diff options
author | Sean Owen <sowen@cloudera.com> | 2016-05-27 11:28:28 -0700 |
---|---|---|
committer | Marcelo Vanzin <vanzin@cloudera.com> | 2016-05-27 11:28:28 -0700 |
commit | 623aae5907f4ba8b7079c21328e0c0b5fef7bb00 (patch) | |
tree | 6e4c55947915ab163cb9ae88e7b5c566a5cb797f /bin/spark-class2.cmd | |
parent | ce756daa4f012ebdc5a41bf5a89ff11b6dfdab8c (diff) | |
download | spark-623aae5907f4ba8b7079c21328e0c0b5fef7bb00.tar.gz spark-623aae5907f4ba8b7079c21328e0c0b5fef7bb00.tar.bz2 spark-623aae5907f4ba8b7079c21328e0c0b5fef7bb00.zip |
[SPARK-15531][DEPLOY] spark-class tries to use too much memory when running Launcher
## What changes were proposed in this pull request?
Explicitly limit launcher JVM memory to modest 128m
## How was this patch tested?
Jenkins tests.
Author: Sean Owen <sowen@cloudera.com>
Closes #13360 from srowen/SPARK-15531.
Diffstat (limited to 'bin/spark-class2.cmd')
-rw-r--r-- | bin/spark-class2.cmd | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/bin/spark-class2.cmd b/bin/spark-class2.cmd index db680218dc..869c0b202f 100644 --- a/bin/spark-class2.cmd +++ b/bin/spark-class2.cmd @@ -55,7 +55,7 @@ if not "x%JAVA_HOME%"=="x" set RUNNER=%JAVA_HOME%\bin\java rem The launcher library prints the command to be executed in a single line suitable for being rem executed by the batch interpreter. So read all the output of the launcher into a variable. set LAUNCHER_OUTPUT=%temp%\spark-class-launcher-output-%RANDOM%.txt -"%RUNNER%" -cp "%LAUNCH_CLASSPATH%" org.apache.spark.launcher.Main %* > %LAUNCHER_OUTPUT% +"%RUNNER%" -Xmx128m -cp "%LAUNCH_CLASSPATH%" org.apache.spark.launcher.Main %* > %LAUNCHER_OUTPUT% for /f "tokens=*" %%i in (%LAUNCHER_OUTPUT%) do ( set SPARK_CMD=%%i ) |