequest.setPriority(pri) + + rsrcRequest.setHostName(hostname) + + rsrcRequest.setNumContainers(java.lang.Math.max(numWorkers, 0)) + rsrcRequest + } + + def createReleasedContainerList(): ArrayBuffer[ContainerId] = { + + val retval = new ArrayBuffer[ContainerId](1) + // Iterator on COW list ... + for (container <- releasedContainerList.iterator()){ + retval += container + } + // Remove from the original list. + if (! retval.isEmpty) { + releasedContainerList.removeAll(retval) + for (v <- retval) pendingReleaseContainers.put(v, true) + logInfo("Releasing " + retval.size + " containers. pendingReleaseContainers : " + + pendingReleaseContainers) + } + + retval + } +} + +object YarnAllocationHandler { + + val ANY_HOST = "*" + // All requests are issued with same priority : we do not (yet) have any distinction between + // request types (like map/reduce in hadoop for example) + val PRIORITY = 1 + + // Additional memory overhead - in mb + val MEMORY_OVERHEAD = 384 + + // Host to rack map - saved from allocation requests + // We are expecting this not to change. + // Note that it is possible for this to change : and RM will indicate that to us via update + // response to allocate. But we are punting on handling that for now. + private val hostToRack = new ConcurrentHashMap[String, String]() + private val rackToHostSet = new ConcurrentHashMap[String, JSet[String]]() + + + def newAllocator( + conf: Configuration, + resourceManager: AMRMProtocol, + appAttemptId: ApplicationAttemptId, + args: ApplicationMasterArguments, + sparkConf: SparkConf): YarnAllocationHandler = { + + new YarnAllocationHandler( + conf, + resourceManager, + appAttemptId, + args.numWorkers, + args.workerMemory, + args.workerCores, + Map[String, Int](), + Map[String, Int](), + sparkConf) + } + + def newAllocator( + conf: Configuration, + resourceManager: AMRMProtocol, + appAttemptId: ApplicationAttemptId, + args: ApplicationMasterArguments, + map: collection.Map[String, + collection.Set[SplitInfo]], + sparkConf: SparkConf): YarnAllocationHandler = { + + val (hostToCount, rackToCount) = generateNodeToWeight(conf, map) + new YarnAllocationHandler( + conf, + resourceManager, + appAttemptId, + args.numWorkers, + args.workerMemory, + args.workerCores, + hostToCount, + rackToCount, + sparkConf) + } + + def newAllocator( + conf: Configuration, + resourceManager: AMRMProtocol, + appAttemptId: ApplicationAttemptId, + maxWorkers: Int, + workerMemory: Int, + workerCores: Int, + map: collection.Map[String, collection.Set[SplitInfo]], + sparkConf: SparkConf): YarnAllocationHandler = { + + val (hostToCount, rackToCount) = generateNodeToWeight(conf, map) + + new YarnAllocationHandler( + conf, + resourceManager, + appAttemptId, + maxWorkers, + workerMemory, + workerCores, + hostToCount, + rackToCount, + sparkConf) + } + + // A simple method to copy the split info map. + private def generateNodeToWeight( + conf: Configuration, + input: collection.Map[String, collection.Set[SplitInfo]]) : + // host to count, rack to count + (Map[String, Int], Map[String, Int]) = { + + if (input == null) return (Map[String, Int](), Map[String, Int]()) + + val hostToCount = new HashMap[String, Int] + val rackToCount = new HashMap[String, Int] + + for ((host, splits) <- input) { + val hostCount = hostToCount.getOrElse(host, 0) + hostToCount.put(host, hostCount + splits.size) + + val rack = lookupRack(conf, host) + if (rack != null){ + val rackCount = rackToCount.getOrElse(host, 0) + rackToCount.put(host, rackCount + splits.size) + } + } + + (hostToCount.toMap, rackToCount.toMap) + } + + def lookupRack(conf: Configuration, host: String): String = { + if (!hostToRack.contains(host)) populateRackInfo(conf, host) + hostToRack.get(host) + } + + def fetchCachedHostsForRack(rack: String): Option[Set[String]] = { + val set = rackToHostSet.get(rack) + if (set == null) return None + + // No better way to get a Set[String] from JSet ? + val convertedSet: collection.mutable.Set[String] = set + Some(convertedSet.toSet) + } + + def populateRackInfo(conf: Configuration, hostname: String) { + Utils.checkHost(hostname) + + if (!hostToRack.containsKey(hostname)) { + // If there are repeated failures to resolve, all to an ignore list ? + val rackInfo = RackResolver.resolve(conf, hostname) + if (rackInfo != null && rackInfo.getNetworkLocation != null) { + val rack = rackInfo.getNetworkLocation + hostToRack.put(hostname, rack) + if (! rackToHostSet.containsKey(rack)) { + rackToHostSet.putIfAbsent(rack, + Collections.newSetFromMap(new ConcurrentHashMap[String, JBoolean]())) + } + rackToHostSet.get(rack).add(hostname) + + // TODO(harvey): Figure out this comment... + // Since RackResolver caches, we are disabling this for now ... + } /* else { + // right ? Else we will keep calling rack resolver in case we cant resolve rack info ... + hostToRack.put(hostname, null) + } */ + } + } +} diff --git a/yarn/pom.xml b/yarn/pom.xml index 3300a465ef..7c145ef362 100644 --- a/yarn/pom.xml +++ b/yarn/pom.xml @@ -89,16 +89,16 @@ - hadoop2-yarn + yarn-alpha - 2.0 + alpha - hadoop2.2-yarn + yarn - 2.2 + stable diff --git a/yarn/stable/pom.xml b/yarn/stable/pom.xml new file mode 100644 index 0000000000..45a14313f6 --- /dev/null +++ b/yarn/stable/pom.xml @@ -0,0 +1,112 @@ + + + + 4.0.0 + + org.apache.spark + spark-yarn-aggregator_2.10 + 0.9.0-incubating-SNAPSHOT + ../pom.xml + + + org.apache.spark + spark-yarn_2.10 + jar + Spark Project YARN Stable API + http://spark.incubator.apache.org/ + + + + + org.codehaus.mojo + build-helper-maven-plugin + + + add-scala-sources + generate-sources + + add-source + + + + src/main/scala + ../common/src/main/scala + + + + + add-scala-test-sources + generate-test-sources + + add-test-source + + + + src/test/scala + ../common/src/test/scala + + + + + + + org.apache.maven.plugins + maven-antrun-plugin + + + test + + run + + + true + + + + + + + + + + + + + + + + + + + + org.scalatest + scalatest-maven-plugin + + + ${basedir}/../.. + 1 + ${spark.classpath} + + + + + + target/scala-${scala.binary.version}/classes + target/scala-${scala.binary.version}/test-classes + + diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala new file mode 100644 index 0000000000..7c32e0ab9b --- /dev/null +++ b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/ApplicationMaster.scala @@ -0,0 +1,428 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.io.IOException +import java.net.Socket +import java.util.concurrent.CopyOnWriteArrayList +import java.util.concurrent.atomic.{AtomicInteger, AtomicReference} + +import scala.collection.JavaConversions._ + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.fs.{FileSystem, Path} +import org.apache.hadoop.net.NetUtils +import org.apache.hadoop.security.UserGroupInformation +import org.apache.hadoop.util.ShutdownHookManager +import org.apache.hadoop.yarn.api._ +import org.apache.hadoop.yarn.api.protocolrecords._ +import org.apache.hadoop.yarn.api.records._ +import org.apache.hadoop.yarn.client.api.AMRMClient +import org.apache.hadoop.yarn.client.api.AMRMClient.ContainerRequest +import org.apache.hadoop.yarn.conf.YarnConfiguration +import org.apache.hadoop.yarn.ipc.YarnRPC +import org.apache.hadoop.yarn.util.{ConverterUtils, Records} + +import org.apache.spark.{SparkConf, SparkContext, Logging} +import org.apache.spark.util.Utils + + +class ApplicationMaster(args: ApplicationMasterArguments, conf: Configuration) extends Logging { + + def this(args: ApplicationMasterArguments) = this(args, new Configuration()) + + private var rpc: YarnRPC = YarnRPC.create(conf) + private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + private var appAttemptId: ApplicationAttemptId = _ + private var userThread: Thread = _ + private val fs = FileSystem.get(yarnConf) + + private var yarnAllocator: YarnAllocationHandler = _ + private var isFinished: Boolean = false + private var uiAddress: String = _ + private val maxAppAttempts: Int = conf.getInt( + YarnConfiguration.RM_AM_MAX_ATTEMPTS, YarnConfiguration.DEFAULT_RM_AM_MAX_ATTEMPTS) + private var isLastAMRetry: Boolean = true + private var amClient: AMRMClient[ContainerRequest] = _ + + private val sparkConf = new SparkConf() + // Default to numWorkers * 2, with minimum of 3 + private val maxNumWorkerFailures = sparkConf.getInt("spark.yarn.max.worker.failures", + math.max(args.numWorkers * 2, 3)) + + def run() { + // Setup the directories so things go to YARN approved directories rather + // than user specified and /tmp. + System.setProperty("spark.local.dir", getLocalDirs()) + + // set the web ui port to be ephemeral for yarn so we don't conflict with + // other spark processes running on the same box + System.setProperty("spark.ui.port", "0") + + // Use priority 30 as it's higher then HDFS. It's same priority as MapReduce is using. + ShutdownHookManager.get().addShutdownHook(new AppMasterShutdownHook(this), 30) + + appAttemptId = getApplicationAttemptId() + isLastAMRetry = appAttemptId.getAttemptId() >= maxAppAttempts + amClient = AMRMClient.createAMRMClient() + amClient.init(yarnConf) + amClient.start() + + // Workaround until hadoop moves to something which has + // https://issues.apache.org/jira/browse/HADOOP-8406 - fixed in (2.0.2-alpha but no 0.23 line) + // org.apache.hadoop.io.compress.CompressionCodecFactory.getCodecClasses(conf) + + ApplicationMaster.register(this) + + // Start the user's JAR + userThread = startUserClass() + + // This a bit hacky, but we need to wait until the spark.driver.port property has + // been set by the Thread executing the user class. + waitForSparkContextInitialized() + + // Do this after Spark master is up and SparkContext is created so that we can register UI Url. + val appMasterResponse: RegisterApplicationMasterResponse = registerApplicationMaster() + + // Allocate all containers + allocateWorkers() + + // Wait for the user class to Finish + userThread.join() + + System.exit(0) + } + + /** Get the Yarn approved local directories. */ + private def getLocalDirs(): String = { + // Hadoop 0.23 and 2.x have different Environment variable names for the + // local dirs, so lets check both. We assume one of the 2 is set. + // LOCAL_DIRS => 2.X, YARN_LOCAL_DIRS => 0.23.X + val localDirs = Option(System.getenv("YARN_LOCAL_DIRS")) + .getOrElse(Option(System.getenv("LOCAL_DIRS")) + .getOrElse("")) + + if (localDirs.isEmpty()) { + throw new Exception("Yarn Local dirs can't be empty") + } + localDirs + } + + private def getApplicationAttemptId(): ApplicationAttemptId = { + val envs = System.getenv() + val containerIdString = envs.get(ApplicationConstants.Environment.CONTAINER_ID.name()) + val containerId = ConverterUtils.toContainerId(containerIdString) + val appAttemptId = containerId.getApplicationAttemptId() + logInfo("ApplicationAttemptId: " + appAttemptId) + appAttemptId + } + + private def registerApplicationMaster(): RegisterApplicationMasterResponse = { + logInfo("Registering the ApplicationMaster") + amClient.registerApplicationMaster(Utils.localHostName(), 0, uiAddress) + } + + private def startUserClass(): Thread = { + logInfo("Starting the user JAR in a separate Thread") + val mainMethod = Class.forName( + args.userClass, + false /* initialize */, + Thread.currentThread.getContextClassLoader).getMethod("main", classOf[Array[String]]) + val t = new Thread { + override def run() { + 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 { + logDebug("finishing main") + isLastAMRetry = true + if (successed) { + ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) + } else { + ApplicationMaster.this.finishApplicationMaster(FinalApplicationStatus.FAILED) + } + } + } + } + t.start() + t + } + + // This need to happen before allocateWorkers() + private def waitForSparkContextInitialized() { + logInfo("Waiting for Spark context initialization") + try { + var sparkContext: SparkContext = null + ApplicationMaster.sparkContextRef.synchronized { + var numTries = 0 + val waitTime = 10000L + val maxNumTries = sparkConf.getInt("spark.yarn.applicationMaster.waitTries", 10) + while (ApplicationMaster.sparkContextRef.get() == null && numTries < maxNumTries) { + logInfo("Waiting for Spark context initialization ... " + numTries) + numTries = numTries + 1 + ApplicationMaster.sparkContextRef.wait(waitTime) + } + sparkContext = ApplicationMaster.sparkContextRef.get() + assert(sparkContext != null || numTries >= maxNumTries) + + if (sparkContext != null) { + uiAddress = sparkContext.ui.appUIAddress + this.yarnAllocator = YarnAllocationHandler.newAllocator( + yarnConf, + amClient, + appAttemptId, + args, + sparkContext.preferredNodeLocationData, + sparkContext.getConf) + } else { + logWarning("Unable to retrieve SparkContext inspite of waiting for %d, maxNumTries = %d". + format(numTries * waitTime, maxNumTries)) + this.yarnAllocator = YarnAllocationHandler.newAllocator( + yarnConf, + amClient, + appAttemptId, + args, + sparkContext.getConf) + } + } + } finally { + // In case of exceptions, etc - ensure that count is at least ALLOCATOR_LOOP_WAIT_COUNT : + // so that the loop (in ApplicationMaster.sparkContextInitialized) breaks. + ApplicationMaster.incrementAllocatorLoop(ApplicationMaster.ALLOCATOR_LOOP_WAIT_COUNT) + } + } + + private def allocateWorkers() { + try { + logInfo("Allocating " + args.numWorkers + " workers.") + // Wait until all containers have finished + // TODO: This is a bit ugly. Can we make it nicer? + // TODO: Handle container failure + yarnAllocator.addResourceRequests(args.numWorkers) + // Exits the loop if the user thread exits. + while (yarnAllocator.getNumWorkersRunning < args.numWorkers && userThread.isAlive) { + if (yarnAllocator.getNumWorkersFailed >= maxNumWorkerFailures) { + finishApplicationMaster(FinalApplicationStatus.FAILED, + "max number of worker failures reached") + } + yarnAllocator.allocateResources() + ApplicationMaster.incrementAllocatorLoop(1) + Thread.sleep(100) + } + } finally { + // In case of exceptions, etc - ensure that count is at least ALLOCATOR_LOOP_WAIT_COUNT, + // so that the loop in ApplicationMaster#sparkContextInitialized() breaks. + ApplicationMaster.incrementAllocatorLoop(ApplicationMaster.ALLOCATOR_LOOP_WAIT_COUNT) + } + logInfo("All workers have launched.") + + // Launch a progress reporter thread, else the app will get killed after expiration + // (def: 10mins) timeout. + if (userThread.isAlive) { + // Ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapses. + val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) + + // we want to be reasonably responsive without causing too many requests to RM. + val schedulerInterval = + sparkConf.getLong("spark.yarn.scheduler.heartbeat.interval-ms", 5000) + + + // must be <= timeoutInterval / 2. + val interval = math.min(timeoutInterval / 2, schedulerInterval) + + launchReporterThread(interval) + } + } + + private def launchReporterThread(_sleepTime: Long): Thread = { + val sleepTime = if (_sleepTime <= 0 ) 0 else _sleepTime + + val t = new Thread { + override def run() { + while (userThread.isAlive) { + if (yarnAllocator.getNumWorkersFailed >= maxNumWorkerFailures) { + finishApplicationMaster(FinalApplicationStatus.FAILED, + "max number of worker failures reached") + } + val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning - + yarnAllocator.getNumPendingAllocate + if (missingWorkerCount > 0) { + logInfo("Allocating %d containers to make up for (potentially) lost containers". + format(missingWorkerCount)) + yarnAllocator.addResourceRequests(missingWorkerCount) + } + sendProgress() + Thread.sleep(sleepTime) + } + } + } + // Setting to daemon status, though this is usually not a good idea. + t.setDaemon(true) + t.start() + logInfo("Started progress reporter thread - sleep time : " + sleepTime) + t + } + + private def sendProgress() { + logDebug("Sending progress") + // Simulated with an allocate request with no nodes requested. + yarnAllocator.allocateResources() + } + + /* + def printContainers(containers: List[Container]) = { + for (container <- containers) { + logInfo("Launching shell command on a new container." + + ", containerId=" + container.getId() + + ", containerNode=" + container.getNodeId().getHost() + + ":" + container.getNodeId().getPort() + + ", containerNodeURI=" + container.getNodeHttpAddress() + + ", containerState" + container.getState() + + ", containerResourceMemory" + + container.getResource().getMemory()) + } + } + */ + + def finishApplicationMaster(status: FinalApplicationStatus, diagnostics: String = "") { + synchronized { + if (isFinished) { + return + } + isFinished = true + } + + logInfo("finishApplicationMaster with " + status) + // Set tracking URL to empty since we don't have a history server. + amClient.unregisterApplicationMaster(status, "" /* appMessage */, "" /* appTrackingUrl */) + } + + /** + * Clean up the staging directory. + */ + private def cleanupStagingDir() { + var stagingDirPath: Path = null + try { + val preserveFiles = sparkConf.get("spark.yarn.preserve.staging.files", "false").toBoolean + if (!preserveFiles) { + stagingDirPath = new Path(System.getenv("SPARK_YARN_STAGING_DIR")) + if (stagingDirPath == null) { + logError("Staging directory is null") + return + } + logInfo("Deleting staging directory " + stagingDirPath) + fs.delete(stagingDirPath, true) + } + } catch { + case ioe: IOException => + logError("Failed to cleanup staging dir " + stagingDirPath, ioe) + } + } + + // The shutdown hook that runs when a signal is received AND during normal close of the JVM. + class AppMasterShutdownHook(appMaster: ApplicationMaster) extends Runnable { + + def run() { + logInfo("AppMaster received a signal.") + // we need to clean up staging dir before HDFS is shut down + // make sure we don't delete it until this is the last AM + if (appMaster.isLastAMRetry) appMaster.cleanupStagingDir() + } + } +} + +object ApplicationMaster { + // Number of times to wait for the allocator loop to complete. + // Each loop iteration waits for 100ms, so maximum of 3 seconds. + // This is to ensure that we have reasonable number of containers before we start + // TODO: Currently, task to container is computed once (TaskSetManager) - which need not be + // optimal as more containers are available. Might need to handle this better. + private val ALLOCATOR_LOOP_WAIT_COUNT = 30 + + private val applicationMasters = new CopyOnWriteArrayList[ApplicationMaster]() + + val sparkContextRef: AtomicReference[SparkContext] = + new AtomicReference[SparkContext](null /* initialValue */) + + val yarnAllocatorLoop: AtomicInteger = new AtomicInteger(0) + + def incrementAllocatorLoop(by: Int) { + val count = yarnAllocatorLoop.getAndAdd(by) + if (count >= ALLOCATOR_LOOP_WAIT_COUNT) { + yarnAllocatorLoop.synchronized { + // to wake threads off wait ... + yarnAllocatorLoop.notifyAll() + } + } + } + + def register(master: ApplicationMaster) { + applicationMasters.add(master) + } + + // TODO(harvey): See whether this should be discarded - it isn't used anywhere atm... + def sparkContextInitialized(sc: SparkContext): Boolean = { + var modified = false + sparkContextRef.synchronized { + modified = sparkContextRef.compareAndSet(null, sc) + sparkContextRef.notifyAll() + } + + // Add a shutdown hook - as a best case effort in case users do not call sc.stop or do + // System.exit. + // Should not really have to do this, but it helps YARN to evict resources earlier. + // Not to mention, prevent the Client from declaring failure even though we exited properly. + // Note that this will unfortunately not properly clean up the staging files because it gets + // called too late, after the filesystem is already shutdown. + if (modified) { + Runtime.getRuntime().addShutdownHook(new Thread with Logging { + // This is not only logs, but also ensures that log system is initialized for this instance + // when we are actually 'run'-ing. + logInfo("Adding shutdown hook for context " + sc) + override def run() { + logInfo("Invoking sc stop from shutdown hook") + sc.stop() + // Best case ... + 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) { + yarnAllocatorLoop.wait(1000L) + } + } + modified + } + + def main(argStrings: Array[String]) { + val args = new ApplicationMasterArguments(argStrings) + new ApplicationMaster(args).run() + } +} diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/Client.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/Client.scala new file mode 100644 index 0000000000..a75066888c --- /dev/null +++ b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/Client.scala @@ -0,0 +1,523 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.{InetAddress, UnknownHostException, URI} +import java.nio.ByteBuffer + +import scala.collection.JavaConversions._ +import scala.collection.mutable.HashMap +import scala.collection.mutable.Map + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.fs.{FileContext, FileStatus, FileSystem, Path, FileUtil} +import org.apache.hadoop.fs.permission.FsPermission; +import org.apache.hadoop.io.DataOutputBuffer +import org.apache.hadoop.mapred.Master +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.protocolrecords._ +import org.apache.hadoop.yarn.api.records._ +import org.apache.hadoop.yarn.client.api.impl.YarnClientImpl +import org.apache.hadoop.yarn.conf.YarnConfiguration +import org.apache.hadoop.yarn.ipc.YarnRPC +import org.apache.hadoop.yarn.util.{Apps, Records} + +import org.apache.spark.