[ https://issues.apache.org/jira/browse/SPARK-11403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14980456#comment-14980456 ]
Apache Spark commented on SPARK-11403: -------------------------------------- User 'darabos' has created a pull request for this issue: https://github.com/apache/spark/pull/9355 > Log something when dying from OnOutOfMemoryError > ------------------------------------------------ > > Key: SPARK-11403 > URL: https://issues.apache.org/jira/browse/SPARK-11403 > Project: Spark > Issue Type: Improvement > Components: YARN > Affects Versions: 1.5.1 > Reporter: Daniel Darabos > Priority: Trivial > > Executors on YARN run with the {{-XX:OnOutOfMemoryError='kill %p'}} flag. The > motivation I think is to avoid getting to an unpredictable state where some > threads may have been lost. > The problem is that when this happens nothing is logged. The executor does > not log anything, it just exits with exit code 143. This is logged in the > NodeManager log. > I'd like to add a tiny log message to make debugging such issues a bit easier. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org