bsidhom opened a new pull request #24462: [SPARK-26268][CORE] Do not resubmit tasks when executors are lost URL: https://github.com/apache/spark/pull/24462 The DAGScheduler assumes that shuffle data is lost if an executor is lost and the Spark external shuffle service is not enabled. However, external shuffle managers other than the default external shuffle service can store shuffle data outside of the Spark cluster itself. In this case, completed map taks are not rerun even if the corresponding executors are lost. The new `spark.shuffle.external.enabled` property allows this new external shuffle behavior to be toggled (independently off the built-in Spark external shuffle service).
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