{Logging, SparkConf} +import org.apache.spark.util.Utils +import org.apache.spark.deploy.SparkHadoopUtil + + +/** + * The entry point (starting in Client#main() and Client#run()) for launching Spark on YARN. The + * Client submits an application to the global ResourceManager to launch Spark's ApplicationMaster, + * which will launch a Spark master process and negotiate resources throughout its duration. + */ +class Client(conf: Configuration, args: ClientArguments) extends YarnClientImpl with Logging { + + var rpc: YarnRPC = YarnRPC.create(conf) + val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + val credentials = UserGroupInformation.getCurrentUser().getCredentials() + private val SPARK_STAGING: String = ".sparkStaging" + private val distCacheMgr = new ClientDistributedCacheManager() + private val sparkConf = new SparkConf + + + // Staging directory is private! -> rwx-------- + val STAGING_DIR_PERMISSION: FsPermission = FsPermission.createImmutable(0700: Short) + // App files are world-wide readable and owner writable -> rw-r--r-- + val APP_FILE_PERMISSION: FsPermission = FsPermission.createImmutable(0644: Short) + + def this(args: ClientArguments) = this(new Configuration(), args) + + def runApp(): ApplicationId = { + validateArgs() + // Initialize and start the client service. + init(yarnConf) + start() + + // Log details about this YARN cluster (e.g, the number of slave machines/NodeManagers). + logClusterResourceDetails() + + // Prepare to submit a request to the ResourcManager (specifically its ApplicationsManager (ASM) + // interface). + + // Get a new client application. + val newApp = super.createApplication() + val newAppResponse = newApp.getNewApplicationResponse() + val appId = newAppResponse.getApplicationId() + + verifyClusterResources(newAppResponse) + + // Set up resource and environment variables. + val appStagingDir = getAppStagingDir(appId) + val localResources = prepareLocalResources(appStagingDir) + val launchEnv = setupLaunchEnv(localResources, appStagingDir) + val amContainer = createContainerLaunchContext(newAppResponse, localResources, launchEnv) + + // Set up an application submission context. + val appContext = newApp.getApplicationSubmissionContext() + appContext.setApplicationName(args.appName) + appContext.setQueue(args.amQueue) + appContext.setAMContainerSpec(amContainer) + + // Memory for the ApplicationMaster. + val memoryResource = Records.newRecord(classOf[Resource]).asInstanceOf[Resource] + memoryResource.setMemory(args.amMemory + YarnAllocationHandler.MEMORY_OVERHEAD) + appContext.setResource(memoryResource) + + // Finally, submit and monitor the application. + submitApp(appContext) + appId + } + + def run() { + val appId = runApp() + monitorApplication(appId) + System.exit(0) + } + + // TODO(harvey): This could just go in ClientArguments. + def validateArgs() = { + Map( + (System.getenv("SPARK_JAR") == null) -> "Error: You must set SPARK_JAR environment variable!", + (args.userJar == null) -> "Error: You must specify a user jar!", + (args.userClass == null) -> "Error: You must specify a user class!", + (args.numWorkers <= 0) -> "Error: You must specify atleast 1 worker!", + (args.amMemory <= YarnAllocationHandler.MEMORY_OVERHEAD) -> ("Error: AM memory size must be" + + "greater than: " + YarnAllocationHandler.MEMORY_OVERHEAD), + (args.workerMemory <= YarnAllocationHandler.MEMORY_OVERHEAD) -> ("Error: Worker memory size" + + "must be greater than: " + YarnAllocationHandler.MEMORY_OVERHEAD.toString) + ).foreach { case(cond, errStr) => + if (cond) { + logError(errStr) + args.printUsageAndExit(1) + } + } + } + + def getAppStagingDir(appId: ApplicationId): String = { + SPARK_STAGING + Path.SEPARATOR + appId.toString() + Path.SEPARATOR + } + + def logClusterResourceDetails() { + val clusterMetrics: YarnClusterMetrics = super.getYarnClusterMetrics + logInfo("Got Cluster metric info from ApplicationsManager (ASM), number of NodeManagers: " + + clusterMetrics.getNumNodeManagers) + + val queueInfo: QueueInfo = super.getQueueInfo(args.amQueue) + logInfo("""Queue info ... queueName: %s, queueCurrentCapacity: %s, queueMaxCapacity: %s, + queueApplicationCount = %s, queueChildQueueCount = %s""".format( + queueInfo.getQueueName, + queueInfo.getCurrentCapacity, + queueInfo.getMaximumCapacity, + queueInfo.getApplications.size, + queueInfo.getChildQueues.size)) + } + + def verifyClusterResources(app: GetNewApplicationResponse) = { + val maxMem = app.getMaximumResourceCapability().getMemory() + logInfo("Max mem capabililty of a single resource in this cluster " + maxMem) + + // If we have requested more then the clusters max for a single resource then exit. + if (args.workerMemory > maxMem) { + logError("Required worker memory (%d MB), is above the max threshold (%d MB) of this cluster.". + format(args.workerMemory, maxMem)) + System.exit(1) + } + val amMem = args.amMemory + YarnAllocationHandler.MEMORY_OVERHEAD + if (amMem > maxMem) { + logError("Required AM memory (%d) is above the max threshold (%d) of this cluster". + format(args.amMemory, maxMem)) + 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. + } + + /** See if two file systems are the same or not. */ + private def compareFs(srcFs: FileSystem, destFs: FileSystem): Boolean = { + val srcUri = srcFs.getUri() + val dstUri = destFs.getUri() + if (srcUri.getScheme() == null) { + return false + } + if (!srcUri.getScheme().equals(dstUri.getScheme())) { + return false + } + var srcHost = srcUri.getHost() + var dstHost = dstUri.getHost() + if ((srcHost != null) && (dstHost != null)) { + try { + srcHost = InetAddress.getByName(srcHost).getCanonicalHostName() + dstHost = InetAddress.getByName(dstHost).getCanonicalHostName() + } catch { + case e: UnknownHostException => + return false + } + if (!srcHost.equals(dstHost)) { + return false + } + } else if (srcHost == null && dstHost != null) { + return false + } else if (srcHost != null && dstHost == null) { + return false + } + //check for ports + if (srcUri.getPort() != dstUri.getPort()) { + return false + } + return true + } + + /** Copy the file into HDFS if needed. */ + private def copyRemoteFile( + dstDir: Path, + originalPath: Path, + replication: Short, + setPerms: Boolean = false): Path = { + val fs = FileSystem.get(conf) + val remoteFs = originalPath.getFileSystem(conf) + var newPath = originalPath + if (! compareFs(remoteFs, fs)) { + newPath = new Path(dstDir, originalPath.getName()) + logInfo("Uploading " + originalPath + " to " + newPath) + FileUtil.copy(remoteFs, originalPath, fs, newPath, false, conf) + fs.setReplication(newPath, replication) + if (setPerms) fs.setPermission(newPath, new FsPermission(APP_FILE_PERMISSION)) + } + // Resolve any symlinks in the URI path so using a "current" symlink to point to a specific + // version shows the specific version in the distributed cache configuration + val qualPath = fs.makeQualified(newPath) + val fc = FileContext.getFileContext(qualPath.toUri(), conf) + val destPath = fc.resolvePath(qualPath) + destPath + } + + def prepareLocalResources(appStagingDir: String): HashMap[String, LocalResource] = { + logInfo("Preparing Local resources") + // Upload Spark and the application JAR to the remote file system if necessary. Add them as + // local resources to the application master. + 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) + } + } + val dst = new Path(fs.getHomeDirectory(), appStagingDir) + val replication = sparkConf.getInt("spark.yarn.submit.file.replication", 3).toShort + + if (UserGroupInformation.isSecurityEnabled()) { + val dstFs = dst.getFileSystem(conf) + dstFs.addDelegationTokens(delegTokenRenewer, credentials) + } + + val localResources = HashMap[String, LocalResource]() + FileSystem.mkdirs(fs, dst, new FsPermission(STAGING_DIR_PERMISSION)) + + val statCache: Map[URI, FileStatus] = HashMap[URI, FileStatus]() + + Map( + Client.SPARK_JAR -> System.getenv("SPARK_JAR"), Client.APP_JAR -> args.userJar, + Client.LOG4J_PROP -> System.getenv("SPARK_LOG4J_CONF") + ).foreach { case(destName, _localPath) => + val localPath: String = if (_localPath != null) _localPath.trim() else "" + if (! localPath.isEmpty()) { + var localURI = new URI(localPath) + // If not specified assume these are in the local filesystem to keep behavior like Hadoop + if (localURI.getScheme() == null) { + localURI = new URI(FileSystem.getLocal(conf).makeQualified(new Path(localPath)).toString) + } + val setPermissions = if (destName.equals(Client.APP_JAR)) true else false + val destPath = copyRemoteFile(dst, new Path(localURI), replication, setPermissions) + distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, + destName, statCache) + } + } + + // Handle jars local to the ApplicationMaster. + if ((args.addJars != null) && (!args.addJars.isEmpty())){ + args.addJars.split(',').foreach { case file: String => + val localURI = new URI(file.trim()) + val localPath = new Path(localURI) + val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) + val destPath = copyRemoteFile(dst, localPath, replication) + // Only add the resource to the Spark ApplicationMaster. + val appMasterOnly = true + distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, + linkname, statCache, appMasterOnly) + } + } + + // Handle any distributed cache files + if ((args.files != null) && (!args.files.isEmpty())){ + args.files.split(',').foreach { case file: String => + val localURI = new URI(file.trim()) + val localPath = new Path(localURI) + val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) + val destPath = copyRemoteFile(dst, localPath, replication) + distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.FILE, + linkname, statCache) + } + } + + // Handle any distributed cache archives + if ((args.archives != null) && (!args.archives.isEmpty())) { + args.archives.split(',').foreach { case file:String => + val localURI = new URI(file.trim()) + val localPath = new Path(localURI) + val linkname = Option(localURI.getFragment()).getOrElse(localPath.getName()) + val destPath = copyRemoteFile(dst, localPath, replication) + distCacheMgr.addResource(fs, conf, destPath, localResources, LocalResourceType.ARCHIVE, + linkname, statCache) + } + } + + UserGroupInformation.getCurrentUser().addCredentials(credentials) + localResources + } + + def setupLaunchEnv( + localResources: HashMap[String, LocalResource], + stagingDir: String): HashMap[String, String] = { + logInfo("Setting up the launch environment") + val log4jConfLocalRes = localResources.getOrElse(Client.LOG4J_PROP, null) + + val env = new HashMap[String, String]() + + Client.populateClasspath(yarnConf, log4jConfLocalRes != null, env) + env("SPARK_YARN_MODE") = "true" + env("SPARK_YARN_STAGING_DIR") = stagingDir + + // Set the environment variables to be passed on to the Workers. + distCacheMgr.setDistFilesEnv(env) + distCacheMgr.setDistArchivesEnv(env) + + // Allow users to specify some environment variables. + Apps.setEnvFromInputString(env, System.getenv("SPARK_YARN_USER_ENV")) + + // Add each SPARK_* key to the environment. + System.getenv().filterKeys(_.startsWith("SPARK")).foreach { case (k,v) => env(k) = v } + + env + } + + def userArgsToString(clientArgs: ClientArguments): String = { + val prefix = " --args " + val args = clientArgs.userArgs + val retval = new StringBuilder() + for (arg <- args){ + retval.append(prefix).append(" '").append(arg).append("' ") + } + retval.toString + } + + def createContainerLaunchContext( + newApp: GetNewApplicationResponse, + localResources: HashMap[String, LocalResource], + env: HashMap[String, String]): ContainerLaunchContext = { + logInfo("Setting up container launch context") + val amContainer = Records.newRecord(classOf[ContainerLaunchContext]) + amContainer.setLocalResources(localResources) + amContainer.setEnvironment(env) + + // TODO: Need a replacement for the following code to fix -Xmx? + // val minResMemory: Int = newApp.getMinimumResourceCapability().getMemory() + // var amMemory = ((args.amMemory / minResMemory) * minResMemory) + + // ((if ((args.amMemory % minResMemory) == 0) 0 else minResMemory) - + // YarnAllocationHandler.MEMORY_OVERHEAD) + + // Extra options for the JVM + var JAVA_OPTS = "" + + // Add Xmx for AM memory + JAVA_OPTS += "-Xmx" + args.amMemory + "m" + + val tmpDir = new Path(Environment.PWD.$(), YarnConfiguration.DEFAULT_CONTAINER_TEMP_DIR) + JAVA_OPTS += " -Djava.io.tmpdir=" + tmpDir + + // TODO: Remove once cpuset version is pushed out. + // The context is, default gc for server class machines ends up using all cores to do gc - + // hence if there are multiple containers in same node, Spark GC affects all other containers' + // performance (which can be that of other Spark containers) + // Instead of using this, rely on cpusets by YARN to enforce "proper" Spark behavior in + // multi-tenant environments. Not sure how default Java GC behaves if it is limited to subset + // of cores on a node. + val useConcurrentAndIncrementalGC = env.isDefinedAt("SPARK_USE_CONC_INCR_GC") && + java.lang.Boolean.parseBoolean(env("SPARK_USE_CONC_INCR_GC")) + if (useConcurrentAndIncrementalGC) { + // In our expts, using (default) throughput collector has severe perf ramifications in + // multi-tenant machines + JAVA_OPTS += " -XX:+UseConcMarkSweepGC " + JAVA_OPTS += " -XX:+CMSIncrementalMode " + JAVA_OPTS += " -XX:+CMSIncrementalPacing " + JAVA_OPTS += " -XX:CMSIncrementalDutyCycleMin=0 " + JAVA_OPTS += " -XX:CMSIncrementalDutyCycle=10 " + } + + if (env.isDefinedAt("SPARK_JAVA_OPTS")) { + JAVA_OPTS += " " + env("SPARK_JAVA_OPTS") + } + + // Command for the ApplicationMaster + 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 " + + JAVA_OPTS + + " " + args.amClass + + " --class " + args.userClass + + " --jar " + args.userJar + + userArgsToString(args) + + " --worker-memory " + args.workerMemory + + " --worker-cores " + args.workerCores + + " --num-workers " + args.numWorkers + + " 1> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stdout" + + " 2> " + ApplicationConstants.LOG_DIR_EXPANSION_VAR + "/stderr") + + logInfo("Command for starting the Spark ApplicationMaster: " + commands(0)) + amContainer.setCommands(commands) + + // Setup security tokens. + val dob = new DataOutputBuffer() + credentials.writeTokenStorageToStream(dob) + amContainer.setTokens(ByteBuffer.wrap(dob.getData())) + + amContainer + } + + def submitApp(appContext: ApplicationSubmissionContext) = { + // Submit the application to the applications manager. + logInfo("Submitting application to ASM") + super.submitApplication(appContext) + } + + def monitorApplication(appId: ApplicationId): Boolean = { + val interval = sparkConf.getLong("spark.yarn.report.interval", 1000) + + while (true) { + Thread.sleep(interval) + val report = super.getApplicationReport(appId) + + logInfo("Application report from ASM: \n" + + "\t application identifier: " + appId.toString() + "\n" + + "\t appId: " + appId.getId() + "\n" + + "\t clientToAMToken: " + report.getClientToAMToken() + "\n" + + "\t appDiagnostics: " + report.getDiagnostics() + "\n" + + "\t appMasterHost: " + report.getHost() + "\n" + + "\t appQueue: " + report.getQueue() + "\n" + + "\t appMasterRpcPort: " + report.getRpcPort() + "\n" + + "\t appStartTime: " + report.getStartTime() + "\n" + + "\t yarnAppState: " + report.getYarnApplicationState() + "\n" + + "\t distributedFinalState: " + report.getFinalApplicationStatus() + "\n" + + "\t appTrackingUrl: " + report.getTrackingUrl() + "\n" + + "\t appUser: " + report.getUser() + ) + + val state = report.getYarnApplicationState() + val dsStatus = report.getFinalApplicationStatus() + if (state == YarnApplicationState.FINISHED || + state == YarnApplicationState.FAILED || + state == YarnApplicationState.KILLED) { + return true + } + } + true + } +} + +object Client { + val SPARK_JAR: String = "spark.jar" + val APP_JAR: String = "app.jar" + val LOG4J_PROP: String = "log4j.properties" + + def main(argStrings: Array[String]) { + // Set an env variable indicating we are running in YARN mode. + // Note: anything env variable with SPARK_ prefix gets propagated to all (remote) processes - + // see Client#setupLaunchEnv(). + System.setProperty("SPARK_YARN_MODE", "true") + + val args = new ClientArguments(argStrings) + + (new Client(args)).run() + } + + // Based on code from org.apache.hadoop.mapreduce.v2.util.MRApps + def populateHadoopClasspath(conf: Configuration, env: HashMap[String, String]) { + for (c <- conf.getStrings(YarnConfiguration.YARN_APPLICATION_CLASSPATH)) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, c.trim) + } + } + + def populateClasspath(conf: Configuration, addLog4j: Boolean, env: HashMap[String, String]) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$()) + // If log4j present, ensure ours overrides all others + if (addLog4j) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + LOG4J_PROP) + } + // Normally the users app.jar is last in case conflicts with spark jars + val userClasspathFirst = new SparkConf().get("spark.yarn.user.classpath.first", "false") + .toBoolean + if (userClasspathFirst) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + APP_JAR) + } + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + SPARK_JAR) + Client.populateHadoopClasspath(conf, env) + + if (!userClasspathFirst) { + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + APP_JAR) + } + Apps.addToEnvironment(env, Environment.CLASSPATH.name, Environment.PWD.$() + + Path.SEPARATOR + "*") + } +} diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala new file mode 100644 index 0000000000..4d9cca0766 --- /dev/null +++ b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerLauncher.scala @@ -0,0 +1,227 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.Socket +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.net.NetUtils +import org.apache.hadoop.yarn.api._ +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.util.{ConverterUtils, Records} +import akka.actor._ +import akka.remote._ +import akka.actor.Terminated +import org.apache.spark.{SparkConf, SparkContext, Logging} +import org.apache.spark.util.{Utils, AkkaUtils} +import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend +import org.apache.spark.scheduler.SplitInfo +import org.apache.hadoop.yarn.client.api.AMRMClient +import org.apache.hadoop.yarn.client.api.AMRMClient.ContainerRequest + +class WorkerLauncher(args: ApplicationMasterArguments, conf: Configuration) extends Logging { + + def this(args: ApplicationMasterArguments) = this(args, new Configuration()) + + private var appAttemptId: ApplicationAttemptId = _ + private var reporterThread: Thread = _ + private val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + + private var yarnAllocator: YarnAllocationHandler = _ + private var driverClosed:Boolean = false + + private var amClient: AMRMClient[ContainerRequest] = _ + private val sparkConf = new SparkConf + + val actorSystem : ActorSystem = AkkaUtils.createActorSystem("sparkYarnAM", Utils.localHostName, 0, + conf = sparkConf)._1 + var actor: ActorRef = _ + + // This actor just working as a monitor to watch on Driver Actor. + class MonitorActor(driverUrl: String) extends Actor { + + var driver: ActorSelection = _ + + override def preStart() { + logInfo("Listen to driver: " + driverUrl) + driver = context.actorSelection(driverUrl) + // Send a hello message thus the connection is actually established, thus we can monitor Lifecycle Events. + driver ! "Hello" + context.system.eventStream.subscribe(self, classOf[RemotingLifecycleEvent]) + } + + override def receive = { + case x: DisassociatedEvent => + logInfo("Driver terminated or disconnected! Shutting down. $x") + driverClosed = true + } + } + + def run() { + + amClient = AMRMClient.createAMRMClient() + amClient.init(yarnConf) + amClient.start() + + appAttemptId = getApplicationAttemptId() + registerApplicationMaster() + + waitForSparkMaster() + + // Allocate all containers + allocateWorkers() + + // Launch a progress reporter thread, else app will get killed after expiration (def: 10mins) timeout + // ensure that progress is sent before YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS elapse. + + val timeoutInterval = yarnConf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS, 120000) + // must be <= timeoutInterval/ 2. + // On other hand, also ensure that we are reasonably responsive without causing too many requests to RM. + // so atleast 1 minute or timeoutInterval / 10 - whichever is higher. + val interval = math.min(timeoutInterval / 2, math.max(timeoutInterval/ 10, 60000L)) + reporterThread = launchReporterThread(interval) + + // Wait for the reporter thread to Finish. + reporterThread.join() + + finishApplicationMaster(FinalApplicationStatus.SUCCEEDED) + actorSystem.shutdown() + + logInfo("Exited") + System.exit(0) + } + + private def getApplicationAttemptId(): ApplicationAttemptId = { + val envs = System.getenv() + val containerIdString = envs.get(ApplicationConstants.Environment.CONTAINER_ID.name()) + val containerId = ConverterUtils.toContainerId(containerIdString) + val appAttemptId = containerId.getApplicationAttemptId() + logInfo("ApplicationAttemptId: " + appAttemptId) + appAttemptId + } + + private def registerApplicationMaster(): RegisterApplicationMasterResponse = { + logInfo("Registering the ApplicationMaster") + // TODO:(Raymond) Find out Spark UI address and fill in here? + amClient.registerApplicationMaster(Utils.localHostName(), 0, "") + } + + private def waitForSparkMaster() { + logInfo("Waiting for Spark driver to be reachable.") + var driverUp = false + val hostport = args.userArgs(0) + val (driverHost, driverPort) = Utils.parseHostPort(hostport) + while(!driverUp) { + try { + val socket = new Socket(driverHost, driverPort) + socket.close() + logInfo("Driver now available: %s:%s".format(driverHost, driverPort)) + driverUp = true + } catch { + case e: Exception => + logError("Failed to connect to driver at %s:%s, retrying ...". + format(driverHost, driverPort)) + Thread.sleep(100) + } + } + sparkConf.set("spark.driver.host", driverHost) + sparkConf.set("spark.driver.port", driverPort.toString) + + val driverUrl = "akka.tcp://spark@%s:%s/user/%s".format( + driverHost, driverPort.toString, CoarseGrainedSchedulerBackend.ACTOR_NAME) + + actor = actorSystem.actorOf(Props(new MonitorActor(driverUrl)), name = "YarnAM") + } + + + private def allocateWorkers() { + + // Fixme: should get preferredNodeLocationData from SparkContext, just fake a empty one for now. + val preferredNodeLocationData: scala.collection.Map[String, scala.collection.Set[SplitInfo]] = + scala.collection.immutable.Map() + + yarnAllocator = YarnAllocationHandler.newAllocator( + yarnConf, + amClient, + appAttemptId, + args, + preferredNodeLocationData, + sparkConf) + + logInfo("Allocating " + args.numWorkers + " workers.") + // Wait until all containers have finished + // TODO: This is a bit ugly. Can we make it nicer? + // TODO: Handle container failure + + yarnAllocator.addResourceRequests(args.numWorkers) + while(yarnAllocator.getNumWorkersRunning < args.numWorkers) { + yarnAllocator.allocateResources() + Thread.sleep(100) + } + + logInfo("All workers have launched.") + + } + + // TODO: We might want to extend this to allocate more containers in case they die ! + private def launchReporterThread(_sleepTime: Long): Thread = { + val sleepTime = if (_sleepTime <= 0 ) 0 else _sleepTime + + val t = new Thread { + override def run() { + while (!driverClosed) { + val missingWorkerCount = args.numWorkers - yarnAllocator.getNumWorkersRunning - + yarnAllocator.getNumPendingAllocate + if (missingWorkerCount > 0) { + logInfo("Allocating %d containers to make up for (potentially) lost containers". + format(missingWorkerCount)) + yarnAllocator.addResourceRequests(missingWorkerCount) + } + sendProgress() + Thread.sleep(sleepTime) + } + } + } + // setting to daemon status, though this is usually not a good idea. + t.setDaemon(true) + t.start() + logInfo("Started progress reporter thread - sleep time : " + sleepTime) + t + } + + private def sendProgress() { + logDebug("Sending progress") + // simulated with an allocate request with no nodes requested ... + yarnAllocator.allocateResources() + } + + def finishApplicationMaster(status: FinalApplicationStatus) { + logInfo("finish ApplicationMaster with " + status) + amClient.unregisterApplicationMaster(status, "" /* appMessage */, "" /* appTrackingUrl */) + } + +} + + +object WorkerLauncher { + def main(argStrings: Array[String]) { + val args = new ApplicationMasterArguments(argStrings) + new WorkerLauncher(args).run() + } +} diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala new file mode 100644 index 0000000000..9f5523c4b9 --- /dev/null +++ b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/WorkerRunnable.scala @@ -0,0 +1,209 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.records.impl.pb.ProtoUtils +import org.apache.hadoop.yarn.api.protocolrecords._ +import org.apache.hadoop.yarn.client.api.NMClient +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.spark.Logging + + +class WorkerRunnable( + container: Container, + conf: Configuration, + masterAddress: String, + slaveId: String, + hostname: String, + workerMemory: Int, + workerCores: Int) + extends Runnable with Logging { + + var rpc: YarnRPC = YarnRPC.create(conf) + var nmClient: NMClient = _ + val yarnConf: YarnConfiguration = new YarnConfiguration(conf) + + def run = { + logInfo("Starting Worker Container") + nmClient = NMClient.createNMClient() + nmClient.init(yarnConf) + nmClient.start() + startContainer + } + + def startContainer = { + logInfo("Setting up ContainerLaunchContext") + + val ctx = Records.newRecord(classOf[ContainerLaunchContext]) + .asInstanceOf[ContainerLaunchContext] + + val localResources = prepareLocalResources + ctx.setLocalResources(localResources) + + val env = prepareEnvironment + ctx.setEnvironment(env) + + // 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 " + } +*/ + + val credentials = UserGroupInformation.getCurrentUser().getCredentials() + val dob = new DataOutputBuffer() + credentials.writeTokenStorageToStream(dob) + ctx.setTokens(ByteBuffer.wrap(dob.getData())) + + 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") + logInfo("Setting up worker with commands: " + commands) + ctx.setCommands(commands) + + // Send the start request to the ContainerManager + nmClient.startContainer(container, ctx) + } + + 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]() + + Client.populateClasspath(yarnConf, 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 + } + +} diff --git a/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala new file mode 100644 index 0000000000..8a9a73f5b4 --- /dev/null +++ b/yarn/stable/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocationHandler.scala @@ -0,0 +1,694 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES 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.lang.{Boolean => JBoolean} +import java.util.{Collections, Set => JSet} +import java.util.concurrent.{CopyOnWriteArrayList, ConcurrentHashMap} +import java.util.concurrent.atomic.AtomicInteger + +import scala.collection +import scala.collection.JavaConversions._ +import scala.collection.mutable.{ArrayBuffer, HashMap, HashSet} + +import org.apache.spark.{Logging, SparkConf} +import org.apache.spark.scheduler.{SplitInfo,TaskSchedulerImpl} +import org.apache.spark.scheduler.cluster.CoarseGrainedSchedulerBackend +import org.apache.spark.util.Utils + +import org.apache.hadoop.conf.Configuration +import org.apache.hadoop.yarn.api.ApplicationMasterProtocol +import org.apache.hadoop.yarn.api.records.ApplicationAttemptId +import org.apache.hadoop.yarn.api.records.{Container, ContainerId, ContainerStatus} +import org.apache.hadoop.yarn.api.records.{Priority, Resource, ResourceRequest} +import org.apache.hadoop.yarn.api.protocolrecords.{AllocateRequest, AllocateResponse} +import org.apache.hadoop.yarn.client.api.AMRMClient +import org.apache.hadoop.yarn.client.api.AMRMClient.ContainerRequest +import org.apache.hadoop.yarn.util.