[
https://issues.apache.org/jira/browse/SPARK-19757?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15886838#comment-15886838
]
Jimmy Xiang commented on SPARK-19757:
-------------------------------------
In class CoarseGrainedSchedulerBackend, killExecutors makes sure an executor is
not busy before killing it. However, makeOffers could schedule a task to the
executor right after killExecutors thinks it is not busy.
I think we should improve makeOffers to add some synchronization for checking
if an executor is alive and schedule task to it.
> Executor with task scheduled could be killed due to idleness
> ------------------------------------------------------------
>
> Key: SPARK-19757
> URL: https://issues.apache.org/jira/browse/SPARK-19757
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
> Affects Versions: 1.6.0
> Reporter: Jimmy Xiang
> Priority: Minor
>
> With dynamic executor allocation enabled on yarn mode, after one job is
> finished for a while, submit another job, then there is race between killing
> idle executors and scheduling new task on these executors. Sometimes, some
> executor is killed right after a task is scheduled on it.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]