{RackResolver, Records} + + +object AllocationType extends Enumeration { + type AllocationType = Value + val HOST, RACK, ANY = Value +} + +// TODO: +// Too many params. +// Needs to be mt-safe +// Need to refactor this to make it 'cleaner' ... right now, all computation is reactive - should +// make it more proactive and decoupled. + +// Note that right now, we assume all node asks as uniform in terms of capabilities and priority +// Refer to http://developer.yahoo.com/blogs/hadoop/posts/2011/03/mapreduce-nextgen-scheduler/ for +// more info on how we are requesting for containers. +private[yarn] class YarnAllocationHandler( + val conf: Configuration, + val amClient: AMRMClient[ContainerRequest], + val appAttemptId: ApplicationAttemptId, + val maxWorkers: Int, + val workerMemory: Int, + val workerCores: Int, + val preferredHostToCount: Map[String, Int], + val preferredRackToCount: Map[String, Int], + val sparkConf: SparkConf) + extends Logging { + // These three are locked on allocatedHostToContainersMap. Complementary data structures + // allocatedHostToContainersMap : containers which are running : host, Set + // allocatedContainerToHostMap: container to host mapping. + private val allocatedHostToContainersMap = + new HashMap[String, collection.mutable.Set[ContainerId]]() + + private val allocatedContainerToHostMap = new HashMap[ContainerId, String]() + + // allocatedRackCount is populated ONLY if allocation happens (or decremented if this is an + // allocated node) + // As with the two data structures above, tightly coupled with them, and to be locked on + // allocatedHostToContainersMap + private val allocatedRackCount = new HashMap[String, Int]() + + // Containers which have been released. + private val releasedContainerList = new CopyOnWriteArrayList[ContainerId]() + // Containers to be released in next request to RM + private val pendingReleaseContainers = new ConcurrentHashMap[ContainerId, Boolean] + + // Number of container requests that have been sent to, but not yet allocated by the + // ApplicationMaster. + private val numPendingAllocate = new AtomicInteger() + private val numWorkersRunning = new AtomicInteger() + // Used to generate a unique id per worker + private val workerIdCounter = new AtomicInteger() + private val lastResponseId = new AtomicInteger() + private val numWorkersFailed = new AtomicInteger() + + def getNumPendingAllocate: Int = numPendingAllocate.intValue + + def getNumWorkersRunning: Int = numWorkersRunning.intValue + + def getNumWorkersFailed: Int = numWorkersFailed.intValue + + def isResourceConstraintSatisfied(container: Container): Boolean = { + container.getResource.getMemory >= (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) + } + + def releaseContainer(container: Container) { + val containerId = container.getId + pendingReleaseContainers.put(containerId, true) + amClient.releaseAssignedContainer(containerId) + } + + def allocateResources() { + // We have already set the container request. Poll the ResourceManager for a response. + // This doubles as a heartbeat if there are no pending container requests. + val progressIndicator = 0.1f + val allocateResponse = amClient.allocate(progressIndicator) + + val allocatedContainers = allocateResponse.getAllocatedContainers() + if (allocatedContainers.size > 0) { + var numPendingAllocateNow = numPendingAllocate.addAndGet(-1 * allocatedContainers.size) + + if (numPendingAllocateNow < 0) { + numPendingAllocateNow = numPendingAllocate.addAndGet(-1 * numPendingAllocateNow) + } + + logDebug(""" + Allocated containers: %d + Current worker count: %d + Containers released: %s + Containers to-be-released: %s + Cluster resources: %s + """.format( + allocatedContainers.size, + numWorkersRunning.get(), + releasedContainerList, + pendingReleaseContainers, + allocateResponse.getAvailableResources)) + + val hostToContainers = new HashMap[String, ArrayBuffer[Container]]() + + for (container <- allocatedContainers) { + if (isResourceConstraintSatisfied(container)) { + // Add the accepted `container` to the host's list of already accepted, + // allocated containers + val host = container.getNodeId.getHost + val containersForHost = hostToContainers.getOrElseUpdate(host, + new ArrayBuffer[Container]()) + containersForHost += container + } else { + // Release container, since it doesn't satisfy resource constraints. + releaseContainer(container) + } + } + + // Find the appropriate containers to use. + // TODO: Cleanup this group-by... + val dataLocalContainers = new HashMap[String, ArrayBuffer[Container]]() + val rackLocalContainers = new HashMap[String, ArrayBuffer[Container]]() + val offRackContainers = new HashMap[String, ArrayBuffer[Container]]() + + for (candidateHost <- hostToContainers.keySet) { + val maxExpectedHostCount = preferredHostToCount.getOrElse(candidateHost, 0) + val requiredHostCount = maxExpectedHostCount - allocatedContainersOnHost(candidateHost) + + val remainingContainersOpt = hostToContainers.get(candidateHost) + assert(remainingContainersOpt.isDefined) + var remainingContainers = remainingContainersOpt.get + + if (requiredHostCount >= remainingContainers.size) { + // Since we have <= required containers, add all remaining containers to + // `dataLocalContainers`. + dataLocalContainers.put(candidateHost, remainingContainers) + // There are no more free containers remaining. + remainingContainers = null + } else if (requiredHostCount > 0) { + // Container list has more containers than we need for data locality. + // Split the list into two: one based on the data local container count, + // (`remainingContainers.size` - `requiredHostCount`), and the other to hold remaining + // containers. + val (dataLocal, remaining) = remainingContainers.splitAt( + remainingContainers.size - requiredHostCount) + dataLocalContainers.put(candidateHost, dataLocal) + + // Invariant: remainingContainers == remaining + + // YARN has a nasty habit of allocating a ton of containers on a host - discourage this. + // Add each container in `remaining` to list of containers to release. If we have an + // insufficient number of containers, then the next allocation cycle will reallocate + // (but won't treat it as data local). + // TODO(harvey): Rephrase this comment some more. + for (container <- remaining) releaseContainer(container) + remainingContainers = null + } + + // For rack local containers + if (remainingContainers != null) { + val rack = YarnAllocationHandler.lookupRack(conf, candidateHost) + if (rack != null) { + val maxExpectedRackCount = preferredRackToCount.getOrElse(rack, 0) + val requiredRackCount = maxExpectedRackCount - allocatedContainersOnRack(rack) - + rackLocalContainers.getOrElse(rack, List()).size + + if (requiredRackCount >= remainingContainers.size) { + // Add all remaining containers to to `dataLocalContainers`. + dataLocalContainers.put(rack, remainingContainers) + remainingContainers = null + } else if (requiredRackCount > 0) { + // Container list has more containers that we need for data locality. + // Split the list into two: one based on the data local container count, + // (`remainingContainers.size` - `requiredHostCount`), and the other to hold remaining + // containers. + val (rackLocal, remaining) = remainingContainers.splitAt( + remainingContainers.size - requiredRackCount) + val existingRackLocal = rackLocalContainers.getOrElseUpdate(rack, + new ArrayBuffer[Container]()) + + existingRackLocal ++= rackLocal + + remainingContainers = remaining + } + } + } + + if (remainingContainers != null) { + // Not all containers have been consumed - add them to the list of off-rack containers. + offRackContainers.put(candidateHost, remainingContainers) + } + } + + // Now that we have split the containers into various groups, go through them in order: + // first host-local, then rack-local, and finally off-rack. + // Note that the list we create below tries to ensure that not all containers end up within + // a host if there is a sufficiently large number of hosts/containers. + val allocatedContainersToProcess = new ArrayBuffer[Container](allocatedContainers.size) + allocatedContainersToProcess ++= TaskSchedulerImpl.prioritizeContainers(dataLocalContainers) + allocatedContainersToProcess ++= TaskSchedulerImpl.prioritizeContainers(rackLocalContainers) + allocatedContainersToProcess ++= TaskSchedulerImpl.prioritizeContainers(offRackContainers) + + // Run each of the allocated containers. + for (container <- allocatedContainersToProcess) { + val numWorkersRunningNow = numWorkersRunning.incrementAndGet() + val workerHostname = container.getNodeId.getHost + val containerId = container.getId + + val workerMemoryOverhead = (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) + assert(container.getResource.getMemory >= workerMemoryOverhead) + + if (numWorkersRunningNow > maxWorkers) { + logInfo("""Ignoring container %s at host %s, since we already have the required number of + containers for it.""".format(containerId, workerHostname)) + releaseContainer(container) + numWorkersRunning.decrementAndGet() + } else { + val workerId = workerIdCounter.incrementAndGet().toString + val driverUrl = "akka.tcp://spark@%s:%s/user/%s".format( + sparkConf.get("spark.driver.host"), + sparkConf.get("spark.driver.port"), + CoarseGrainedSchedulerBackend.ACTOR_NAME) + + logInfo("Launching container %s for on host %s".format(containerId, workerHostname)) + + // To be safe, remove the container from `pendingReleaseContainers`. + pendingReleaseContainers.remove(containerId) + + val rack = YarnAllocationHandler.lookupRack(conf, workerHostname) + allocatedHostToContainersMap.synchronized { + val containerSet = allocatedHostToContainersMap.getOrElseUpdate(workerHostname, + new HashSet[ContainerId]()) + + containerSet += containerId + allocatedContainerToHostMap.put(containerId, workerHostname) + + if (rack != null) { + allocatedRackCount.put(rack, allocatedRackCount.getOrElse(rack, 0) + 1) + } + } + logInfo("Launching WorkerRunnable. driverUrl: %s, workerHostname: %s".format(driverUrl, workerHostname)) + val workerRunnable = new WorkerRunnable( + container, + conf, + driverUrl, + workerId, + workerHostname, + workerMemory, + workerCores) + new Thread(workerRunnable).start() + } + } + logDebug(""" + Finished allocating %s containers (from %s originally). + Current number of workers running: %d, + releasedContainerList: %s, + pendingReleaseContainers: %s + """.format( + allocatedContainersToProcess, + allocatedContainers, + numWorkersRunning.get(), + releasedContainerList, + pendingReleaseContainers)) + } + + val completedContainers = allocateResponse.getCompletedContainersStatuses() + if (completedContainers.size > 0) { + logDebug("Completed %d containers".format(completedContainers.size)) + + for (completedContainer <- completedContainers) { + val containerId = completedContainer.getContainerId + + if (pendingReleaseContainers.containsKey(containerId)) { + // YarnAllocationHandler already marked the container for release, so remove it from + // `pendingReleaseContainers`. + pendingReleaseContainers.remove(containerId) + } else { + // Decrement the number of workers running. The next iteration of the ApplicationMaster's + // reporting thread will take care of allocating. + numWorkersRunning.decrementAndGet() + logInfo("Completed container %s (state: %s, exit status: %s)".format( + containerId, + completedContainer.getState, + completedContainer.getExitStatus())) + // Hadoop 2.2.X added a ContainerExitStatus we should switch to use + // there are some exit status' we shouldn't necessarily count against us, but for + // now I think its ok as none of the containers are expected to exit + if (completedContainer.getExitStatus() != 0) { + logInfo("Container marked as failed: " + containerId) + numWorkersFailed.incrementAndGet() + } + } + + allocatedHostToContainersMap.synchronized { + if (allocatedContainerToHostMap.containsKey(containerId)) { + val hostOpt = allocatedContainerToHostMap.get(containerId) + assert(hostOpt.isDefined) + val host = hostOpt.get + + val containerSetOpt = allocatedHostToContainersMap.get(host) + assert(containerSetOpt.isDefined) + val containerSet = containerSetOpt.get + + containerSet.remove(containerId) + if (containerSet.isEmpty) { + allocatedHostToContainersMap.remove(host) + } else { + allocatedHostToContainersMap.update(host, containerSet) + } + + allocatedContainerToHostMap.remove(containerId) + + // TODO: Move this part outside the synchronized block? + val rack = YarnAllocationHandler.lookupRack(conf, host) + if (rack != null) { + val rackCount = allocatedRackCount.getOrElse(rack, 0) - 1 + if (rackCount > 0) { + allocatedRackCount.put(rack, rackCount) + } else { + allocatedRackCount.remove(rack) + } + } + } + } + } + logDebug(""" + Finished processing %d completed containers. + Current number of workers running: %d, + releasedContainerList: %s, + pendingReleaseContainers: %s + """.format( + completedContainers.size, + numWorkersRunning.get(), + releasedContainerList, + pendingReleaseContainers)) + } + } + + def createRackResourceRequests( + hostContainers: ArrayBuffer[ContainerRequest] + ): ArrayBuffer[ContainerRequest] = { + // Generate modified racks and new set of hosts under it before issuing requests. + val rackToCounts = new HashMap[String, Int]() + + for (container <- hostContainers) { + val candidateHost = container.getNodes.last + assert(YarnAllocationHandler.ANY_HOST != candidateHost) + + val rack = YarnAllocationHandler.lookupRack(conf, candidateHost) + if (rack != null) { + var count = rackToCounts.getOrElse(rack, 0) + count += 1 + rackToCounts.put(rack, count) + } + } + + val requestedContainers = new ArrayBuffer[ContainerRequest](rackToCounts.size) + for ((rack, count) <- rackToCounts) { + requestedContainers ++= createResourceRequests( + AllocationType.RACK, + rack, + count, + YarnAllocationHandler.PRIORITY) + } + + requestedContainers + } + + def allocatedContainersOnHost(host: String): Int = { + var retval = 0 + allocatedHostToContainersMap.synchronized { + retval = allocatedHostToContainersMap.getOrElse(host, Set()).size + } + retval + } + + def allocatedContainersOnRack(rack: String): Int = { + var retval = 0 + allocatedHostToContainersMap.synchronized { + retval = allocatedRackCount.getOrElse(rack, 0) + } + retval + } + + def addResourceRequests(numWorkers: Int) { + val containerRequests: List[ContainerRequest] = + if (numWorkers <= 0 || preferredHostToCount.isEmpty) { + logDebug("numWorkers: " + numWorkers + ", host preferences: " + + preferredHostToCount.isEmpty) + createResourceRequests( + AllocationType.ANY, + resource = null, + numWorkers, + YarnAllocationHandler.PRIORITY).toList + } else { + // Request for all hosts in preferred nodes and for numWorkers - + // candidates.size, request by default allocation policy. + val hostContainerRequests = new ArrayBuffer[ContainerRequest](preferredHostToCount.size) + for ((candidateHost, candidateCount) <- preferredHostToCount) { + val requiredCount = candidateCount - allocatedContainersOnHost(candidateHost) + + if (requiredCount > 0) { + hostContainerRequests ++= createResourceRequests( + AllocationType.HOST, + candidateHost, + requiredCount, + YarnAllocationHandler.PRIORITY) + } + } + val rackContainerRequests: List[ContainerRequest] = createRackResourceRequests( + hostContainerRequests).toList + + val anyContainerRequests = createResourceRequests( + AllocationType.ANY, + resource = null, + numWorkers, + YarnAllocationHandler.PRIORITY) + + val containerRequestBuffer = new ArrayBuffer[ContainerRequest]( + hostContainerRequests.size + rackContainerRequests.size() + anyContainerRequests.size) + + containerRequestBuffer ++= hostContainerRequests + containerRequestBuffer ++= rackContainerRequests + containerRequestBuffer ++= anyContainerRequests + containerRequestBuffer.toList + } + + for (request <- containerRequests) { + amClient.addContainerRequest(request) + } + + if (numWorkers > 0) { + numPendingAllocate.addAndGet(numWorkers) + logInfo("Will Allocate %d worker containers, each with %d memory".format( + numWorkers, + (workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD))) + } else { + logDebug("Empty allocation request ...") + } + + for (request <- containerRequests) { + val nodes = request.getNodes + var hostStr = if (nodes == null || nodes.isEmpty) { + "Any" + } else { + nodes.last + } + logInfo("Container request (host: %s, priority: %s, capability: %s".format( + hostStr, + request.getPriority().getPriority, + request.getCapability)) + } + } + + private def createResourceRequests( + requestType: AllocationType.AllocationType, + resource: String, + numWorkers: Int, + priority: Int + ): ArrayBuffer[ContainerRequest] = { + + // If hostname is specified, then we need at least two requests - node local and rack local. + // There must be a third request, which is ANY. That will be specially handled. + requestType match { + case AllocationType.HOST => { + assert(YarnAllocationHandler.ANY_HOST != resource) + val hostname = resource + val nodeLocal = constructContainerRequests( + Array(hostname), + racks = null, + numWorkers, + priority) + + // Add `hostname` to the global (singleton) host->rack mapping in YarnAllocationHandler. + YarnAllocationHandler.populateRackInfo(conf, hostname) + nodeLocal + } + case AllocationType.RACK => { + val rack = resource + constructContainerRequests(hosts = null, Array(rack), numWorkers, priority) + } + case AllocationType.ANY => constructContainerRequests( + hosts = null, racks = null, numWorkers, priority) + case _ => throw new IllegalArgumentException( + "Unexpected/unsupported request type: " + requestType) + } + } + + private def constructContainerRequests( + hosts: Array[String], + racks: Array[String], + numWorkers: Int, + priority: Int + ): ArrayBuffer[ContainerRequest] = { + + val memoryResource = Records.newRecord(classOf[Resource]) + memoryResource.setMemory(workerMemory + YarnAllocationHandler.MEMORY_OVERHEAD) + + val prioritySetting = Records.newRecord(classOf[Priority]) + prioritySetting.setPriority(priority) + + val requests = new ArrayBuffer[ContainerRequest]() + for (i <- 0 until numWorkers) { + requests += new ContainerRequest(memoryResource, hosts, racks, prioritySetting) + } + requests + } +} + +object YarnAllocationHandler { + + val ANY_HOST = "*" + // All requests are issued with same priority : we do not (yet) have any distinction between + // request types (like map/reduce in hadoop for example) + val PRIORITY = 1 + + // Additional memory overhead - in mb. + val MEMORY_OVERHEAD = 384 + + // Host to rack map - saved from allocation requests. We are expecting this not to change. + // Note that it is possible for this to change : and ResurceManager will indicate that to us via + // update response to allocate. But we are punting on handling that for now. + private val hostToRack = new ConcurrentHashMap[String, String]() + private val rackToHostSet = new ConcurrentHashMap[String, JSet[String]]() + + + def newAllocator( + conf: Configuration, + amClient: AMRMClient[ContainerRequest], + appAttemptId: ApplicationAttemptId, + args: ApplicationMasterArguments, + sparkConf: SparkConf + ): YarnAllocationHandler = { + new YarnAllocationHandler( + conf, + amClient, + appAttemptId, + args.numWorkers, + args.workerMemory, + args.workerCores, + Map[String, Int](), + Map[String, Int](), + sparkConf) + } + + def newAllocator( + conf: Configuration, + amClient: AMRMClient[ContainerRequest], + appAttemptId: ApplicationAttemptId, + args: ApplicationMasterArguments, + map: collection.Map[String, + collection.Set[SplitInfo]], + sparkConf: SparkConf + ): YarnAllocationHandler = { + val (hostToSplitCount, rackToSplitCount) = generateNodeToWeight(conf, map) + new YarnAllocationHandler( + conf, + amClient, + appAttemptId, + args.numWorkers, + args.workerMemory, + args.workerCores, + hostToSplitCount, + rackToSplitCount, + sparkConf) + } + + def newAllocator( + conf: Configuration, + amClient: AMRMClient[ContainerRequest], + appAttemptId: ApplicationAttemptId, + maxWorkers: Int, + workerMemory: Int, + workerCores: Int, + map: collection.Map[String, collection.Set[SplitInfo]], + sparkConf: SparkConf + ): YarnAllocationHandler = { + val (hostToCount, rackToCount) = generateNodeToWeight(conf, map) + new YarnAllocationHandler( + conf, + amClient, + appAttemptId, + maxWorkers, + workerMemory, + workerCores, + hostToCount, + rackToCount, + sparkConf) + } + + // A simple method to copy the split info map. + private def generateNodeToWeight( + conf: Configuration, + input: collection.Map[String, collection.Set[SplitInfo]] + ): (Map[String, Int], Map[String, Int]) = { + + if (input == null) { + return (Map[String, Int](), Map[String, Int]()) + } + + val hostToCount = new HashMap[String, Int] + val rackToCount = new HashMap[String, Int] + + for ((host, splits) <- input) { + val hostCount = hostToCount.getOrElse(host, 0) + hostToCount.put(host, hostCount + splits.size) + + val rack = lookupRack(conf, host) + if (rack != null){ + val rackCount = rackToCount.getOrElse(host, 0) + rackToCount.put(host, rackCount + splits.size) + } + } + + (hostToCount.toMap, rackToCount.toMap) + } + + def lookupRack(conf: Configuration, host: String): String = { + if (!hostToRack.contains(host)) { + populateRackInfo(conf, host) + } + hostToRack.get(host) + } + + def fetchCachedHostsForRack(rack: String): Option[Set[String]] = { + Option(rackToHostSet.get(rack)).map { set => + val convertedSet: collection.mutable.Set[String] = set + // TODO: Better way to get a Set[String] from JSet. + convertedSet.toSet + } + } + + def populateRackInfo(conf: Configuration, hostname: String) { + Utils.checkHost(hostname) + + if (!hostToRack.containsKey(hostname)) { + // If there are repeated failures to resolve, all to an ignore list. + val rackInfo = RackResolver.resolve(conf, hostname) + if (rackInfo != null && rackInfo.getNetworkLocation != null) { + val rack = rackInfo.getNetworkLocation + hostToRack.put(hostname, rack) + if (! rackToHostSet.containsKey(rack)) { + rackToHostSet.putIfAbsent(rack, + Collections.newSetFromMap(new ConcurrentHashMap[String, JBoolean]())) + } + rackToHostSet.get(rack).add(hostname) + + // TODO(harvey): Figure out what this comment means... + // Since RackResolver caches, we are disabling this for now ... + } /* else { + // right ? Else we will keep calling rack resolver in case we cant resolve rack info ... + hostToRack.put(hostname, null) + } */ + } + } +} -- cgit v1.2.3 From f442afc22ef2e8a10ea22d5a7b392d41a1c7fdf8 Mon Sep 17 00:00:00 2001 From: Raymond Liu Date: Thu, 2 Jan 2014 14:24:48 +0800 Subject: fix docs for yarn --- docs/running-on-yarn.md | 3 --- yarn/README.md | 4 ++-- 2 files changed, 2 insertions(+), 5 deletions(-) (limited to 'docs') diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index f47098554e..2c1b2cc294 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -116,8 +116,6 @@ For example: # Building Spark for Hadoop/YARN 2.2.x -Hadoop 2.2.x users must build Spark and publish it locally. The SBT build process handles Hadoop 2.2.x as a special case. This version of Hadoop has new YARN API changes and depends on a Protobuf version (2.5). - See [Building Spark with Maven](building-with-maven.html) for instructions on how to build Spark using the Maven process. # Important Notes @@ -126,4 +124,3 @@ See [Building Spark with Maven](building-with-maven.html) for instructions on ho - The local directories used for spark will be the local directories configured for YARN (Hadoop Yarn config yarn.nodemanager.local-dirs). If the user specifies spark.local.dir, it will be ignored. - The --files and --archives options support specifying file names with the # similar to Hadoop. For example you can specify: --files localtest.txt#appSees.txt and this will upload the file you have locally named localtest.txt into HDFS but this will be linked to by the name appSees.txt and your application should use the name as appSees.txt to reference it when running on YARN. - The --addJars option allows the SparkContext.addJar function to work if you are using it with local files. It does not need to be used if you are using it with HDFS, HTTP, HTTPS, or FTP files. -- YARN 2.2.x users cannot simply depend on the Spark packages without building Spark, as the published Spark artifacts are compiled to work with the pre 2.2 API. Those users must build Spark and publish it locally. diff --git a/yarn/README.md b/yarn/README.md index 9a7a1dd838..65ee85447e 100644 --- a/yarn/README.md +++ b/yarn/README.md @@ -1,12 +1,12 @@ # YARN DIRECTORY LAYOUT -Hadoop Yarn related codes are organized in separate directories for easy management. +Hadoop Yarn related codes are organized in separate directories to minimize duplicated code. * common : Common codes that do not depending on specific version of Hadoop. * alpha / stable : Codes that involve specific version of Hadoop YARN API. alpha represents 0.23 and 2.0.x - stable represents 2.2 and later, until the API is break again. + stable represents 2.2 and later, until the API changes again. alpha / stable will build together with common dir into a single jar -- cgit v1.2.3 From 94f2fffa23436ed66a24c705f88dabe59bf54037 Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Fri, 3 Jan 2014 14:37:42 +0530 Subject: fixed review comments --- README.md | 19 +++++++++++-------- .../main/scala/org/apache/spark/SparkContext.scala | 7 ++++++- .../test/scala/org/apache/spark/DriverSuite.scala | 9 ++++++--- .../spark/deploy/worker/ExecutorRunnerTest.scala | 8 +++++++- docs/running-on-yarn.md | 4 ++-- project/SparkBuild.scala | 14 +++++++++----- python/lib/py4j-0.8.1.zip | Bin 0 -> 809541 bytes python/lib/py4j0.7.egg | Bin 191756 -> 0 bytes .../scala/org/apache/spark/repl/SparkILoop.scala | 4 +++- 9 files changed, 44 insertions(+), 21 deletions(-) create mode 100644 python/lib/py4j-0.8.1.zip delete mode 100644 python/lib/py4j0.7.egg (limited to 'docs') diff --git a/README.md b/README.md index 22e7ab8245..8f686743f0 100644 --- a/README.md +++ b/README.md @@ -13,7 +13,7 @@ This README file only contains basic setup instructions. ## Building Spark requires Scala 2.10. The project is built using Simple Build Tool (SBT), -which can be obtained from [here](http://www.scala-sbt.org/release/docs/Getting-Started/Setup.html). To build Spark and its example programs, run: +which can be obtained from [here](http://www.scala-sbt.org). To build Spark and its example programs, run: sbt assembly @@ -38,19 +38,22 @@ locally with one thread, or "local[N]" to run locally with N threads. ## Running tests -### With sbt. (you need sbt installed) -Once you have built spark with `sbt assembly` mentioned in [Building](#Building) section. Test suits can be run as follows on *nix based systems using sbt. +### With sbt (Much faster to run compared to maven) +Once you have built spark with `sbt assembly` mentioned in [Building](#Building) section. Test suits can be run as follows using sbt. -`SPARK_HOME=$(pwd) SPARK_TESTING=1 sbt test` - -TODO: figure out instructions for windows. +`sbt test` ### With maven. +1. Export these necessary environment variables as follows. + + `export SCALA_HOME=` + + `export MAVEN_OPTS="-Xmx1512m -XX:MaxPermSize=512m"` -1. Build assembly by +2. Build assembly by `mvn package -DskipTests` -2. Run tests +3. Run tests `mvn test` ## A Note About Hadoop Versions diff --git a/core/src/main/scala/org/apache/spark/SparkContext.scala b/core/src/main/scala/org/apache/spark/SparkContext.scala index 4d6a97e255..c6f6883b01 100644 --- a/core/src/main/scala/org/apache/spark/SparkContext.scala +++ b/core/src/main/scala/org/apache/spark/SparkContext.scala @@ -169,10 +169,15 @@ 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_CLASSPATH", "SPARK_LIBRARY_PATH", "SPARK_JAVA_OPTS", "SPARK_TESTING"); + for (key <- Seq("SPARK_CLASSPATH", "SPARK_LIBRARY_PATH", "SPARK_JAVA_OPTS"); value <- Option(System.getenv(key))) { executorEnvs(key) = value } + // A workaround for SPARK_TESTING and SPARK_HOME + for { (envKey, propKey) <- Seq(("SPARK_HOME", "spark.home"), ("SPARK_TESTING", "spark.testing")) + value <- Option(System.getenv(envKey)).orElse(Option(System.getProperty(propKey)))} { + executorEnvs(envKey) = value + } // Since memory can be set with a system property too, use that executorEnvs("SPARK_MEM") = executorMemory + "m" executorEnvs ++= conf.getExecutorEnv diff --git a/core/src/test/scala/org/apache/spark/DriverSuite.scala b/core/src/test/scala/org/apache/spark/DriverSuite.scala index 6d1695eae7..89c5631ad8 100644 --- a/core/src/test/scala/org/apache/spark/DriverSuite.scala +++ b/core/src/test/scala/org/apache/spark/DriverSuite.scala @@ -30,13 +30,16 @@ import org.apache.spark.util.Utils class DriverSuite extends FunSuite with Timeouts { test("driver should exit after finishing") { - assert(System.getenv("SPARK_HOME") != null) + val sparkHome = Option(System.getenv("SPARK_HOME")) + .orElse(Option(System.getProperty("spark.home"))).get // Regression test for SPARK-530: "Spark driver process doesn't exit after finishing" val masters = Table(("master"), ("local"), ("local-cluster[2,1,512]")) forAll(masters) { (master: String) => failAfter(60 seconds) { - Utils.execute(Seq("./spark-class", "org.apache.spark.DriverWithoutCleanup", master), - new File(System.getenv("SPARK_HOME"))) + Utils.executeAndGetOutput( + Seq("./spark-class", "org.apache.spark.DriverWithoutCleanup", master), + new File(sparkHome), + Map("SPARK_TESTING" -> "1", "SPARK_HOME" -> sparkHome)) } } } diff --git a/core/src/test/scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala b/core/src/test/scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala index 4cb4ddc9cd..894a72284b 100644 --- a/core/src/test/scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala +++ b/core/src/test/scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala @@ -18,13 +18,19 @@ package org.apache.spark.deploy.worker import java.io.File + +import scala.util.Try + import org.scalatest.FunSuite + import org.apache.spark.deploy.{ExecutorState, Command, ApplicationDescription} + class ExecutorRunnerTest extends FunSuite { test("command includes appId") { def f(s:String) = new File(s) - val sparkHome = sys.env("SPARK_HOME") + val sparkHome = Try(sys.env("SPARK_HOME")).toOption + .orElse(Option(System.getProperty("spark.home"))).get val appDesc = new ApplicationDescription("app name", 8, 500, Command("foo", Seq(),Map()), sparkHome, "appUiUrl") val appId = "12345-worker321-9876" diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index 13d5fd3685..aded6430b3 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -12,7 +12,7 @@ was added to Spark in version 0.6.0, and improved in 0.7.0 and 0.8.0. We need a consolidated Spark JAR (which bundles all the required dependencies) to run Spark jobs on a YARN cluster. This can be built by setting the Hadoop version and `SPARK_YARN` environment variable, as follows: - SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true ./sbt assembly + SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt assembly The assembled JAR will be something like this: `./assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly_{{site.SPARK_VERSION}}-hadoop2.0.5.jar`. @@ -72,7 +72,7 @@ The command to launch the YARN Client is as follows: For example: # Build the Spark assembly JAR and the Spark examples JAR - $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true ./sbt assembly + $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt assembly # Configure logging $ cp conf/log4j.properties.template conf/log4j.properties diff --git a/project/SparkBuild.scala b/project/SparkBuild.scala index b335b5a20a..8290e7cf43 100644 --- a/project/SparkBuild.scala +++ b/project/SparkBuild.scala @@ -20,6 +20,7 @@ import sbt.Classpaths.publishTask import Keys._ import sbtassembly.Plugin._ import AssemblyKeys._ +import scala.util.Properties // For Sonatype publishing //import com.jsuereth.pgp.sbtplugin.PgpKeys._ @@ -68,10 +69,12 @@ object SparkBuild extends Build { // A configuration to set an alternative publishLocalConfiguration lazy val MavenCompile = config("m2r") extend(Compile) lazy val publishLocalBoth = TaskKey[Unit]("publish-local", "publish local for m2 and ivy") - + val sparkHome = System.getProperty("user.dir") + System.setProperty("spark.home", sparkHome) + System.setProperty("spark.testing", "1") // Allows build configuration to be set through environment variables - lazy val hadoopVersion = scala.util.Properties.envOrElse("SPARK_HADOOP_VERSION", DEFAULT_HADOOP_VERSION) - lazy val isNewHadoop = scala.util.Properties.envOrNone("SPARK_IS_NEW_HADOOP") match { + lazy val hadoopVersion = Properties.envOrElse("SPARK_HADOOP_VERSION", DEFAULT_HADOOP_VERSION) + lazy val isNewHadoop = Properties.envOrNone("SPARK_IS_NEW_HADOOP") match { case None => { val isNewHadoopVersion = "2.[2-9]+".r.findFirstIn(hadoopVersion).isDefined (isNewHadoopVersion|| DEFAULT_IS_NEW_HADOOP) @@ -79,7 +82,7 @@ object SparkBuild extends Build { case Some(v) => v.toBoolean } - lazy val isYarnEnabled = scala.util.Properties.envOrNone("SPARK_YARN") match { + lazy val isYarnEnabled = Properties.envOrNone("SPARK_YARN") match { case None => DEFAULT_YARN case Some(v) => v.toBoolean } @@ -112,8 +115,9 @@ object SparkBuild extends Build { // Fork new JVMs for tests and set Java options for those fork := true, + javaOptions += "-Dspark.home=" + sparkHome, + javaOptions += "-Dspark.testing=1", javaOptions += "-Xmx3g", - // Show full stack trace and duration in test cases. testOptions in Test += Tests.Argument("-oDF"), diff --git a/python/lib/py4j-0.8.1.zip b/python/lib/py4j-0.8.1.zip new file mode 100644 index 0000000000..3231e31164 Binary files /dev/null and b/python/lib/py4j-0.8.1.zip differ diff --git a/python/lib/py4j0.7.egg b/python/lib/py4j0.7.egg deleted file mode 100644 index f8a339d8ee..0000000000 Binary files a/python/lib/py4j0.7.egg and /dev/null differ diff --git a/repl/src/main/scala/org/apache/spark/repl/SparkILoop.scala b/repl/src/main/scala/org/apache/spark/repl/SparkILoop.scala index f108c70f21..7e54421b52 100644 --- a/repl/src/main/scala/org/apache/spark/repl/SparkILoop.scala +++ b/repl/src/main/scala/org/apache/spark/repl/SparkILoop.scala @@ -942,12 +942,14 @@ class SparkILoop(in0: Option[BufferedReader], protected val out: JPrintWriter, val conf = new SparkConf() .setMaster(master) .setAppName("Spark shell") - .setSparkHome(System.getenv("SPARK_HOME")) .setJars(jars) .set("spark.repl.class.uri", intp.classServer.uri) if (execUri != null) { conf.set("spark.executor.uri", execUri) } + if (System.getenv("SPARK_HOME") != null) { + conf.setSparkHome(System.getenv("SPARK_HOME")) + } sparkContext = new SparkContext(conf) echo("Created spark context..") sparkContext -- cgit v1.2.3 From 74ba97fcf75ae61833295a896dcdf775036ae1d8 Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Fri, 3 Jan 2014 15:08:01 +0530 Subject: sbin/spark-class* -> bin/spark-class* --- bin/run-example | 2 +- bin/run-example2.cmd | 2 +- bin/spark-class | 154 +++++++++++++++++++++ bin/spark-class.cmd | 23 +++ bin/spark-class2.cmd | 85 ++++++++++++ bin/spark-shell | 4 +- .../mesos/CoarseMesosSchedulerBackend.scala | 4 +- .../org/apache/spark/ui/UIWorkloadGenerator.scala | 4 +- .../test/scala/org/apache/spark/DriverSuite.scala | 2 +- docs/running-on-yarn.md | 4 +- docs/spark-standalone.md | 2 +- python/pyspark/java_gateway.py | 2 +- sbin/spark-class | 154 --------------------- sbin/spark-class.cmd | 23 --- sbin/spark-class2.cmd | 85 ------------ sbin/spark-daemon.sh | 2 +- sbin/spark-executor | 2 +- 17 files changed, 277 insertions(+), 277 deletions(-) create mode 100755 bin/spark-class create mode 100644 bin/spark-class.cmd create mode 100644 bin/spark-class2.cmd delete mode 100755 sbin/spark-class delete mode 100644 sbin/spark-class.cmd delete mode 100644 sbin/spark-class2.cmd (limited to 'docs') diff --git a/bin/run-example b/bin/run-example index a3976beb12..f2699c38a9 100755 --- a/bin/run-example +++ b/bin/run-example @@ -36,7 +36,7 @@ if [ -e "$FWDIR/conf/spark-env.sh" ] ; then fi if [ -z "$1" ]; then - echo "Usage: bin/run-example []" >&2 + echo "Usage: run-example []" >&2 exit 1 fi diff --git a/bin/run-example2.cmd b/bin/run-example2.cmd index d9c10545e7..6861334cb0 100644 --- a/bin/run-example2.cmd +++ b/bin/run-example2.cmd @@ -30,7 +30,7 @@ if exist "%FWDIR%conf\spark-env.cmd" call "%FWDIR%conf\spark-env.cmd" rem Test that an argument was given if not "x%1"=="x" goto arg_given - echo Usage: bin/run-example ^ [^] + echo Usage: run-example ^ [^] goto exit :arg_given diff --git a/bin/spark-class b/bin/spark-class new file mode 100755 index 0000000000..4e440d8729 --- /dev/null +++ b/bin/spark-class @@ -0,0 +1,154 @@ +#!/usr/bin/env bash + +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# + +cygwin=false +case "`uname`" in + CYGWIN*) cygwin=true;; +esac + +SCALA_VERSION=2.10 + +# Figure out where the Scala framework is installed +FWDIR="$(cd `dirname $0`/..; pwd)" + +# Export this as SPARK_HOME +export SPARK_HOME="$FWDIR" + +# Load environment variables from conf/spark-env.sh, if it exists +if [ -e "$FWDIR/conf/spark-env.sh" ] ; then + . $FWDIR/conf/spark-env.sh +fi + +if [ -z "$1" ]; then + echo "Usage: spark-class []" >&2 + exit 1 +fi + +# If this is a standalone cluster daemon, reset SPARK_JAVA_OPTS and SPARK_MEM to reasonable +# values for that; it doesn't need a lot +if [ "$1" = "org.apache.spark.deploy.master.Master" -o "$1" = "org.apache.spark.deploy.worker.Worker" ]; then + SPARK_MEM=${SPARK_DAEMON_MEMORY:-512m} + SPARK_DAEMON_JAVA_OPTS="$SPARK_DAEMON_JAVA_OPTS -Dspark.akka.logLifecycleEvents=true" + # Do not overwrite SPARK_JAVA_OPTS environment variable in this script + OUR_JAVA_OPTS="$SPARK_DAEMON_JAVA_OPTS" # Empty by default +else + OUR_JAVA_OPTS="$SPARK_JAVA_OPTS" +fi + + +# Add java opts for master, worker, executor. The opts maybe null +case "$1" in + 'org.apache.spark.deploy.master.Master') + OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_MASTER_OPTS" + ;; + 'org.apache.spark.deploy.worker.Worker') + OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_WORKER_OPTS" + ;; + 'org.apache.spark.executor.CoarseGrainedExecutorBackend') + OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_EXECUTOR_OPTS" + ;; + 'org.apache.spark.executor.MesosExecutorBackend') + OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_EXECUTOR_OPTS" + ;; + 'org.apache.spark.repl.Main') + OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_REPL_OPTS" + ;; +esac + +# Find the java binary +if [ -n "${JAVA_HOME}" ]; then + RUNNER="${JAVA_HOME}/bin/java" +else + if [ `command -v java` ]; then + RUNNER="java" + else + echo "JAVA_HOME is not set" >&2 + exit 1 + fi +fi + +# Set SPARK_MEM if it isn't already set since we also use it for this process +SPARK_MEM=${SPARK_MEM:-512m} +export SPARK_MEM + +# Set JAVA_OPTS to be able to load native libraries and to set heap size +JAVA_OPTS="$OUR_JAVA_OPTS" +JAVA_OPTS="$JAVA_OPTS -Djava.library.path=$SPARK_LIBRARY_PATH" +JAVA_OPTS="$JAVA_OPTS -Xms$SPARK_MEM -Xmx$SPARK_MEM" +# Load extra JAVA_OPTS from conf/java-opts, if it exists +if [ -e "$FWDIR/conf/java-opts" ] ; then + JAVA_OPTS="$JAVA_OPTS `cat $FWDIR/conf/java-opts`" +fi +export JAVA_OPTS +# Attention: when changing the way the JAVA_OPTS are assembled, the change must be reflected in ExecutorRunner.scala! + +if [ ! -f "$FWDIR/RELEASE" ]; then + # Exit if the user hasn't compiled Spark + num_jars=$(ls "$FWDIR"/assembly/target/scala-$SCALA_VERSION/ | grep "spark-assembly.*hadoop.*.jar" | wc -l) + jars_list=$(ls "$FWDIR"/assembly/target/scala-$SCALA_VERSION/ | grep "spark-assembly.*hadoop.*.jar") + if [ "$num_jars" -eq "0" ]; then + echo "Failed to find Spark assembly in $FWDIR/assembly/target/scala-$SCALA_VERSION/" >&2 + echo "You need to build Spark with 'sbt/sbt assembly' before running this program." >&2 + exit 1 + fi + if [ "$num_jars" -gt "1" ]; then + echo "Found multiple Spark assembly jars in $FWDIR/assembly/target/scala-$SCALA_VERSION:" >&2 + echo "$jars_list" + echo "Please remove all but one jar." + exit 1 + fi +fi + +TOOLS_DIR="$FWDIR"/tools +SPARK_TOOLS_JAR="" +if [ -e "$TOOLS_DIR"/target/scala-$SCALA_VERSION/*assembly*[0-9Tg].jar ]; then + # Use the JAR from the SBT build + export SPARK_TOOLS_JAR=`ls "$TOOLS_DIR"/target/scala-$SCALA_VERSION/*assembly*[0-9Tg].jar` +fi +if [ -e "$TOOLS_DIR"/target/spark-tools*[0-9Tg].jar ]; then + # Use the JAR from the Maven build + # TODO: this also needs to become an assembly! + export SPARK_TOOLS_JAR=`ls "$TOOLS_DIR"/target/spark-tools*[0-9Tg].jar` +fi + +# Compute classpath using external script +CLASSPATH=`$FWDIR/sbin/compute-classpath.sh` + +if [ "$1" == "org.apache.spark.tools.JavaAPICompletenessChecker" ]; then + CLASSPATH="$CLASSPATH:$SPARK_TOOLS_JAR" +fi + +if $cygwin; then + CLASSPATH=`cygpath -wp $CLASSPATH` + if [ "$1" == "org.apache.spark.tools.JavaAPICompletenessChecker" ]; then + export SPARK_TOOLS_JAR=`cygpath -w $SPARK_TOOLS_JAR` + fi +fi +export CLASSPATH + +if [ "$SPARK_PRINT_LAUNCH_COMMAND" == "1" ]; then + echo -n "Spark Command: " + echo "$RUNNER" -cp "$CLASSPATH" $JAVA_OPTS "$@" + echo "========================================" + echo +fi + +exec "$RUNNER" -cp "$CLASSPATH" $JAVA_OPTS "$@" + + diff --git a/bin/spark-class.cmd b/bin/spark-class.cmd new file mode 100644 index 0000000000..19850db9e1 --- /dev/null +++ b/bin/spark-class.cmd @@ -0,0 +1,23 @@ +@echo off + +rem +rem Licensed to the Apache Software Foundation (ASF) under one or more +rem contributor license agreements. See the NOTICE file distributed with +rem this work for additional information regarding copyright ownership. +rem The ASF licenses this file to You under the Apache License, Version 2.0 +rem (the "License"); you may not use this file except in compliance with +rem the License. You may obtain a copy of the License at +rem +rem http://www.apache.org/licenses/LICENSE-2.0 +rem +rem Unless required by applicable law or agreed to in writing, software +rem distributed under the License is distributed on an "AS IS" BASIS, +rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +rem See the License for the specific language governing permissions and +rem limitations under the License. +rem + +rem This is the entry point for running a Spark class. To avoid polluting +rem the environment, it just launches a new cmd to do the real work. + +cmd /V /E /C %~dp0spark-class2.cmd %* diff --git a/bin/spark-class2.cmd b/bin/spark-class2.cmd new file mode 100644 index 0000000000..460e661476 --- /dev/null +++ b/bin/spark-class2.cmd @@ -0,0 +1,85 @@ +@echo off + +rem +rem Licensed to the Apache Software Foundation (ASF) under one or more +rem contributor license agreements. See the NOTICE file distributed with +rem this work for additional information regarding copyright ownership. +rem The ASF licenses this file to You under the Apache License, Version 2.0 +rem (the "License"); you may not use this file except in compliance with +rem the License. You may obtain a copy of the License at +rem +rem http://www.apache.org/licenses/LICENSE-2.0 +rem +rem Unless required by applicable law or agreed to in writing, software +rem distributed under the License is distributed on an "AS IS" BASIS, +rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +rem See the License for the specific language governing permissions and +rem limitations under the License. +rem + +set SCALA_VERSION=2.10 + +rem Figure out where the Spark framework is installed +set FWDIR=%~dp0..\ + +rem Export this as SPARK_HOME +set SPARK_HOME=%FWDIR% + +rem Load environment variables from conf\spark-env.cmd, if it exists +if exist "%FWDIR%conf\spark-env.cmd" call "%FWDIR%conf\spark-env.cmd" + +rem Test that an argument was given +if not "x%1"=="x" goto arg_given + echo Usage: spark-class ^ [^] + goto exit +:arg_given + +set RUNNING_DAEMON=0 +if "%1"=="spark.deploy.master.Master" set RUNNING_DAEMON=1 +if "%1"=="spark.deploy.worker.Worker" set RUNNING_DAEMON=1 +if "x%SPARK_DAEMON_MEMORY%" == "x" set SPARK_DAEMON_MEMORY=512m +set SPARK_DAEMON_JAVA_OPTS=%SPARK_DAEMON_JAVA_OPTS% -Dspark.akka.logLifecycleEvents=true +if "%RUNNING_DAEMON%"=="1" set SPARK_MEM=%SPARK_DAEMON_MEMORY% +rem Do not overwrite SPARK_JAVA_OPTS environment variable in this script +if "%RUNNING_DAEMON%"=="0" set OUR_JAVA_OPTS=%SPARK_JAVA_OPTS% +if "%RUNNING_DAEMON%"=="1" set OUR_JAVA_OPTS=%SPARK_DAEMON_JAVA_OPTS% + +rem Figure out how much memory to use per executor and set it as an environment +rem variable so that our process sees it and can report it to Mesos +if "x%SPARK_MEM%"=="x" set SPARK_MEM=512m + +rem Set JAVA_OPTS to be able to load native libraries and to set heap size +set JAVA_OPTS=%OUR_JAVA_OPTS% -Djava.library.path=%SPARK_LIBRARY_PATH% -Xms%SPARK_MEM% -Xmx%SPARK_MEM% +rem Attention: when changing the way the JAVA_OPTS are assembled, the change must be reflected in ExecutorRunner.scala! + +rem Test whether the user has built Spark +if exist "%FWDIR%RELEASE" goto skip_build_test +set FOUND_JAR=0 +for %%d in ("%FWDIR%assembly\target\scala-%SCALA_VERSION%\spark-assembly*hadoop*.jar") do ( + set FOUND_JAR=1 +) +if "%FOUND_JAR%"=="0" ( + echo Failed to find Spark assembly JAR. + echo You need to build Spark with sbt\sbt assembly before running this program. + goto exit +) +:skip_build_test + +set TOOLS_DIR=%FWDIR%tools +set SPARK_TOOLS_JAR= +for %%d in ("%TOOLS_DIR%\target\scala-%SCALA_VERSION%\spark-tools*assembly*.jar") do ( + set SPARK_TOOLS_JAR=%%d +) + +rem Compute classpath using external script +set DONT_PRINT_CLASSPATH=1 +call "%FWDIR%sbin\compute-classpath.cmd" +set DONT_PRINT_CLASSPATH=0 +set CLASSPATH=%CLASSPATH%;%SPARK_TOOLS_JAR% + +rem Figure out where java is. +set RUNNER=java +if not "x%JAVA_HOME%"=="x" set RUNNER=%JAVA_HOME%\bin\java + +"%RUNNER%" -cp "%CLASSPATH%" %JAVA_OPTS% %* +:exit diff --git a/bin/spark-shell b/bin/spark-shell index bc7386db4d..e6885b51ef 100755 --- a/bin/spark-shell +++ b/bin/spark-shell @@ -90,10 +90,10 @@ if $cygwin; then # "Backspace sends ^H" setting in "Keys" section of the Mintty options # (see https://github.com/sbt/sbt/issues/562). stty -icanon min 1 -echo > /dev/null 2>&1 - $FWDIR/sbin/spark-class -Djline.terminal=unix $OPTIONS org.apache.spark.repl.Main "$@" + $FWDIR/bin/spark-class -Djline.terminal=unix $OPTIONS org.apache.spark.repl.Main "$@" stty icanon echo > /dev/null 2>&1 else - $FWDIR/sbin/spark-class $OPTIONS org.apache.spark.repl.Main "$@" + $FWDIR/bin/spark-class $OPTIONS org.apache.spark.repl.Main "$@" fi # record the exit status lest it be overwritten: diff --git a/core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/CoarseMesosSchedulerBackend.scala b/core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/CoarseMesosSchedulerBackend.scala index 0494ca8726..d46fceba89 100644 --- a/core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/CoarseMesosSchedulerBackend.scala +++ b/core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/CoarseMesosSchedulerBackend.scala @@ -127,7 +127,7 @@ private[spark] class CoarseMesosSchedulerBackend( CoarseGrainedSchedulerBackend.ACTOR_NAME) val uri = conf.get("spark.executor.uri", null) if (uri == null) { - val runScript = new File(sparkHome, "./sbin/spark-class").getCanonicalPath + val runScript = new File(sparkHome, "./bin/spark-class").getCanonicalPath command.setValue( "\"%s\" org.apache.spark.executor.CoarseGrainedExecutorBackend %s %s %s %d".format( runScript, driverUrl, offer.getSlaveId.getValue, offer.getHostname, numCores)) @@ -136,7 +136,7 @@ private[spark] class CoarseMesosSchedulerBackend( // glob the directory "correctly". val basename = uri.split('/').last.split('.').head command.setValue( - "cd %s*; ./sbin/spark-class org.apache.spark.executor.CoarseGrainedExecutorBackend %s %s %s %d" + "cd %s*; ./bin/spark-class org.apache.spark.executor.CoarseGrainedExecutorBackend %s %s %s %d" .format(basename, driverUrl, offer.getSlaveId.getValue, offer.getHostname, numCores)) command.addUris(CommandInfo.URI.newBuilder().setValue(uri)) } diff --git a/core/src/main/scala/org/apache/spark/ui/UIWorkloadGenerator.scala b/core/src/main/scala/org/apache/spark/ui/UIWorkloadGenerator.scala index f207627e52..6ba15187d9 100644 --- a/core/src/main/scala/org/apache/spark/ui/UIWorkloadGenerator.scala +++ b/core/src/main/scala/org/apache/spark/ui/UIWorkloadGenerator.scala @@ -27,7 +27,7 @@ import org.apache.spark.scheduler.SchedulingMode /** * Continuously generates jobs that expose various features of the WebUI (internal testing tool). * - * Usage: ./run spark.ui.UIWorkloadGenerator [master] + * Usage: ./bin/spark-class org.apache.spark.ui.UIWorkloadGenerator [master] [FIFO|FAIR] */ private[spark] object UIWorkloadGenerator { @@ -36,7 +36,7 @@ private[spark] object UIWorkloadGenerator { def main(args: Array[String]) { if (args.length < 2) { - println("usage: ./sbin/spark-class org.apache.spark.ui.UIWorkloadGenerator [master] [FIFO|FAIR]") + println("usage: ./bin/spark-class org.apache.spark.ui.UIWorkloadGenerator [master] [FIFO|FAIR]") System.exit(1) } diff --git a/core/src/test/scala/org/apache/spark/DriverSuite.scala b/core/src/test/scala/org/apache/spark/DriverSuite.scala index c37fd9ab62..605588f7f6 100644 --- a/core/src/test/scala/org/apache/spark/DriverSuite.scala +++ b/core/src/test/scala/org/apache/spark/DriverSuite.scala @@ -35,7 +35,7 @@ class DriverSuite extends FunSuite with Timeouts { val masters = Table(("master"), ("local"), ("local-cluster[2,1,512]")) forAll(masters) { (master: String) => failAfter(60 seconds) { - Utils.execute(Seq("./sbin/spark-class", "org.apache.spark.DriverWithoutCleanup", master), + Utils.execute(Seq("./bin/spark-class", "org.apache.spark.DriverWithoutCleanup", master), new File(System.getenv("SPARK_HOME"))) } } diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index e2b21f9cde..e21812378c 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -54,7 +54,7 @@ There are two scheduler mode that can be used to launch spark application on YAR The command to launch the YARN Client is as follows: - SPARK_JAR= ./sbin/spark-class org.apache.spark.deploy.yarn.Client \ + SPARK_JAR= ./bin/spark-class org.apache.spark.deploy.yarn.Client \ --jar \ --class \ --args \ @@ -79,7 +79,7 @@ For example: # Submit Spark's ApplicationMaster to YARN's ResourceManager, and instruct Spark to run the SparkPi example $ SPARK_JAR=./assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop2.0.5-alpha.jar \ - ./sbin/spark-class org.apache.spark.deploy.yarn.Client \ + ./bin/spark-class org.apache.spark.deploy.yarn.Client \ --jar examples/target/scala-{{site.SCALA_VERSION}}/spark-examples-assembly-{{site.SPARK_VERSION}}.jar \ --class org.apache.spark.examples.SparkPi \ --args yarn-standalone \ diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index baa0a062f7..c851833a18 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -28,7 +28,7 @@ the master's web UI, which is [http://localhost:8080](http://localhost:8080) by Similarly, you can start one or more workers and connect them to the master via: - ./sbin/spark-class org.apache.spark.deploy.worker.Worker spark://IP:PORT + ./bin/spark-class org.apache.spark.deploy.worker.Worker spark://IP:PORT Once you have started a worker, look at the master's web UI ([http://localhost:8080](http://localhost:8080) by default). You should see the new node listed there, along with its number of CPUs and memory (minus one gigabyte left for the OS). diff --git a/python/pyspark/java_gateway.py b/python/pyspark/java_gateway.py index 7243ee6861..c15add5237 100644 --- a/python/pyspark/java_gateway.py +++ b/python/pyspark/java_gateway.py @@ -31,7 +31,7 @@ def launch_gateway(): # Launch the Py4j gateway using Spark's run command so that we pick up the # proper classpath and SPARK_MEM settings from spark-env.sh on_windows = platform.system() == "Windows" - script = "./sbin/spark-class.cmd" if on_windows else "./sbin/spark-class" + script = "./bin/spark-class.cmd" if on_windows else "./bin/spark-class" command = [os.path.join(SPARK_HOME, script), "py4j.GatewayServer", "--die-on-broken-pipe", "0"] if not on_windows: diff --git a/sbin/spark-class b/sbin/spark-class deleted file mode 100755 index 4e440d8729..0000000000 --- a/sbin/spark-class +++ /dev/null @@ -1,154 +0,0 @@ -#!/usr/bin/env bash - -# -# Licensed to the Apache Software Foundation (ASF) under one or more -# contributor license agreements. See the NOTICE file distributed with -# this work for additional information regarding copyright ownership. -# The ASF licenses this file to You under the Apache License, Version 2.0 -# (the "License"); you may not use this file except in compliance with -# the License. You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -# - -cygwin=false -case "`uname`" in - CYGWIN*) cygwin=true;; -esac - -SCALA_VERSION=2.10 - -# Figure out where the Scala framework is installed -FWDIR="$(cd `dirname $0`/..; pwd)" - -# Export this as SPARK_HOME -export SPARK_HOME="$FWDIR" - -# Load environment variables from conf/spark-env.sh, if it exists -if [ -e "$FWDIR/conf/spark-env.sh" ] ; then - . $FWDIR/conf/spark-env.sh -fi - -if [ -z "$1" ]; then - echo "Usage: spark-class []" >&2 - exit 1 -fi - -# If this is a standalone cluster daemon, reset SPARK_JAVA_OPTS and SPARK_MEM to reasonable -# values for that; it doesn't need a lot -if [ "$1" = "org.apache.spark.deploy.master.Master" -o "$1" = "org.apache.spark.deploy.worker.Worker" ]; then - SPARK_MEM=${SPARK_DAEMON_MEMORY:-512m} - SPARK_DAEMON_JAVA_OPTS="$SPARK_DAEMON_JAVA_OPTS -Dspark.akka.logLifecycleEvents=true" - # Do not overwrite SPARK_JAVA_OPTS environment variable in this script - OUR_JAVA_OPTS="$SPARK_DAEMON_JAVA_OPTS" # Empty by default -else - OUR_JAVA_OPTS="$SPARK_JAVA_OPTS" -fi - - -# Add java opts for master, worker, executor. The opts maybe null -case "$1" in - 'org.apache.spark.deploy.master.Master') - OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_MASTER_OPTS" - ;; - 'org.apache.spark.deploy.worker.Worker') - OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_WORKER_OPTS" - ;; - 'org.apache.spark.executor.CoarseGrainedExecutorBackend') - OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_EXECUTOR_OPTS" - ;; - 'org.apache.spark.executor.MesosExecutorBackend') - OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_EXECUTOR_OPTS" - ;; - 'org.apache.spark.repl.Main') - OUR_JAVA_OPTS="$OUR_JAVA_OPTS $SPARK_REPL_OPTS" - ;; -esac - -# Find the java binary -if [ -n "${JAVA_HOME}" ]; then - RUNNER="${JAVA_HOME}/bin/java" -else - if [ `command -v java` ]; then - RUNNER="java" - else - echo "JAVA_HOME is not set" >&2 - exit 1 - fi -fi - -# Set SPARK_MEM if it isn't already set since we also use it for this process -SPARK_MEM=${SPARK_MEM:-512m} -export SPARK_MEM - -# Set JAVA_OPTS to be able to load native libraries and to set heap size -JAVA_OPTS="$OUR_JAVA_OPTS" -JAVA_OPTS="$JAVA_OPTS -Djava.library.path=$SPARK_LIBRARY_PATH" -JAVA_OPTS="$JAVA_OPTS -Xms$SPARK_MEM -Xmx$SPARK_MEM" -# Load extra JAVA_OPTS from conf/java-opts, if it exists -if [ -e "$FWDIR/conf/java-opts" ] ; then - JAVA_OPTS="$JAVA_OPTS `cat $FWDIR/conf/java-opts`" -fi -export JAVA_OPTS -# Attention: when changing the way the JAVA_OPTS are assembled, the change must be reflected in ExecutorRunner.scala! - -if [ ! -f "$FWDIR/RELEASE" ]; then - # Exit if the user hasn't compiled Spark - num_jars=$(ls "$FWDIR"/assembly/target/scala-$SCALA_VERSION/ | grep "spark-assembly.*hadoop.*.jar" | wc -l) - jars_list=$(ls "$FWDIR"/assembly/target/scala-$SCALA_VERSION/ | grep "spark-assembly.*hadoop.*.jar") - if [ "$num_jars" -eq "0" ]; then - echo "Failed to find Spark assembly in $FWDIR/assembly/target/scala-$SCALA_VERSION/" >&2 - echo "You need to build Spark with 'sbt/sbt assembly' before running this program." >&2 - exit 1 - fi - if [ "$num_jars" -gt "1" ]; then - echo "Found multiple Spark assembly jars in $FWDIR/assembly/target/scala-$SCALA_VERSION:" >&2 - echo "$jars_list" - echo "Please remove all but one jar." - exit 1 - fi -fi - -TOOLS_DIR="$FWDIR"/tools -SPARK_TOOLS_JAR="" -if [ -e "$TOOLS_DIR"/target/scala-$SCALA_VERSION/*assembly*[0-9Tg].jar ]; then - # Use the JAR from the SBT build - export SPARK_TOOLS_JAR=`ls "$TOOLS_DIR"/target/scala-$SCALA_VERSION/*assembly*[0-9Tg].jar` -fi -if [ -e "$TOOLS_DIR"/target/spark-tools*[0-9Tg].jar ]; then - # Use the JAR from the Maven build - # TODO: this also needs to become an assembly! - export SPARK_TOOLS_JAR=`ls "$TOOLS_DIR"/target/spark-tools*[0-9Tg].jar` -fi - -# Compute classpath using external script -CLASSPATH=`$FWDIR/sbin/compute-classpath.sh` - -if [ "$1" == "org.apache.spark.tools.JavaAPICompletenessChecker" ]; then - CLASSPATH="$CLASSPATH:$SPARK_TOOLS_JAR" -fi - -if $cygwin; then - CLASSPATH=`cygpath -wp $CLASSPATH` - if [ "$1" == "org.apache.spark.tools.JavaAPICompletenessChecker" ]; then - export SPARK_TOOLS_JAR=`cygpath -w $SPARK_TOOLS_JAR` - fi -fi -export CLASSPATH - -if [ "$SPARK_PRINT_LAUNCH_COMMAND" == "1" ]; then - echo -n "Spark Command: " - echo "$RUNNER" -cp "$CLASSPATH" $JAVA_OPTS "$@" - echo "========================================" - echo -fi - -exec "$RUNNER" -cp "$CLASSPATH" $JAVA_OPTS "$@" - - diff --git a/sbin/spark-class.cmd b/sbin/spark-class.cmd deleted file mode 100644 index 19850db9e1..0000000000 --- a/sbin/spark-class.cmd +++ /dev/null @@ -1,23 +0,0 @@ -@echo off - -rem -rem Licensed to the Apache Software Foundation (ASF) under one or more -rem contributor license agreements. See the NOTICE file distributed with -rem this work for additional information regarding copyright ownership. -rem The ASF licenses this file to You under the Apache License, Version 2.0 -rem (the "License"); you may not use this file except in compliance with -rem the License. You may obtain a copy of the License at -rem -rem http://www.apache.org/licenses/LICENSE-2.0 -rem -rem Unless required by applicable law or agreed to in writing, software -rem distributed under the License is distributed on an "AS IS" BASIS, -rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -rem See the License for the specific language governing permissions and -rem limitations under the License. -rem - -rem This is the entry point for running a Spark class. To avoid polluting -rem the environment, it just launches a new cmd to do the real work. - -cmd /V /E /C %~dp0spark-class2.cmd %* diff --git a/sbin/spark-class2.cmd b/sbin/spark-class2.cmd deleted file mode 100644 index 460e661476..0000000000 --- a/sbin/spark-class2.cmd +++ /dev/null @@ -1,85 +0,0 @@ -@echo off - -rem -rem Licensed to the Apache Software Foundation (ASF) under one or more -rem contributor license agreements. See the NOTICE file distributed with -rem this work for additional information regarding copyright ownership. -rem The ASF licenses this file to You under the Apache License, Version 2.0 -rem (the "License"); you may not use this file except in compliance with -rem the License. You may obtain a copy of the License at -rem -rem http://www.apache.org/licenses/LICENSE-2.0 -rem -rem Unless required by applicable law or agreed to in writing, software -rem distributed under the License is distributed on an "AS IS" BASIS, -rem WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -rem See the License for the specific language governing permissions and -rem limitations under the License. -rem - -set SCALA_VERSION=2.10 - -rem Figure out where the Spark framework is installed -set FWDIR=%~dp0..\ - -rem Export this as SPARK_HOME -set SPARK_HOME=%FWDIR% - -rem Load environment variables from conf\spark-env.cmd, if it exists -if exist "%FWDIR%conf\spark-env.cmd" call "%FWDIR%conf\spark-env.cmd" - -rem Test that an argument was given -if not "x%1"=="x" goto arg_given - echo Usage: spark-class ^ [^] - goto exit -:arg_given - -set RUNNING_DAEMON=0 -if "%1"=="spark.deploy.master.Master" set RUNNING_DAEMON=1 -if "%1"=="spark.deploy.worker.Worker" set RUNNING_DAEMON=1 -if "x%SPARK_DAEMON_MEMORY%" == "x" set SPARK_DAEMON_MEMORY=512m -set SPARK_DAEMON_JAVA_OPTS=%SPARK_DAEMON_JAVA_OPTS% -Dspark.akka.logLifecycleEvents=true -if "%RUNNING_DAEMON%"=="1" set SPARK_MEM=%SPARK_DAEMON_MEMORY% -rem Do not overwrite SPARK_JAVA_OPTS environment variable in this script -if "%RUNNING_DAEMON%"=="0" set OUR_JAVA_OPTS=%SPARK_JAVA_OPTS% -if "%RUNNING_DAEMON%"=="1" set OUR_JAVA_OPTS=%SPARK_DAEMON_JAVA_OPTS% - -rem Figure out how much memory to use per executor and set it as an environment -rem variable so that our process sees it and can report it to Mesos -if "x%SPARK_MEM%"=="x" set SPARK_MEM=512m - -rem Set JAVA_OPTS to be able to load native libraries and to set heap size -set JAVA_OPTS=%OUR_JAVA_OPTS% -Djava.library.path=%SPARK_LIBRARY_PATH% -Xms%SPARK_MEM% -Xmx%SPARK_MEM% -rem Attention: when changing the way the JAVA_OPTS are assembled, the change must be reflected in ExecutorRunner.scala! - -rem Test whether the user has built Spark -if exist "%FWDIR%RELEASE" goto skip_build_test -set FOUND_JAR=0 -for %%d in ("%FWDIR%assembly\target\scala-%SCALA_VERSION%\spark-assembly*hadoop*.jar") do ( - set FOUND_JAR=1 -) -if "%FOUND_JAR%"=="0" ( - echo Failed to find Spark assembly JAR. - echo You need to build Spark with sbt\sbt assembly before running this program. - goto exit -) -:skip_build_test - -set TOOLS_DIR=%FWDIR%tools -set SPARK_TOOLS_JAR= -for %%d in ("%TOOLS_DIR%\target\scala-%SCALA_VERSION%\spark-tools*assembly*.jar") do ( - set SPARK_TOOLS_JAR=%%d -) - -rem Compute classpath using external script -set DONT_PRINT_CLASSPATH=1 -call "%FWDIR%sbin\compute-classpath.cmd" -set DONT_PRINT_CLASSPATH=0 -set CLASSPATH=%CLASSPATH%;%SPARK_TOOLS_JAR% - -rem Figure out where java is. -set RUNNER=java -if not "x%JAVA_HOME%"=="x" set RUNNER=%JAVA_HOME%\bin\java - -"%RUNNER%" -cp "%CLASSPATH%" %JAVA_OPTS% %* -:exit diff --git a/sbin/spark-daemon.sh b/sbin/spark-daemon.sh index ca6b893b9b..2be2b3d7c0 100755 --- a/sbin/spark-daemon.sh +++ b/sbin/spark-daemon.sh @@ -147,7 +147,7 @@ case $startStop in spark_rotate_log "$log" echo starting $command, logging to $log cd "$SPARK_PREFIX" - nohup nice -n $SPARK_NICENESS "$SPARK_PREFIX"/sbin/spark-class $command "$@" >> "$log" 2>&1 < /dev/null & + nohup nice -n $SPARK_NICENESS "$SPARK_PREFIX"/bin/spark-class $command "$@" >> "$log" 2>&1 < /dev/null & newpid=$! echo $newpid > $pid sleep 2 diff --git a/sbin/spark-executor b/sbin/spark-executor index 214e00f6f8..de5bfab563 100755 --- a/sbin/spark-executor +++ b/sbin/spark-executor @@ -20,4 +20,4 @@ FWDIR="$(cd `dirname $0`/..; pwd)" echo "Running spark-executor with framework dir = $FWDIR" -exec $FWDIR/sbin/spark-class org.apache.spark.executor.MesosExecutorBackend +exec $FWDIR/bin/spark-class org.apache.spark.executor.MesosExecutorBackend -- cgit v1.2.3 From 8b5be0675245e206943574b8c6f6b77018b3561a Mon Sep 17 00:00:00 2001 From: Hossein Falaki Date: Fri, 3 Jan 2014 16:38:33 -0800 Subject: Added table of contents and minor fixes --- docs/mllib-guide.md | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) (limited to 'docs') diff --git a/docs/mllib-guide.md b/docs/mllib-guide.md index 0bebc41137..3fd3c91e2a 100644 --- a/docs/mllib-guide.md +++ b/docs/mllib-guide.md @@ -3,6 +3,9 @@ layout: global title: Machine Learning Library (MLlib) --- +* Table of contests +{:toc} + MLlib is a Spark implementation of some common machine learning (ML) functionality, as well associated tests and data generators. MLlib currently supports four common types of machine learning problem settings, @@ -297,9 +300,17 @@ val numIterations = 20 val model = ALS.train(ratings, 1, 20, 0.01) // Evaluate the model on rating data -//val ratesAndPreds = ratings.map{ case Rating(user, item, rate) => (rate, model.predict(user, item))} -//val MSE = ratesAndPreds.map{ case(v, p) => math.pow((v - p), 2)}.reduce(_ + _)/ratesAndPreds.count -//println("Mean Squared Error = " + MSE) +val usersProducts = ratings.map{ case Rating(user, product, rate) => (user, product)} +val predictions = model.predict(usersProducts).map{ + case Rating(user, product, rate) => ((user, product), rate) +} +val ratesAndPreds = ratings.map{ + case Rating(user, product, rate) => ((user, product), rate) +}.join(predictions) +val MSE = ratesAndPreds.map{ + case ((user, product), (r1, r2)) => math.pow((r1- r2), 2) +}.reduce(_ + _)/ratesAndPreds.count +println("Mean Squared Error = " + MSE) {% endhighlight %} If the rating matrix is derived from other source of information (i.e., it is inferred from @@ -393,7 +404,7 @@ Errors. ## Collaborative Filtering In the following example we load rating data. Each row consists of a user, a product and a rating. We use the default ALS.train() method which assumes ratings are explicit. We evaluate the recommendation -model by measuring the Mean Squared Error of rating prediction. +on one example. {% highlight python %} from pyspark.mllib.recommendation import ALS @@ -407,10 +418,7 @@ ratings = data.map(lambda line: array([float(x) for x in line.split(',')])) model = ALS.train(sc, ratings, 1, 20) # Evaluate the model on training data -#ratesAndPreds = ratings.map(lambda p: (p[2], model.predict(int(p[0]), int(p[1])))) -#MSE = valuesAndPreds.map(lambda (v, p): (v - p)**2).reduce(lambda x, y: x + y)/valuesAndPreds.count() -#print("Mean Squared Error = " + str(MSE)) - +print("predicted rating of user {0} for item {1} is {2:.6}".format(1, 2, model.predict(1, 2))) {% endhighlight %} If the rating matrix is derived from other source of information (i.e., it is inferred from other -- cgit v1.2.3 From d86dc74d796121b61ff43c632791c52dd49ff8ad Mon Sep 17 00:00:00 2001 From: Holden Karau Date: Sun, 5 Jan 2014 22:05:30 -0800 Subject: Code review feedback --- README.md | 14 ++++++++------ docs/README.md | 4 ++-- docs/_plugins/copy_api_dirs.rb | 4 ++-- docs/api.md | 2 +- docs/hadoop-third-party-distributions.md | 2 +- docs/index.md | 6 +++--- docs/python-programming-guide.md | 2 +- docs/quick-start.md | 8 ++++---- docs/running-on-yarn.md | 6 +++--- docs/scala-programming-guide.md | 2 +- make-distribution.sh | 7 +++++-- sbt/sbt | 13 ++++--------- 12 files changed, 35 insertions(+), 35 deletions(-) (limited to 'docs') diff --git a/README.md b/README.md index 2c08a4ac63..b91e4cf867 100644 --- a/README.md +++ b/README.md @@ -13,7 +13,9 @@ This README file only contains basic setup instructions. ## Building Spark requires Scala 2.10. The project is built using Simple Build Tool (SBT), -which can be obtained [here](http://www.scala-sbt.org). To build Spark and its example programs, run: +which can be obtained [here](http://www.scala-sbt.org). If SBT is installed we +will use the system version of sbt otherwise we will attempt to download it +automatically. To build Spark and its example programs, run: ./sbt/sbt assembly @@ -55,22 +57,22 @@ For Apache Hadoop versions 1.x, Cloudera CDH MRv1, and other Hadoop versions without YARN, use: # Apache Hadoop 1.2.1 - $ SPARK_HADOOP_VERSION=1.2.1 sbt assembly + $ SPARK_HADOOP_VERSION=1.2.1 sbt/sbt assembly # Cloudera CDH 4.2.0 with MapReduce v1 - $ SPARK_HADOOP_VERSION=2.0.0-mr1-cdh4.2.0 sbt assembly + $ SPARK_HADOOP_VERSION=2.0.0-mr1-cdh4.2.0 sbt/sbt assembly For Apache Hadoop 2.2.X, 2.1.X, 2.0.X, 0.23.x, Cloudera CDH MRv2, and other Hadoop versions with YARN, also set `SPARK_YARN=true`: # Apache Hadoop 2.0.5-alpha - $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt assembly + $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt/sbt assembly # Cloudera CDH 4.2.0 with MapReduce v2 - $ SPARK_HADOOP_VERSION=2.0.0-cdh4.2.0 SPARK_YARN=true sbt assembly + $ SPARK_HADOOP_VERSION=2.0.0-cdh4.2.0 SPARK_YARN=true sbt/sbt assembly # Apache Hadoop 2.2.X and newer - $ SPARK_HADOOP_VERSION=2.2.0 SPARK_YARN=true sbt assembly + $ SPARK_HADOOP_VERSION=2.2.0 SPARK_YARN=true sbt/sbt assembly When developing a Spark application, specify the Hadoop version by adding the "hadoop-client" artifact to your project's dependencies. For example, if you're diff --git a/docs/README.md b/docs/README.md index e3d6c9a5bc..dfcf753553 100644 --- a/docs/README.md +++ b/docs/README.md @@ -27,10 +27,10 @@ To mark a block of code in your markdown to be syntax highlighted by jekyll duri ## API Docs (Scaladoc and Epydoc) -You can build just the Spark scaladoc by running `sbt doc` from the SPARK_PROJECT_ROOT directory. +You can build just the Spark scaladoc by running `sbt/sbt doc` from the SPARK_PROJECT_ROOT directory. Similarly, you can build just the PySpark epydoc by running `epydoc --config epydoc.conf` from the SPARK_PROJECT_ROOT/pyspark directory. -When you run `jekyll` in the docs directory, it will also copy over the scaladoc for the various Spark subprojects into the docs directory (and then also into the _site directory). We use a jekyll plugin to run `sbt doc` before building the site so if you haven't run it (recently) it may take some time as it generates all of the scaladoc. The jekyll plugin also generates the PySpark docs using [epydoc](http://epydoc.sourceforge.net/). +When you run `jekyll` in the docs directory, it will also copy over the scaladoc for the various Spark subprojects into the docs directory (and then also into the _site directory). We use a jekyll plugin to run `sbt/sbt doc` before building the site so if you haven't run it (recently) it may take some time as it generates all of the scaladoc. The jekyll plugin also generates the PySpark docs using [epydoc](http://epydoc.sourceforge.net/). NOTE: To skip the step of building and copying over the Scala and Python API docs, run `SKIP_API=1 jekyll`. diff --git a/docs/_plugins/copy_api_dirs.rb b/docs/_plugins/copy_api_dirs.rb index ef9912c808..431de909cb 100644 --- a/docs/_plugins/copy_api_dirs.rb +++ b/docs/_plugins/copy_api_dirs.rb @@ -26,8 +26,8 @@ if not (ENV['SKIP_API'] == '1' or ENV['SKIP_SCALADOC'] == '1') curr_dir = pwd cd("..") - puts "Running sbt doc from " + pwd + "; this may take a few minutes..." - puts `sbt doc` + puts "Running sbt/sbt doc from " + pwd + "; this may take a few minutes..." + puts `sbt/sbt doc` puts "Moving back into docs dir." cd("docs") diff --git a/docs/api.md b/docs/api.md index 11e2c15324..e86d07770a 100644 --- a/docs/api.md +++ b/docs/api.md @@ -3,7 +3,7 @@ layout: global title: Spark API documentation (Scaladoc) --- -Here you can find links to the Scaladoc generated for the Spark sbt subprojects. If the following links don't work, try running `sbt doc` from the Spark project home directory. +Here you can find links to the Scaladoc generated for the Spark sbt subprojects. If the following links don't work, try running `sbt/sbt doc` from the Spark project home directory. - [Spark](api/core/index.html) - [Spark Examples](api/examples/index.html) diff --git a/docs/hadoop-third-party-distributions.md b/docs/hadoop-third-party-distributions.md index 141d475ba6..de6a2b0a43 100644 --- a/docs/hadoop-third-party-distributions.md +++ b/docs/hadoop-third-party-distributions.md @@ -12,7 +12,7 @@ with these distributions: When compiling Spark, you'll need to [set the SPARK_HADOOP_VERSION flag](index.html#a-note-about-hadoop-versions): - SPARK_HADOOP_VERSION=1.0.4 sbt assembly + SPARK_HADOOP_VERSION=1.0.4 sbt/sbt assembly The table below lists the corresponding `SPARK_HADOOP_VERSION` code for each CDH/HDP release. Note that some Hadoop releases are binary compatible across client versions. This means the pre-built Spark diff --git a/docs/index.md b/docs/index.md index bf8d1c3375..86d574daaa 100644 --- a/docs/index.md +++ b/docs/index.md @@ -17,7 +17,7 @@ Spark runs on both Windows and UNIX-like systems (e.g. Linux, Mac OS). All you n Spark uses [Simple Build Tool](http://www.scala-sbt.org), which is bundled with it. To compile the code, go into the top-level Spark directory and run - sbt assembly + sbt/sbt assembly For its Scala API, Spark {{site.SPARK_VERSION}} depends on Scala {{site.SCALA_VERSION}}. If you write applications in Scala, you will need to use this same version of Scala in your own program -- newer major versions may not work. You can get the right version of Scala from [scala-lang.org](http://www.scala-lang.org/download/). @@ -56,12 +56,12 @@ Hadoop, you must build Spark against the same version that your cluster uses. By default, Spark links to Hadoop 1.0.4. You can change this by setting the `SPARK_HADOOP_VERSION` variable when compiling: - SPARK_HADOOP_VERSION=2.2.0 sbt assembly + SPARK_HADOOP_VERSION=2.2.0 sbt/sbt assembly In addition, if you wish to run Spark on [YARN](running-on-yarn.html), set `SPARK_YARN` to `true`: - SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt assembly + SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt/sbt assembly Note that on Windows, you need to set the environment variables on separate lines, e.g., `set SPARK_HADOOP_VERSION=1.2.1`. diff --git a/docs/python-programming-guide.md b/docs/python-programming-guide.md index 5d48cb676a..dc187b3efe 100644 --- a/docs/python-programming-guide.md +++ b/docs/python-programming-guide.md @@ -69,7 +69,7 @@ The script automatically adds the `bin/pyspark` package to the `PYTHONPATH`. The `bin/pyspark` script launches a Python interpreter that is configured to run PySpark applications. To use `pyspark` interactively, first build Spark, then launch it directly from the command line without any options: {% highlight bash %} -$ sbt assembly +$ sbt/sbt assembly $ ./bin/pyspark {% endhighlight %} diff --git a/docs/quick-start.md b/docs/quick-start.md index 9b9261cfff..153081bdaa 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -12,7 +12,7 @@ See the [programming guide](scala-programming-guide.html) for a more complete re To follow along with this guide, you only need to have successfully built Spark on one machine. Simply go into your Spark directory and run: {% highlight bash %} -$ sbt assembly +$ sbt/sbt assembly {% endhighlight %} # Interactive Analysis with the Spark Shell @@ -146,7 +146,7 @@ If you also wish to read data from Hadoop's HDFS, you will also need to add a de libraryDependencies += "org.apache.hadoop" % "hadoop-client" % "" {% endhighlight %} -Finally, for sbt to work correctly, we'll need to layout `SimpleApp.scala` and `simple.sbt` according to the typical directory structure. Once that is in place, we can create a JAR package containing the application's code, then use `sbt run` to execute our program. +Finally, for sbt to work correctly, we'll need to layout `SimpleApp.scala` and `simple.sbt` according to the typical directory structure. Once that is in place, we can create a JAR package containing the application's code, then use `sbt/sbt run` to execute our program. {% highlight bash %} $ find . @@ -157,8 +157,8 @@ $ find . ./src/main/scala ./src/main/scala/SimpleApp.scala -$ sbt package -$ sbt run +$ sbt/sbt package +$ sbt/sbt run ... Lines with a: 46, Lines with b: 23 {% endhighlight %} diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index a35e003cdc..717071d72c 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -12,7 +12,7 @@ was added to Spark in version 0.6.0, and improved in 0.7.0 and 0.8.0. We need a consolidated Spark JAR (which bundles all the required dependencies) to run Spark jobs on a YARN cluster. This can be built by setting the Hadoop version and `SPARK_YARN` environment variable, as follows: - SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt assembly + SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt/sbt assembly The assembled JAR will be something like this: `./assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly_{{site.SPARK_VERSION}}-hadoop2.0.5.jar`. @@ -25,7 +25,7 @@ The build process now also supports new YARN versions (2.2.x). See below. - The assembled jar can be installed into HDFS or used locally. - Your application code must be packaged into a separate JAR file. -If you want to test out the YARN deployment mode, you can use the current Spark examples. A `spark-examples_{{site.SCALA_VERSION}}-{{site.SPARK_VERSION}}` file can be generated by running `sbt assembly`. NOTE: since the documentation you're reading is for Spark version {{site.SPARK_VERSION}}, we are assuming here that you have downloaded Spark {{site.SPARK_VERSION}} or checked it out of source control. If you are using a different version of Spark, the version numbers in the jar generated by the sbt package command will obviously be different. +If you want to test out the YARN deployment mode, you can use the current Spark examples. A `spark-examples_{{site.SCALA_VERSION}}-{{site.SPARK_VERSION}}` file can be generated by running `sbt/sbt assembly`. NOTE: since the documentation you're reading is for Spark version {{site.SPARK_VERSION}}, we are assuming here that you have downloaded Spark {{site.SPARK_VERSION}} or checked it out of source control. If you are using a different version of Spark, the version numbers in the jar generated by the sbt package command will obviously be different. # Configuration @@ -72,7 +72,7 @@ The command to launch the YARN Client is as follows: For example: # Build the Spark assembly JAR and the Spark examples JAR - $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt assembly + $ SPARK_HADOOP_VERSION=2.0.5-alpha SPARK_YARN=true sbt/sbt assembly # Configure logging $ cp conf/log4j.properties.template conf/log4j.properties diff --git a/docs/scala-programming-guide.md b/docs/scala-programming-guide.md index 3d0e8923d5..c1ef46a1cd 100644 --- a/docs/scala-programming-guide.md +++ b/docs/scala-programming-guide.md @@ -31,7 +31,7 @@ In addition, if you wish to access an HDFS cluster, you need to add a dependency artifactId = hadoop-client version = -For other build systems, you can run `sbt assembly` to pack Spark and its dependencies into one JAR (`assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop*.jar`), then add this to your CLASSPATH. Set the HDFS version as described [here](index.html#a-note-about-hadoop-versions). +For other build systems, you can run `sbt/sbt assembly` to pack Spark and its dependencies into one JAR (`assembly/target/scala-{{site.SCALA_VERSION}}/spark-assembly-{{site.SPARK_VERSION}}-hadoop*.jar`), then add this to your CLASSPATH. Set the HDFS version as described [here](index.html#a-note-about-hadoop-versions). Finally, you need to import some Spark classes and implicit conversions into your program. Add the following lines: diff --git a/make-distribution.sh b/make-distribution.sh index 6c466c8a06..61e6654dcb 100755 --- a/make-distribution.sh +++ b/make-distribution.sh @@ -44,13 +44,16 @@ DISTDIR="$FWDIR/dist" # Get version from SBT export TERM=dumb # Prevents color codes in SBT output -if ! test `which sbt` ;then +VERSIONSTRING=$FWDIR/sbt/sbt "show version" + +if [ $? == -1 ] ;then echo -e "You need sbt installed and available on your path." echo -e "Download sbt from http://www.scala-sbt.org/" exit -1; fi -VERSION=$(sbt "show version" | tail -1 | cut -f 2 | sed 's/^\([a-zA-Z0-9.-]*\).*/\1/') +VERSION=$(echo "${VERSIONSTRING}" | tail -1 | cut -f 2 | sed 's/^\([a-zA-Z0-9.-]*\).*/\1/') +echo "Version is ${VERSION}" # Initialize defaults SPARK_HADOOP_VERSION=1.0.4 diff --git a/sbt/sbt b/sbt/sbt index 6d2caca120..09cc5a0b4a 100755 --- a/sbt/sbt +++ b/sbt/sbt @@ -27,22 +27,17 @@ else wget --progress=bar ${URL1} -O ${JAR} || wget --progress=bar ${URL2} -O ${JAR} else printf "You do not have curl or wget installed, please install sbt manually from http://www.scala-sbt.org/\n" - exit + exit -1 fi fi if [ ! -f ${JAR} ]; then # We failed to download - printf "Our attempt to download sbt locally to {$JAR} failed. Please install sbt manually from http://www.scala-sbt.org/\n" - exit + printf "Our attempt to download sbt locally to ${JAR} failed. Please install sbt manually from http://www.scala-sbt.org/\n" + exit -1 fi printf "Launching sbt from .sbtlib\n" java \ - -Duser.timezone=UTC \ - -Djava.awt.headless=true \ - -Dfile.encoding=UTF-8 \ - -XX:MaxPermSize=256m \ - -Xmx1g \ - -noverify \ + -Xmx1200m -XX:MaxPermSize=350m -XX:ReservedCodeCacheSize=256m \ -jar ${JAR} \ "$@" fi -- cgit v1.2.3 From 2dd4fb5698220bc33acb878254d41704221573bd Mon Sep 17 00:00:00 2001 From: Andrew Ash Date: Mon, 6 Jan 2014 09:01:46 -0800 Subject: Clarify spark.cores.max It controls the count of cores across the cluster, not on a per-machine basis. --- docs/configuration.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'docs') diff --git a/docs/configuration.md b/docs/configuration.md index 567aba07f0..09342fedfc 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -81,7 +81,8 @@ there are at least five properties that you will commonly want to control: When running on a standalone deploy cluster or a Mesos cluster in "coarse-grained" - sharing mode, how many CPU cores to request at most. The default will use all available cores + sharing mode, the maximum amount of CPU cores to request for the application from + across the cluster (not from each machine). The default will use all available cores offered by the cluster manager. -- cgit v1.2.3 From 150089dae12bbba693db4edbfcea360b443637df Mon Sep 17 00:00:00 2001 From: Hossein Falaki Date: Mon, 6 Jan 2014 12:43:17 -0800 Subject: Added proper evaluation example for collaborative filtering and fixed typo --- docs/mllib-guide.md | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'docs') diff --git a/docs/mllib-guide.md b/docs/mllib-guide.md index 3fd3c91e2a..5f3b676126 100644 --- a/docs/mllib-guide.md +++ b/docs/mllib-guide.md @@ -3,7 +3,7 @@ layout: global title: Machine Learning Library (MLlib) --- -* Table of contests +* Table of contents {:toc} MLlib is a Spark implementation of some common machine learning (ML) @@ -403,8 +403,8 @@ Errors. ## Collaborative Filtering In the following example we load rating data. Each row consists of a user, a product and a rating. -We use the default ALS.train() method which assumes ratings are explicit. We evaluate the recommendation -on one example. +We use the default ALS.train() method which assumes ratings are explicit. We evaluate the +recommendation by measuring the Mean Squared Error of rating prediction. {% highlight python %} from pyspark.mllib.recommendation import ALS @@ -418,7 +418,11 @@ ratings = data.map(lambda line: array([float(x) for x in line.split(',')])) model = ALS.train(sc, ratings, 1, 20) # Evaluate the model on training data -print("predicted rating of user {0} for item {1} is {2:.6}".format(1, 2, model.predict(1, 2))) +testdata = ratings.map(lambda p: (int(p[0]), int(p[1]))) +predictions = model.predictAll(testdata).map(lambda r: ((r[0], r[1]), r[2])) +ratesAndPreds = ratings.map(lambda r: ((r[0], r[1]), r[2])).join(predictions) +MSE = ratesAndPreds.map(lambda r: (r[1][0] - r[1][1])**2).reduce(lambda x, y: x + y)/ratesAndPreds.count() +print("Mean Squared Error = " + str(MSE)) {% endhighlight %} If the rating matrix is derived from other source of information (i.e., it is inferred from other -- cgit v1.2.3 From 67af8031368c4a3558aeb46c290b7b7a09516539 Mon Sep 17 00:00:00 2001 From: Raymond Liu Date: Mon, 6 Jan 2014 16:12:25 +0800 Subject: Export --file for YarnClient mode to support sending extra files to worker on yarn cluster --- docs/running-on-yarn.md | 2 ++ .../apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala | 4 +++- 2 files changed, 5 insertions(+), 1 deletion(-) (limited to 'docs') diff --git a/docs/running-on-yarn.md b/docs/running-on-yarn.md index a35e003cdc..da5a502fd5 100644 --- a/docs/running-on-yarn.md +++ b/docs/running-on-yarn.md @@ -114,6 +114,8 @@ For example: SPARK_YARN_APP_JAR=examples/target/scala-{{site.SCALA_VERSION}}/spark-examples-assembly-{{site.SPARK_VERSION}}.jar \ MASTER=yarn-client ./bin/spark-shell +You can also send extra files to yarn cluster for worker to use by exporting SPARK_YARN_DIST_FILES=file1,file2... etc. + # Building Spark for Hadoop/YARN 2.2.x See [Building Spark with Maven](building-with-maven.html) for instructions on how to build Spark using the Maven process. diff --git a/yarn/common/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala b/yarn/common/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala index 324ef4616f..4b1b5da048 100644 --- a/yarn/common/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala +++ b/yarn/common/src/main/scala/org/apache/spark/scheduler/cluster/YarnClientSchedulerBackend.scala @@ -39,6 +39,7 @@ private[spark] class YarnClientSchedulerBackend( val defaultWorkerNumber = "1" val userJar = System.getenv("SPARK_YARN_APP_JAR") + val distFiles = System.getenv("SPARK_YARN_DIST_FILES") var workerCores = System.getenv("SPARK_WORKER_CORES") var workerMemory = System.getenv("SPARK_WORKER_MEMORY") var workerNumber = System.getenv("SPARK_WORKER_INSTANCES") @@ -64,7 +65,8 @@ private[spark] class YarnClientSchedulerBackend( "--worker-memory", workerMemory, "--worker-cores", workerCores, "--num-workers", workerNumber, - "--master-class", "org.apache.spark.deploy.yarn.WorkerLauncher" + "--master-class", "org.apache.spark.deploy.yarn.WorkerLauncher", + "--files", distFiles ) val args = new ClientArguments(argsArray, conf) -- cgit v1.2.3 From b72cceba2727586c1e1f89c58b66417628e1afa7 Mon Sep 17 00:00:00 2001 From: Patrick Wendell Date: Mon, 6 Jan 2014 22:05:53 -0800 Subject: Some doc fixes --- docs/spark-standalone.md | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'docs') diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index f426db0141..7da64749b7 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -157,8 +157,7 @@ You may also run your application entirely inside of the cluster by submitting y [application-options] cluster-url: The URL of the master node. - application-jar-url: Path to a bundled jar including your application and all dependencies. - Accepts hdfs://, file://, and http:// paths. + application-jar-url: Path to a bundled jar including your application and all dependencies. Currently, the URL must be visible from inside of your cluster, for instance, in an HDFS directory. main-class: The entry point for your application. Client Options: @@ -170,7 +169,7 @@ Keep in mind that your driver program will be executed on a remote worker machin * _Environment variables_: These will be captured from the environment in which you launch the client and applied when launching the driver program. * _Java options_: You can add java options by setting `SPARK_JAVA_OPTS` in the environment in which you launch the submission client. - * _Dependencies_: You'll still need to call `sc.addJar` inside of your driver program to add your application jar and any dependencies. If you submit a local application jar to the client (e.g one with a `file://` URL), it will be uploaded into the working directory of your driver program. Then, you can add it using `sc.addJar("jar-name.jar")`. + * _Dependencies_: You'll still need to call `sc.addJar` inside of your program to make your bundled application jar visible on all worker nodes. Once you submit a driver program, it will appear in the cluster management UI at port 8080 and be assigned an identifier. If you'd like to prematurely terminate the program, you can do so using -- cgit v1.2.3 From b3018811e106e6414816380a35c07a8564945d37 Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Mon, 6 Jan 2014 15:47:40 +0530 Subject: Allow users to set arbitrary akka configurations via spark conf. --- core/src/main/scala/org/apache/spark/SparkConf.scala | 7 +++++++ core/src/main/scala/org/apache/spark/util/AkkaUtils.scala | 8 +++++--- docs/configuration.md | 8 ++++++++ 3 files changed, 20 insertions(+), 3 deletions(-) (limited to 'docs') diff --git a/core/src/main/scala/org/apache/spark/SparkConf.scala b/core/src/main/scala/org/apache/spark/SparkConf.scala index 55f27033b5..2d437f1b21 100644 --- a/core/src/main/scala/org/apache/spark/SparkConf.scala +++ b/core/src/main/scala/org/apache/spark/SparkConf.scala @@ -172,6 +172,13 @@ class SparkConf(loadDefaults: Boolean) extends Serializable with Cloneable with .map{case (k, v) => (k.substring(prefix.length), v)} } + /** Get all akka conf variables set on this SparkConf */ + def getAkkaConf: Seq[(String, String)] = { + getAll.filter { + case (k, v) => k.startsWith("akka.") + } + } + /** Does the configuration contain a given parameter? */ def contains(key: String): Boolean = settings.contains(key) diff --git a/core/src/main/scala/org/apache/spark/util/AkkaUtils.scala b/core/src/main/scala/org/apache/spark/util/AkkaUtils.scala index 7df7e3d8e5..2ee37815de 100644 --- a/core/src/main/scala/org/apache/spark/util/AkkaUtils.scala +++ b/core/src/main/scala/org/apache/spark/util/AkkaUtils.scala @@ -17,12 +17,13 @@ package org.apache.spark.util +import scala.collection.JavaConversions.mapAsJavaMap import scala.concurrent.duration.{Duration, FiniteDuration} import akka.actor.{ActorSystem, ExtendedActorSystem, IndestructibleActorSystem} import com.typesafe.config.ConfigFactory -import org.apache.log4j.{Level, Logger} +import org.apache.log4j.{Level, Logger} import org.apache.spark.SparkConf /** @@ -64,7 +65,8 @@ private[spark] object AkkaUtils { conf.get("spark.akka.failure-detector.threshold", "300.0").toDouble val akkaHeartBeatInterval = conf.get("spark.akka.heartbeat.interval", "1000").toInt - val akkaConf = ConfigFactory.parseString( + val akkaConf = ConfigFactory.parseMap(conf.getAkkaConf.toMap[String, String]).withFallback( + ConfigFactory.parseString( s""" |akka.daemonic = on |akka.loggers = [""akka.event.slf4j.Slf4jLogger""] @@ -86,7 +88,7 @@ private[spark] object AkkaUtils { |akka.remote.log-remote-lifecycle-events = $lifecycleEvents |akka.log-dead-letters = $lifecycleEvents |akka.log-dead-letters-during-shutdown = $lifecycleEvents - """.stripMargin) + """.stripMargin)) val actorSystem = if (indestructible) { IndestructibleActorSystem(name, akkaConf) diff --git a/docs/configuration.md b/docs/configuration.md index 09342fedfc..8a8857bb3b 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -360,6 +360,14 @@ Apart from these, the following properties are also available, and may be useful Too large a value decreases parallelism during broadcast (makes it slower); however, if it is too small, BlockManager might take a performance hit. + + akka.x.y.... + value + + An arbitrary akka configuration can be set directly on spark conf and it is applied for all the ActorSystems created spark wide for that spark context and its assigned executors as well. + + + spark.shuffle.consolidateFiles false -- cgit v1.2.3 From b84dc780d3ad3fc2f0f1e9ba3fd1676447c65e96 Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Mon, 6 Jan 2014 15:51:32 +0530 Subject: Allow configuration to be printed in logs for diagnosis. --- core/src/main/scala/org/apache/spark/SparkContext.scala | 4 ++++ docs/configuration.md | 7 +++++++ 2 files changed, 11 insertions(+) (limited to 'docs') diff --git a/core/src/main/scala/org/apache/spark/SparkContext.scala b/core/src/main/scala/org/apache/spark/SparkContext.scala index e80e43af6d..99dcced7d7 100644 --- a/core/src/main/scala/org/apache/spark/SparkContext.scala +++ b/core/src/main/scala/org/apache/spark/SparkContext.scala @@ -116,6 +116,10 @@ class SparkContext( throw new SparkException("An application must be set in your configuration") } + if (conf.get("spark.log-conf", "false").toBoolean) { + logInfo("Spark configuration:\n" + conf.toDebugString) + } + // Set Spark driver host and port system properties conf.setIfMissing("spark.driver.host", Utils.localHostName()) conf.setIfMissing("spark.driver.port", "0") diff --git a/docs/configuration.md b/docs/configuration.md index 8a8857bb3b..f00fe05476 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -403,6 +403,13 @@ Apart from these, the following properties are also available, and may be useful How many times slower a task is than the median to be considered for speculation. + + spark.log-conf + false + + Log the supplied SparkConf as INFO at start of spark context. + + ## Viewing Spark Properties -- cgit v1.2.3 From c729fa7c8ed733a778a7201ed17bf74f3e132845 Mon Sep 17 00:00:00 2001 From: Prashant Sharma Date: Tue, 7 Jan 2014 13:08:16 +0530 Subject: formatting related fixes suggested by Patrick. --- core/src/main/scala/org/apache/spark/SparkConf.scala | 6 +----- docs/configuration.md | 2 +- 2 files changed, 2 insertions(+), 6 deletions(-) (limited to 'docs') diff --git a/core/src/main/scala/org/apache/spark/SparkConf.scala b/core/src/main/scala/org/apache/spark/SparkConf.scala index 2d437f1b21..b166527614 100644 --- a/core/src/main/scala/org/apache/spark/SparkConf.scala +++ b/core/src/main/scala/org/apache/spark/SparkConf.scala @@ -173,11 +173,7 @@ class SparkConf(loadDefaults: Boolean) extends Serializable with Cloneable with } /** Get all akka conf variables set on this SparkConf */ - def getAkkaConf: Seq[(String, String)] = { - getAll.filter { - case (k, v) => k.startsWith("akka.") - } - } + def getAkkaConf: Seq[(String, String)] = getAll.filter {case (k, v) => k.startsWith("akka.")} /** Does the configuration contain a given parameter? */ def contains(key: String): Boolean = settings.contains(key) diff --git a/docs/configuration.md b/docs/configuration.md index f00fe05476..1d36ecb9c1 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -364,7 +364,7 @@ Apart from these, the following properties are also available, and may be useful akka.x.y.... value - An arbitrary akka configuration can be set directly on spark conf and it is applied for all the ActorSystems created spark wide for that spark context and its assigned executors as well. + An arbitrary akka configuration can be set directly on spark conf and it is applied for all the ActorSystems created spark wide for that SparkContext and its assigned executors as well. -- cgit v1.2.3 From d8bcc8e9a095c1b20dd7a17b6535800d39bff80e Mon Sep 17 00:00:00 2001 From: Matei Zaharia Date: Tue, 7 Jan 2014 14:35:52 -0500 Subject: Add way to limit default # of cores used by applications on standalone mode Also documents the spark.deploy.spreadOut option. --- .../main/scala/org/apache/spark/SparkConf.scala | 7 ++++- .../main/scala/org/apache/spark/SparkContext.scala | 2 +- .../spark/deploy/master/ApplicationInfo.scala | 7 +++-- .../org/apache/spark/deploy/master/Master.scala | 8 ++++-- docs/configuration.md | 33 +++++++++++++++++++--- docs/css/bootstrap.min.css | 2 +- docs/job-scheduling.md | 5 ++-- docs/spark-standalone.md | 10 +++++++ 8 files changed, 60 insertions(+), 14 deletions(-) (limited to 'docs') diff --git a/core/src/main/scala/org/apache/spark/SparkConf.scala b/core/src/main/scala/org/apache/spark/SparkConf.scala index b166527614..2de32231e8 100644 --- a/core/src/main/scala/org/apache/spark/SparkConf.scala +++ b/core/src/main/scala/org/apache/spark/SparkConf.scala @@ -67,7 +67,7 @@ class SparkConf(loadDefaults: Boolean) extends Serializable with Cloneable with /** Set JAR files to distribute to the cluster. */ def setJars(jars: Seq[String]): SparkConf = { - for (jar <- jars if (jar == null)) logWarning("null jar passed to SparkContext constructor") + for (jar <- jars if (jar == null)) logWarning("null jar passed to SparkContext constructor") set("spark.jars", jars.filter(_ != null).mkString(",")) } @@ -165,6 +165,11 @@ class SparkConf(loadDefaults: Boolean) extends Serializable with Cloneable with getOption(key).map(_.toDouble).getOrElse(defaultValue) } + /** Get a parameter as a boolean, falling back to a default if not set */ + def getBoolean(key: String, defaultValue: Boolean): Boolean = { + getOption(key).map(_.toBoolean).getOrElse(defaultValue) + } + /** Get all executor environment variables set on this SparkConf */ def getExecutorEnv: Seq[(String, String)] = { val prefix = "spark.executorEnv." diff --git a/core/src/main/scala/org/apache/spark/SparkContext.scala b/core/src/main/scala/org/apache/spark/SparkContext.scala index 99dcced7d7..0e47f4e442 100644 --- a/core/src/main/scala/org/apache/spark/SparkContext.scala +++ b/core/src/main/scala/org/apache/spark/SparkContext.scala @@ -116,7 +116,7 @@ class SparkContext( throw new SparkException("An application must be set in your configuration") } - if (conf.get("spark.log-conf", "false").toBoolean) { + if (conf.get("spark.logConf", "false").toBoolean) { logInfo("Spark configuration:\n" + conf.toDebugString) } diff --git a/core/src/main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala b/core/src/main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala index 5150b7c7de..1321d9200b 100644 --- a/core/src/main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala +++ b/core/src/main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala @@ -28,7 +28,8 @@ private[spark] class ApplicationInfo( val desc: ApplicationDescription, val submitDate: Date, val driver: ActorRef, - val appUiUrl: String) + val appUiUrl: String, + defaultCores: Int) extends Serializable { @transient var state: ApplicationState.Value = _ @@ -81,7 +82,9 @@ private[spark] class ApplicationInfo( } } - def coresLeft: Int = desc.maxCores - coresGranted + private val myMaxCores = if (desc.maxCores == Int.MaxValue) defaultCores else desc.maxCores + + def coresLeft: Int = myMaxCores - coresGranted private var _retryCount = 0 diff --git a/core/src/main/scala/org/apache/spark/deploy/master/Master.scala b/core/src/main/scala/org/apache/spark/deploy/master/Master.scala index 7b696cfcca..ee01fb11df 100644 --- a/core/src/main/scala/org/apache/spark/deploy/master/Master.scala +++ b/core/src/main/scala/org/apache/spark/deploy/master/Master.scala @@ -88,7 +88,10 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act // As a temporary workaround before better ways of configuring memory, we allow users to set // a flag that will perform round-robin scheduling across the nodes (spreading out each app // among all the nodes) instead of trying to consolidate each app onto a small # of nodes. - val spreadOutApps = conf.get("spark.deploy.spreadOut", "true").toBoolean + val spreadOutApps = conf.getBoolean("spark.deploy.spreadOut", true) + + // Default maxCores for applications that don't specify it (i.e. pass Int.MaxValue) + val defaultCores = conf.getInt("spark.deploy.defaultCores", Int.MaxValue) override def preStart() { logInfo("Starting Spark master at " + masterUrl) @@ -426,7 +429,8 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act def createApplication(desc: ApplicationDescription, driver: ActorRef): ApplicationInfo = { val now = System.currentTimeMillis() val date = new Date(now) - new ApplicationInfo(now, newApplicationId(date), desc, date, driver, desc.appUiUrl) + new ApplicationInfo( + now, newApplicationId(date), desc, date, driver, desc.appUiUrl, defaultCores) } def registerApplication(app: ApplicationInfo): Unit = { diff --git a/docs/configuration.md b/docs/configuration.md index 1d36ecb9c1..52ed59be30 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -77,13 +77,14 @@ there are at least five properties that you will commonly want to control: spark.cores.max - (infinite) + (not set) When running on a standalone deploy cluster or a Mesos cluster in "coarse-grained" sharing mode, the maximum amount of CPU cores to request for the application from - across the cluster (not from each machine). The default will use all available cores - offered by the cluster manager. + across the cluster (not from each machine). If not set, the default will be + spark.deploy.defaultCores on Spark's standalone cluster manager, or + infinite (all available cores) on Mesos. @@ -404,12 +405,36 @@ Apart from these, the following properties are also available, and may be useful - spark.log-conf + spark.logConf false Log the supplied SparkConf as INFO at start of spark context. + + spark.deploy.spreadOut + true + + Whether the standalone cluster manager should spread applications out across nodes or try + to consolidate them onto as few nodes as possible. Spreading out is usually better for + data locality in HDFS, but consolidating is more efficient for compute-intensive workloads.
+ Note: this setting needs to be configured in the cluster master, not in individual + applications; you can set it through SPARK_JAVA_OPTS in spark-env.sh. + + + + spark.deploy.defaultCores + (infinite) + + Default number of cores to give to applications in Spark's standalone mode if they don't + set spark.cores.max. If not set, applications always get all available + cores unless they configure spark.cores.max themselves. + Set this lower on a shared cluster to prevent users from grabbing + the whole cluster by default.
+ Note: this setting needs to be configured in the cluster master, not in individual + applications; you can set it through SPARK_JAVA_OPTS in spark-env.sh. + + ## Viewing Spark Properties diff --git a/docs/css/bootstrap.min.css b/docs/css/bootstrap.min.css index 119bb5dba7..30ea440c00 100644 --- a/docs/css/bootstrap.min.css +++ b/docs/css/bootstrap.min.css @@ -6,4 +6,4 @@ * 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{height:auto;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;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}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{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;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 .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:20px;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}h1,h2,h3,h4,h5,h6{margin:10px 0;font-family:inherit;font-weight:bold;line-height:1;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{font-size:36px;line-height:40px}h2{font-size:30px;line-height:40px}h3{font-size:24px;line-height:40px}h4{font-size:18px;line-height:20px}h5{font-size:14px;line-height:20px}h6{font-size:12px;line-height:20px}h1 small{font-size:24px}h2 small{font-size:18px}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}dl{margin-bottom:20px}dt,dd{line-height:20px}dt{font-weight:bold}dd{margin-left:10px}.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:20px 0;border:0;border-top:1px solid #eee;border-bottom:1px solid #fff}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 20px;border-left:5px solid #eee}blockquote p{margin-bottom:0;font-size:16px;font-weight:300;line-height:25px}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;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;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:9px;font-size:14px;line-height:20px;color:#555;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}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-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;cursor:pointer}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 #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}.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: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[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"]{float:left}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>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;-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 .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: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>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;-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 .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: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>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;-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 .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: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}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: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{margin-bottom:5px;font-size:0;white-space:nowrap}.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;font-size:14px;vertical-align:top;-webkit-border-radius:0 3px 3px 0;-moz-border-radius:0 3px 3px 0;border-radius:0 3px 3px 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{margin-left:-1px;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: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 .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}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: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:10px;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: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-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{-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,.table-bordered tfoot: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,.table-bordered tfoot: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-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-right-topleft: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 [class*=span],.row-fluid table [class*=span]{display:table-cell;float:none;margin-left:0}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}.table tbody tr.success td{background-color:#dff0d8}.table tbody tr.error td{background-color:#f2dede}.table tbody tr.info td{background-color:#d9edf7}[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>.active>a>[class^="icon-"],.nav>.active>a>[class*=" icon-"],.dropdown-menu>li>a:hover>[class^="icon-"],.dropdown-menu>li>a:hover>[class*=" icon-"],.dropdown-menu>.active>a>[class^="icon-"],.dropdown-menu>.active>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{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 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{color:#fff;text-decoration:none;background-color:#0088cc;background-color:#0088cc;background-image:-moz-linear-gradient(top,#0088cc,#0087b3);background-image:-webkit-gradient(linear,0 0,0 100%,from(#0088cc),to(#0087b3));background-image:-webkit-linear-gradient(top,#0088cc,#0087b3);background-image:-o-linear-gradient(top,#0088cc,#0087b3);background-image:linear-gradient(to bottom,#0088cc,#0087b3);background-repeat:repeat-x;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#ff0088cc',endColorstr='#ff0087b3',GradientType=0)}.dropdown-menu .active>a,.dropdown-menu .active>a:hover{color:#fff;text-decoration:none;background-color:#0088cc;background-color:#0081c2;background-image:linear-gradient(to bottom,#0088cc,#0087b3);background-image:-moz-linear-gradient(top,#0088cc,#0087b3);background-image:-webkit-gradient(linear,0 0,0 100%,from(#0088cc),to(#0087b3));background-image:-webkit-linear-gradient(top,#0088cc,#0087b3);background-image:-o-linear-gradient(top,#0088cc,#0087b3);background-repeat:repeat-x;outline:0;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#ff0088cc',endColorstr='#ff0087b3',GradientType=0)}.dropdown-menu .disabled>a,.dropdown-menu .disabled>a:hover{color:#999}.dropdown-menu .disabled>a:hover{text-decoration:none;cursor:default;background-color:transparent}.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}.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}.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 .dropdown-menu .nav-header{padding-right:20px;padding-left:20px}.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 #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;overflow:visible \9;-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{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 14px;margin-bottom:0;*margin-left:.3em;font-size:14px;line-height:20px;*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:-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-image:-moz-linear-gradient(top,#fff,#e6e6e6);background-repeat:repeat-x;border:1px solid #bbb;*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:#a2a2a2;-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: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{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;-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:16px;line-height:normal;-webkit-border-radius:5px;-moz-border-radius:5px;border-radius:5px}.btn-large [class^="icon-"]{margin-top:2px}.btn-small{padding:3px 9px;font-size:12px;line-height:18px}.btn-small [class^="icon-"]{margin-top:0}.btn-mini{padding:2px 6px;font-size:11px;line-height:16px}.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}.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:#c5c5c5;border-color:rgba(0,0,0,0.15) rgba(0,0,0,0.15) rgba(0,0,0,0.25)}.btn-primary{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#006dcc;*background-color:#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-image:-moz-linear-gradient(top,#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: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:-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-image:-moz-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='#fffbb450',endColorstr='#fff89406',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-warning:hover,.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:-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-image:-moz-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='#ffee5f5b',endColorstr='#ffbd362f',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-danger:hover,.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:-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-image:-moz-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='#ff62c462',endColorstr='#ff51a351',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-success:hover,.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:-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-image:-moz-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='#ff5bc0de',endColorstr='#ff2f96b4',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-info:hover,.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:-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-image:-moz-linear-gradient(top,#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: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{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{color:#005580;text-decoration:underline;background-color:transparent}.btn-group{position:relative;*margin-left:.3em;font-size:0;white-space:nowrap}.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-group{display:inline-block;*display:inline;*zoom:1}.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{font-size:14px}.btn-group>.btn-mini{font-size:11px}.btn-group>.btn-small{font-size:12px}.btn-group>.btn-large{font-size:16px}.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-mini .caret,.btn-small .caret,.btn-large .caret{margin-top:6px}.btn-large .caret{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}.btn-group-vertical{display:inline-block;*display:inline;*zoom:1}.btn-group-vertical .btn{display:block;float:none;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;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 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-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:20px;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-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{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: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{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-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{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{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{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;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{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}.nav>.disabled>a{color:#999}.nav>.disabled>a:hover{text-decoration:none;cursor:default;background-color:transparent}.navbar{*position:relative;*z-index:2;margin-bottom:20px;overflow:visible;color:#555}.navbar-inner{min-height:40px;padding-right:20px;padding-left:20px;background-color:#fafafa;background-image:-moz-linear-gradient(top,#fff,#feecda);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fff),to(#feecda));background-image:-webkit-linear-gradient(top,#fff,#feecda);background-image:-o-linear-gradient(top,#fff,#feecda);background-image:linear-gradient(to bottom,#fff,#feecda);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='#fffeecda',GradientType=0);-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 .container{width:auto}.nav-collapse.collapse{height:auto}.navbar .brand{display:block;float:left;padding:10px 20px 10px;margin-left:-20px;font-size:20px;font-weight:200;color:#555;text-shadow:0 1px 0 #fff}.navbar .brand:hover{text-decoration:none}.navbar-text{margin-bottom:0;line-height:40px}.navbar-link{color:#555}.navbar-link:hover{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:6px}.navbar .btn-group .btn{margin: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: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: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;width:100%;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-fixed-bottom .navbar-inner,.navbar-static-top .navbar-inner{border: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:inset 0 -1px 0 rgba(0,0,0,0.1),0 1px 10px rgba(0,0,0,0.1);-moz-box-shadow:inset 0 -1px 0 rgba(0,0,0,0.1),0 1px 10px rgba(0,0,0,0.1);box-shadow:inset 0 -1px 0 rgba(0,0,0,0.1),0 1px 10px rgba(0,0,0,0.1)}.navbar-fixed-bottom{bottom:0}.navbar-fixed-bottom .navbar-inner{-webkit-box-shadow:inset 0 1px 0 rgba(0,0,0,0.1),0 -1px 10px rgba(0,0,0,0.1);-moz-box-shadow:inset 0 1px 0 rgba(0,0,0,0.1),0 -1px 10px rgba(0,0,0,0.1);box-shadow:inset 0 1px 0 rgba(0,0,0,0.1),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}.navbar .nav>li{float:left}.navbar .nav>li>a{float:none;padding:10px 15px 10px;color:#555;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:-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-image:-moz-linear-gradient(top,#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: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.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:#555;border-bottom-color:#555}.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{color:#999}.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{color:#fff}.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{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>.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:-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-image:-moz-linear-gradient(top,#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: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 .divider{padding:0 5px;color:#ccc}.breadcrumb .active{color:#999}.pagination{height:40px;margin:20px 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,.pagination span{float:left;padding:0 14px;line-height:38px;text-decoration:none;background-color:#fff;border:1px solid #ddd;border-left-width:0}.pagination a:hover,.pagination .active a,.pagination .active span{background-color:#f5f5f5}.pagination .active a,.pagination .active span{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,.pagination li:first-child span{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,.pagination li:last-child span{-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: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 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:2080}.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;-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-header h3{margin:0;line-height:30px}.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;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}.tooltip{position:absolute;z-index:1030;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:-3px}.tooltip.right{margin-left:3px}.tooltip.bottom{margin-top:3px}.tooltip.left{margin-left:-3px}.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;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;width:236px;padding:1px;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-bottom:10px}.popover.right{margin-left:10px}.popover.bottom{margin-top:10px}.popover.left{margin-right: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-content{padding:9px 14px}.popover-content p,.popover-content ul,.popover-content ol{margin-bottom:0}.popover .arrow,.popover .arrow:after{position:absolute;display:inline-block;width:0;height:0;border-color:transparent;border-style:solid}.popover .arrow:after{z-index:-1;content:""}.popover.top .arrow{bottom:-10px;left:50%;margin-left:-10px;border-top-color:#fff;border-width:10px 10px 0}.popover.top .arrow:after{bottom:-1px;left:-11px;border-top-color:rgba(0,0,0,0.25);border-width:11px 11px 0}.popover.right .arrow{top:50%;left:-10px;margin-top:-10px;border-right-color:#fff;border-width:10px 10px 10px 0}.popover.right .arrow:after{bottom:-11px;left:-1px;border-right-color:rgba(0,0,0,0.25);border-width:11px 11px 11px 0}.popover.bottom .arrow{top:-10px;left:50%;margin-left:-10px;border-bottom-color:#fff;border-width:0 10px 10px}.popover.bottom .arrow:after{top:-1px;left:-11px;border-bottom-color:rgba(0,0,0,0.25);border-width:0 11px 11px}.popover.left .arrow{top:50%;right:-10px;margin-top:-10px;border-left-color:#fff;border-width:10px 0 10px 10px}.popover.left .arrow:after{right:-1px;bottom:-11px;border-left-color:rgba(0,0,0,0.25);border-width:11px 0 11px 11px}.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{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}.label,.badge{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{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}.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 .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 .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: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;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:30px;color:inherit}.pull-right{float:right}.pull-left{float:left}.hide{display:none}.show{display:block}.invisible{visibility:hidden}.affix{position:fixed} + */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{height:auto;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;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}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{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;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 .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:20px;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}h1,h2,h3,h4,h5,h6{margin:10px 0;font-family:inherit;font-weight:bold;line-height:1;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{font-size:36px;line-height:40px}h2{font-size:30px;line-height:40px}h3{font-size:24px;line-height:40px}h4{font-size:18px;line-height:20px}h5{font-size:14px;line-height:20px}h6{font-size:12px;line-height:20px}h1 small{font-size:24px}h2 small{font-size:18px}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}dl{margin-bottom:20px}dt,dd{line-height:20px}dt{font-weight:bold}dd{margin-left:10px}.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:20px 0;border:0;border-top:1px solid #eee;border-bottom:1px solid #fff}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 20px;border-left:5px solid #eee}blockquote p{margin-bottom:0;font-size:16px;font-weight:300;line-height:25px}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;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;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:9px;font-size:14px;line-height:20px;color:#555;-webkit-border-radius:3px;-moz-border-radius:3px;border-radius:3px}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-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;cursor:pointer}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 #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}.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: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[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"]{float:left}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>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;-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 .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: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>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;-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 .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: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>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;-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 .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: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}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: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{margin-bottom:5px;font-size:0;white-space:nowrap}.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;font-size:14px;vertical-align:top;-webkit-border-radius:0 3px 3px 0;-moz-border-radius:0 3px 3px 0;border-radius:0 3px 3px 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{margin-left:-1px;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: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 .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}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: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:10px;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: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-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{-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,.table-bordered tfoot: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,.table-bordered tfoot: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-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-right-topleft: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 [class*=span],.row-fluid table [class*=span]{display:table-cell;float:none;margin-left:0}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}.table tbody tr.success td{background-color:#dff0d8}.table tbody tr.error td{background-color:#f2dede}.table tbody tr.info td{background-color:#d9edf7}[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>.active>a>[class^="icon-"],.nav>.active>a>[class*=" icon-"],.dropdown-menu>li>a:hover>[class^="icon-"],.dropdown-menu>li>a:hover>[class*=" icon-"],.dropdown-menu>.active>a>[class^="icon-"],.dropdown-menu>.active>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{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 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{color:#fff;text-decoration:none;background-color:#0088cc;background-color:#0088cc;background-image:-moz-linear-gradient(top,#0088cc,#0087b3);background-image:-webkit-gradient(linear,0 0,0 100%,from(#0088cc),to(#0087b3));background-image:-webkit-linear-gradient(top,#0088cc,#0087b3);background-image:-o-linear-gradient(top,#0088cc,#0087b3);background-image:linear-gradient(to bottom,#0088cc,#0087b3);background-repeat:repeat-x;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#ff0088cc',endColorstr='#ff0087b3',GradientType=0)}.dropdown-menu .active>a,.dropdown-menu .active>a:hover{color:#fff;text-decoration:none;background-color:#0088cc;background-color:#0081c2;background-image:linear-gradient(to bottom,#0088cc,#0087b3);background-image:-moz-linear-gradient(top,#0088cc,#0087b3);background-image:-webkit-gradient(linear,0 0,0 100%,from(#0088cc),to(#0087b3));background-image:-webkit-linear-gradient(top,#0088cc,#0087b3);background-image:-o-linear-gradient(top,#0088cc,#0087b3);background-repeat:repeat-x;outline:0;filter:progid:dximagetransform.microsoft.gradient(startColorstr='#ff0088cc',endColorstr='#ff0087b3',GradientType=0)}.dropdown-menu .disabled>a,.dropdown-menu .disabled>a:hover{color:#999}.dropdown-menu .disabled>a:hover{text-decoration:none;cursor:default;background-color:transparent}.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}.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}.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 .dropdown-menu .nav-header{padding-right:20px;padding-left:20px}.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 #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;overflow:visible \9;-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{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 14px;margin-bottom:0;*margin-left:.3em;font-size:14px;line-height:20px;*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:-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-image:-moz-linear-gradient(top,#fff,#e6e6e6);background-repeat:repeat-x;border:1px solid #bbb;*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:#a2a2a2;-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: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{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;-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:16px;line-height:normal;-webkit-border-radius:5px;-moz-border-radius:5px;border-radius:5px}.btn-large [class^="icon-"]{margin-top:2px}.btn-small{padding:3px 9px;font-size:12px;line-height:18px}.btn-small [class^="icon-"]{margin-top:0}.btn-mini{padding:2px 6px;font-size:11px;line-height:16px}.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}.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:#c5c5c5;border-color:rgba(0,0,0,0.15) rgba(0,0,0,0.15) rgba(0,0,0,0.25)}.btn-primary{color:#fff;text-shadow:0 -1px 0 rgba(0,0,0,0.25);background-color:#006dcc;*background-color:#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-image:-moz-linear-gradient(top,#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: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:-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-image:-moz-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='#fffbb450',endColorstr='#fff89406',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-warning:hover,.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:-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-image:-moz-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='#ffee5f5b',endColorstr='#ffbd362f',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-danger:hover,.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:-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-image:-moz-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='#ff62c462',endColorstr='#ff51a351',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-success:hover,.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:-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-image:-moz-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='#ff5bc0de',endColorstr='#ff2f96b4',GradientType=0);filter:progid:dximagetransform.microsoft.gradient(enabled=false)}.btn-info:hover,.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:-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-image:-moz-linear-gradient(top,#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: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{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{color:#005580;text-decoration:underline;background-color:transparent}.btn-group{position:relative;*margin-left:.3em;font-size:0;white-space:nowrap}.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-group{display:inline-block;*display:inline;*zoom:1}.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{font-size:14px}.btn-group>.btn-mini{font-size:11px}.btn-group>.btn-small{font-size:12px}.btn-group>.btn-large{font-size:16px}.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-mini .caret,.btn-small .caret,.btn-large .caret{margin-top:6px}.btn-large .caret{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}.btn-group-vertical{display:inline-block;*display:inline;*zoom:1}.btn-group-vertical .btn{display:block;float:none;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;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 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-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:20px;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-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{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: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{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-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{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{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{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;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{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}.nav>.disabled>a{color:#999}.nav>.disabled>a:hover{text-decoration:none;cursor:default;background-color:transparent}.navbar{*position:relative;*z-index:2;margin-bottom:20px;overflow:visible;color:#555}.navbar-inner{min-height:40px;padding-right:20px;padding-left:20px;background-color:#fafafa;background-image:-moz-linear-gradient(top,#fff,#feead8);background-image:-webkit-gradient(linear,0 0,0 100%,from(#fff),to(#feead8));background-image:-webkit-linear-gradient(top,#fff,#feead8);background-image:-o-linear-gradient(top,#fff,#feead8);background-image:linear-gradient(to bottom,#fff,#feead8);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='#fffeead8',GradientType=0);-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 .container{width:auto}.nav-collapse.collapse{height:auto}.navbar .brand{display:block;float:left;padding:10px 20px 10px;margin-left:-20px;font-size:20px;font-weight:200;color:#555;text-shadow:0 1px 0 #fff}.navbar .brand:hover{text-decoration:none}.navbar-text{margin-bottom:0;line-height:40px}.navbar-link{color:#555}.navbar-link:hover{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:6px}.navbar .btn-group .btn{margin: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: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: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;width:100%;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-fixed-bottom .navbar-inner,.navbar-static-top .navbar-inner{border: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:inset 0 -1px 0 rgba(0,0,0,0.1),0 1px 10px rgba(0,0,0,0.1);-moz-box-shadow:inset 0 -1px 0 rgba(0,0,0,0.1),0 1px 10px rgba(0,0,0,0.1);box-shadow:inset 0 -1px 0 rgba(0,0,0,0.1),0 1px 10px rgba(0,0,0,0.1)}.navbar-fixed-bottom{bottom:0}.navbar-fixed-bottom .navbar-inner{-webkit-box-shadow:inset 0 1px 0 rgba(0,0,0,0.1),0 -1px 10px rgba(0,0,0,0.1);-moz-box-shadow:inset 0 1px 0 rgba(0,0,0,0.1),0 -1px 10px rgba(0,0,0,0.1);box-shadow:inset 0 1px 0 rgba(0,0,0,0.1),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}.navbar .nav>li{float:left}.navbar .nav>li>a{float:none;padding:10px 15px 10px;color:#555;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:-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-image:-moz-linear-gradient(top,#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: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.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:#555;border-bottom-color:#555}.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{color:#999}.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{color:#fff}.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{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>.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:-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-image:-moz-linear-gradient(top,#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: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 .divider{padding:0 5px;color:#ccc}.breadcrumb .active{color:#999}.pagination{height:40px;margin:20px 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,.pagination span{float:left;padding:0 14px;line-height:38px;text-decoration:none;background-color:#fff;border:1px solid #ddd;border-left-width:0}.pagination a:hover,.pagination .active a,.pagination .active span{background-color:#f5f5f5}.pagination .active a,.pagination .active span{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,.pagination li:first-child span{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,.pagination li:last-child span{-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: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 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:2080}.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;-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-header h3{margin:0;line-height:30px}.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;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}.tooltip{position:absolute;z-index:1030;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:-3px}.tooltip.right{margin-left:3px}.tooltip.bottom{margin-top:3px}.tooltip.left{margin-left:-3px}.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;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;width:236px;padding:1px;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-bottom:10px}.popover.right{margin-left:10px}.popover.bottom{margin-top:10px}.popover.left{margin-right: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-content{padding:9px 14px}.popover-content p,.popover-content ul,.popover-content ol{margin-bottom:0}.popover .arrow,.popover .arrow:after{position:absolute;display:inline-block;width:0;height:0;border-color:transparent;border-style:solid}.popover .arrow:after{z-index:-1;content:""}.popover.top .arrow{bottom:-10px;left:50%;margin-left:-10px;border-top-color:#fff;border-width:10px 10px 0}.popover.top .arrow:after{bottom:-1px;left:-11px;border-top-color:rgba(0,0,0,0.25);border-width:11px 11px 0}.popover.right .arrow{top:50%;left:-10px;margin-top:-10px;border-right-color:#fff;border-width:10px 10px 10px 0}.popover.right .arrow:after{bottom:-11px;left:-1px;border-right-color:rgba(0,0,0,0.25);border-width:11px 11px 11px 0}.popover.bottom .arrow{top:-10px;left:50%;margin-left:-10px;border-bottom-color:#fff;border-width:0 10px 10px}.popover.bottom .arrow:after{top:-1px;left:-11px;border-bottom-color:rgba(0,0,0,0.25);border-width:0 11px 11px}.popover.left .arrow{top:50%;right:-10px;margin-top:-10px;border-left-color:#fff;border-width:10px 0 10px 10px}.popover.left .arrow:after{right:-1px;bottom:-11px;border-left-color:rgba(0,0,0,0.25);border-width:11px 0 11px 11px}.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{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}.label,.badge{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{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}.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 .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 .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: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;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:30px;color:inherit}.pull-right{float:right}.pull-left{float:left}.hide{display:none}.show{display:block}.invisible{visibility:hidden}.affix{position:fixed} diff --git a/docs/job-scheduling.md b/docs/job-scheduling.md index 5951155fe3..df2faa5e41 100644 --- a/docs/job-scheduling.md +++ b/docs/job-scheduling.md @@ -32,9 +32,8 @@ Resource allocation can be configured as follows, based on the cluster type: * **Standalone mode:** By default, applications submitted to the standalone mode cluster will run in FIFO (first-in-first-out) order, and each application will try to use all available nodes. You can limit - the number of nodes an application uses by setting the `spark.cores.max` configuration property in it. This - will allow multiple users/applications to run concurrently. For example, you might launch a long-running - server that uses 10 cores, and allow users to launch shells that use 20 cores each. + the number of nodes an application uses by setting the `spark.cores.max` configuration property in it, + or change the default for applications that don't set this setting through `spark.deploy.defaultCores`. Finally, in addition to controlling cores, each application's `spark.executor.memory` setting controls its memory use. * **Mesos:** To use static partitioning on Mesos, set the `spark.mesos.coarse` configuration property to `true`, diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index c851833a18..f47d41f966 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -167,6 +167,16 @@ val conf = new SparkConf() val sc = new SparkContext(conf) {% endhighlight %} +In addition, you can configure `spark.deploy.defaultCores` on the cluster master process to change the +default for applications that don't set `spark.cores.max` to something less than infinite. +Do this by adding the following to `conf/spark-env.sh`: + +{% highlight bash %} +export SPARK_JAVA_OPTS="-Dspark.deploy.defaultCores=" +{% endhighlight %} + +This is useful on shared clusters where users might not have configured a maximum number of cores +individually. # Monitoring and Logging -- cgit v1.2.3 From 2c421749eae1e3945ca34ce006addd98a0c1a00b Mon Sep 17 00:00:00 2001 From: Matei Zaharia Date: Tue, 7 Jan 2014 19:30:23 -0500 Subject: Address review comments --- .../main/scala/org/apache/spark/deploy/ApplicationDescription.scala | 2 +- core/src/main/scala/org/apache/spark/deploy/client/TestClient.scala | 2 +- .../main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala | 2 +- core/src/main/scala/org/apache/spark/deploy/master/Master.scala | 3 +++ .../apache/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala | 2 +- core/src/test/scala/org/apache/spark/deploy/JsonProtocolSuite.scala | 2 +- .../scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala | 2 +- docs/configuration.md | 4 ++-- 8 files changed, 11 insertions(+), 8 deletions(-) (limited to 'docs') diff --git a/core/src/main/scala/org/apache/spark/deploy/ApplicationDescription.scala b/core/src/main/scala/org/apache/spark/deploy/ApplicationDescription.scala index 19d393a0db..e38459b883 100644 --- a/core/src/main/scala/org/apache/spark/deploy/ApplicationDescription.scala +++ b/core/src/main/scala/org/apache/spark/deploy/ApplicationDescription.scala @@ -19,7 +19,7 @@ package org.apache.spark.deploy private[spark] class ApplicationDescription( val name: String, - val maxCores: Int, /* Integer.MAX_VALUE denotes an unlimited number of cores */ + val maxCores: Option[Int], val memoryPerSlave: Int, val command: Command, val sparkHome: String, diff --git a/core/src/main/scala/org/apache/spark/deploy/client/TestClient.scala b/core/src/main/scala/org/apache/spark/deploy/client/TestClient.scala index ef649fd80c..28ebbdc66b 100644 --- a/core/src/main/scala/org/apache/spark/deploy/client/TestClient.scala +++ b/core/src/main/scala/org/apache/spark/deploy/client/TestClient.scala @@ -48,7 +48,7 @@ private[spark] object TestClient { val (actorSystem, port) = AkkaUtils.createActorSystem("spark", Utils.localIpAddress, 0, conf = new SparkConf) val desc = new ApplicationDescription( - "TestClient", 1, 512, Command("spark.deploy.client.TestExecutor", Seq(), Map()), + "TestClient", Some(1), 512, Command("spark.deploy.client.TestExecutor", Seq(), Map()), "dummy-spark-home", "ignored") val listener = new TestListener val client = new Client(actorSystem, Array(url), desc, listener, new SparkConf) diff --git a/core/src/main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala b/core/src/main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala index 1321d9200b..3e26379166 100644 --- a/core/src/main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala +++ b/core/src/main/scala/org/apache/spark/deploy/master/ApplicationInfo.scala @@ -82,7 +82,7 @@ private[spark] class ApplicationInfo( } } - private val myMaxCores = if (desc.maxCores == Int.MaxValue) defaultCores else desc.maxCores + private val myMaxCores = desc.maxCores.getOrElse(defaultCores) def coresLeft: Int = myMaxCores - coresGranted diff --git a/core/src/main/scala/org/apache/spark/deploy/master/Master.scala b/core/src/main/scala/org/apache/spark/deploy/master/Master.scala index ee01fb11df..6617b7100f 100644 --- a/core/src/main/scala/org/apache/spark/deploy/master/Master.scala +++ b/core/src/main/scala/org/apache/spark/deploy/master/Master.scala @@ -92,6 +92,9 @@ private[spark] class Master(host: String, port: Int, webUiPort: Int) extends Act // Default maxCores for applications that don't specify it (i.e. pass Int.MaxValue) val defaultCores = conf.getInt("spark.deploy.defaultCores", Int.MaxValue) + if (defaultCores < 1) { + throw new SparkException("spark.deploy.defaultCores must be positive") + } override def preStart() { logInfo("Starting Spark master at " + masterUrl) diff --git a/core/src/main/scala/org/apache/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala b/core/src/main/scala/org/apache/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala index 9858717d13..73fc37444e 100644 --- a/core/src/main/scala/org/apache/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala +++ b/core/src/main/scala/org/apache/spark/scheduler/cluster/SparkDeploySchedulerBackend.scala @@ -38,7 +38,7 @@ private[spark] class SparkDeploySchedulerBackend( var stopping = false var shutdownCallback : (SparkDeploySchedulerBackend) => Unit = _ - val maxCores = conf.get("spark.cores.max", Int.MaxValue.toString).toInt + val maxCores = conf.getOption("spark.cores.max").map(_.toInt) override def start() { super.start() diff --git a/core/src/test/scala/org/apache/spark/deploy/JsonProtocolSuite.scala b/core/src/test/scala/org/apache/spark/deploy/JsonProtocolSuite.scala index 810ebf4140..331fa3a642 100644 --- a/core/src/test/scala/org/apache/spark/deploy/JsonProtocolSuite.scala +++ b/core/src/test/scala/org/apache/spark/deploy/JsonProtocolSuite.scala @@ -70,7 +70,7 @@ class JsonProtocolSuite extends FunSuite { def createAppDesc() : ApplicationDescription = { val cmd = new Command("mainClass", List("arg1", "arg2"), Map()) - new ApplicationDescription("name", 4, 1234, cmd, "sparkHome", "appUiUrl") + new ApplicationDescription("name", Some(4), 1234, cmd, "sparkHome", "appUiUrl") } def createAppInfo() : ApplicationInfo = { new ApplicationInfo( diff --git a/core/src/test/scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala b/core/src/test/scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala index 7e5aaa3f98..be93074b7b 100644 --- a/core/src/test/scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala +++ b/core/src/test/scala/org/apache/spark/deploy/worker/ExecutorRunnerTest.scala @@ -27,7 +27,7 @@ class ExecutorRunnerTest extends FunSuite { test("command includes appId") { def f(s:String) = new File(s) val sparkHome = sys.env.get("SPARK_HOME").orElse(sys.props.get("spark.home")).get - val appDesc = new ApplicationDescription("app name", 8, 500, Command("foo", Seq(),Map()), + val appDesc = new ApplicationDescription("app name", Some(8), 500, Command("foo", Seq(),Map()), sparkHome, "appUiUrl") val appId = "12345-worker321-9876" val er = new ExecutorRunner(appId, 1, appDesc, 8, 500, null, "blah", "worker321", f(sparkHome), diff --git a/docs/configuration.md b/docs/configuration.md index 52ed59be30..1d6c3d1633 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -418,7 +418,7 @@ Apart from these, the following properties are also available, and may be useful Whether the standalone cluster manager should spread applications out across nodes or try to consolidate them onto as few nodes as possible. Spreading out is usually better for data locality in HDFS, but consolidating is more efficient for compute-intensive workloads.
- Note: this setting needs to be configured in the cluster master, not in individual + Note: this setting needs to be configured in the standalone cluster master, not in individual applications; you can set it through SPARK_JAVA_OPTS in spark-env.sh. @@ -431,7 +431,7 @@ Apart from these, the following properties are also available, and may be useful cores unless they configure spark.cores.max themselves. Set this lower on a shared cluster to prevent users from grabbing the whole cluster by default.
- Note: this setting needs to be configured in the cluster master, not in individual + Note: this setting needs to be configured in the standalone cluster master, not in individual applications; you can set it through SPARK_JAVA_OPTS in spark-env.sh. -- cgit v1.2.3 From 82a1d38aea3b10930a2659b9c0e7ad2fb2c2ab4a Mon Sep 17 00:00:00 2001 From: Patrick Wendell Date: Tue, 7 Jan 2014 17:46:02 -0800 Subject: Simplify and fix pyspark script. This patch removes compatibility for IPython < 1.0 but fixes the launch script and makes it much simpler. I tested this using the three commands in the PySpark documentation page: 1. IPYTHON=1 ./pyspark 2. IPYTHON_OPTS="notebook" ./pyspark 3. IPYTHON_OPTS="notebook --pylab inline" ./pyspark There are two changes: - We rely on PYTHONSTARTUP env var to start PySpark - Removed the quotes around $IPYTHON_OPTS... having quotes gloms them together as a single argument passed to `exec` which seemed to cause ipython to fail (it instead expects them as multiple arguments). --- bin/pyspark | 8 +------- docs/python-programming-guide.md | 5 +++-- 2 files changed, 4 insertions(+), 9 deletions(-) (limited to 'docs') diff --git a/bin/pyspark b/bin/pyspark index d6810f4686..a70da5674e 100755 --- a/bin/pyspark +++ b/bin/pyspark @@ -51,7 +51,6 @@ export PYSPARK_PYTHON export PYTHONPATH=$SPARK_HOME/python/:$PYTHONPATH # Load the PySpark shell.py script when ./pyspark is used interactively: -export OLD_PYTHONSTARTUP=$PYTHONSTARTUP export PYTHONSTARTUP=$FWDIR/python/pyspark/shell.py if [ -n "$IPYTHON_OPTS" ]; then @@ -59,12 +58,7 @@ if [ -n "$IPYTHON_OPTS" ]; then fi if [[ "$IPYTHON" = "1" ]] ; then - # IPython <1.0.0 doesn't honor PYTHONSTARTUP, while 1.0.0+ does. - # Hence we clear PYTHONSTARTUP and use the -c "%run $IPYTHONSTARTUP" command which works on all versions - # We also force interactive mode with "-i" - IPYTHONSTARTUP=$PYTHONSTARTUP - PYTHONSTARTUP= - exec ipython "$IPYTHON_OPTS" -i -c "%run $IPYTHONSTARTUP" + exec ipython $IPYTHON_OPTS else exec "$PYSPARK_PYTHON" "$@" fi diff --git a/docs/python-programming-guide.md b/docs/python-programming-guide.md index dc187b3efe..c4236f8312 100644 --- a/docs/python-programming-guide.md +++ b/docs/python-programming-guide.md @@ -99,8 +99,9 @@ $ MASTER=local[4] ./bin/pyspark ## IPython -It is also possible to launch PySpark in [IPython](http://ipython.org), the enhanced Python interpreter. -To do this, set the `IPYTHON` variable to `1` when running `bin/pyspark`: +It is also possible to launch PySpark in [IPython](http://ipython.org), the +enhanced Python interpreter. PySpark works with IPython 1.0.0 and later. To +use IPython, set the `IPYTHON` variable to `1` when running `bin/pyspark`: {% highlight bash %} $ IPYTHON=1 ./bin/pyspark -- cgit v1.2.3 From c78b381e91c9902a1510a2ed4ec5c898b51adfe8 Mon Sep 17 00:00:00 2001 From: Patrick Wendell Date: Tue, 7 Jan 2014 23:56:04 -0800 Subject: Fixes --- core/src/main/scala/org/apache/spark/deploy/Client.scala | 2 +- .../main/scala/org/apache/spark/deploy/worker/ExecutorRunner.scala | 2 +- docs/spark-standalone.md | 5 +++-- 3 files changed, 5 insertions(+), 4 deletions(-) (limited to 'docs') diff --git a/core/src/main/scala/org/apache/spark/deploy/Client.scala b/core/src/main/scala/org/apache/spark/deploy/Client.scala index 0475bb17c0..43b9b1cff9 100644 --- a/core/src/main/scala/org/apache/spark/deploy/Client.scala +++ b/core/src/main/scala/org/apache/spark/deploy/Client.scala @@ -55,7 +55,7 @@ class DriverActor(master: String, response: Promise[(Boolean, String)]) extends /** * Executable utility for starting and terminating drivers inside of a standalone cluster. */ -object DriverClient { +object Client { def main(args: Array[String]) { val driverArgs = new ClientArguments(args) diff --git a/core/src/main/scala/org/apache/spark/deploy/worker/ExecutorRunner.scala b/core/src/main/scala/org/apache/spark/deploy/worker/ExecutorRunner.scala index a9cb998cc2..18885d7ca6 100644 --- a/core/src/main/scala/org/apache/spark/deploy/worker/ExecutorRunner.scala +++ b/core/src/main/scala/org/apache/spark/deploy/worker/ExecutorRunner.scala @@ -100,7 +100,7 @@ private[spark] class ExecutorRunner( def getCommandSeq = { val command = Command(appDesc.command.mainClass, - appDesc.command.arguments.map(substituteVariables), appDesc.command.environment) + appDesc.command.arguments.map(substituteVariables) ++ Seq(appId), appDesc.command.environment) CommandUtils.buildCommandSeq(command, memory, sparkHome.getAbsolutePath) } diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index 7da64749b7..ecd642cc60 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -151,19 +151,20 @@ You can also pass an option `-c ` to control the number of cores that You may also run your application entirely inside of the cluster by submitting your application driver using the submission client. The syntax for submitting applications is as follows: - ./spark-class org.apache.spark.deploy.client.DriverClient launch + ./spark-class org.apache.spark.deploy.Client launch [client-options] \ \ [application-options] cluster-url: The URL of the master node. - application-jar-url: Path to a bundled jar including your application and all dependencies. Currently, the URL must be visible from inside of your cluster, for instance, in an HDFS directory. + application-jar-url: Path to a bundled jar including your application and all dependencies. Currently, the URL must be globally visible inside of your cluster, for instance, an `hdfs://` path or a `file://` path that is present on all nodes. main-class: The entry point for your application. Client Options: --memory (amount of memory, in MB, allocated for your driver program) --cores (number of cores allocated for your driver program) --supervise (whether to automatically restart your driver on application or node failure) + --verbose (prints increased logging output) Keep in mind that your driver program will be executed on a remote worker machine. You can control the execution environment in the following ways: -- cgit v1.2.3 From 112c0a1776bbc866a1026a9579c6f72f293414c4 Mon Sep 17 00:00:00 2001 From: Patrick Wendell Date: Wed, 8 Jan 2014 21:16:16 -0800 Subject: Fixing config option "retained_stages" => "retainedStages". This is a very esoteric option and it's out of sync with the style we use. So it seems fitting to fix it for 0.9.0. --- core/src/main/scala/org/apache/spark/ui/jobs/JobProgressListener.scala | 2 +- docs/configuration.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'docs') diff --git a/core/src/main/scala/org/apache/spark/ui/jobs/JobProgressListener.scala b/core/src/main/scala/org/apache/spark/ui/jobs/JobProgressListener.scala index b7b87250b9..d6d9f0cedf 100644 --- a/core/src/main/scala/org/apache/spark/ui/jobs/JobProgressListener.scala +++ b/core/src/main/scala/org/apache/spark/ui/jobs/JobProgressListener.scala @@ -33,7 +33,7 @@ import org.apache.spark.scheduler._ */ private[spark] class JobProgressListener(val sc: SparkContext) extends SparkListener { // How many stages to remember - val RETAINED_STAGES = sc.conf.get("spark.ui.retained_stages", "1000").toInt + val RETAINED_STAGES = sc.conf.get("spark.ui.retainedStages", "1000").toInt val DEFAULT_POOL_NAME = "default" val stageIdToPool = new HashMap[Int, String]() diff --git a/docs/configuration.md b/docs/configuration.md index 1d6c3d1633..6717757781 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -130,7 +130,7 @@ Apart from these, the following properties are also available, and may be useful - spark.ui.retained_stages + spark.ui.retainedStages 1000 How many stages the Spark UI remembers before garbage collecting. -- cgit v1.2.3