[
https://issues.apache.org/jira/browse/SPARK-27228?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16798841#comment-16798841
]
Lukas Waldmann commented on SPARK-27228:
----------------------------------------
Executors management seems to behave strangely.
After calling spark.stop()
See this:
{quote}19/03/21 09:51:39 INFO YarnSchedulerBackend$YarnDriverEndpoint:
Disabling executor 332.
19/03/21 09:51:39 INFO DAGScheduler: Executor lost: 332 (epoch 446)
19/03/21 09:51:39 INFO BlockManagerMasterEndpoint: Trying to remove executor
332 from BlockManagerMaster.
19/03/21 09:51:39 INFO BlockManagerMasterEndpoint: Removing block manager
BlockManagerId(332, data-10.bdp.gin.merck.com, 38713, None)
19/03/21 09:51:39 INFO BlockManagerMaster: Removed 332 successfully in
removeExecutor
{quote}
and few minutes later:
{quote}19/03/21 09:54:26 WARN HeartbeatReceiver: Removing executor 332 with no
recent heartbeats: 173942 ms exceeds timeout 120000 ms
19/03/21 09:54:26 ERROR YarnClusterScheduler: Lost an executor 332 (already
removed): Executor heartbeat timed out after 173942 ms
19/03/21 09:54:26 INFO YarnClusterSchedulerBackend: Requesting to kill
executor(s) 332
19/03/21 09:54:26 WARN YarnClusterSchedulerBackend: Executor to kill 332 does
not exist!
19/03/21 09:54:26 INFO YarnClusterSchedulerBackend: Actual list of executor(s)
to be killed is{quote}
> Spark long delay on close, possible problem with killing executors
> ------------------------------------------------------------------
>
> Key: SPARK-27228
> URL: https://issues.apache.org/jira/browse/SPARK-27228
> Project: Spark
> Issue Type: Bug
> Components: Block Manager
> Affects Versions: 2.3.0
> Reporter: Lukas Waldmann
> Priority: Major
> Attachments: log.html
>
>
> When using dynamic allocations after all jobs finishes spark delays for
> several minutes before finally finishes. Log suggest that executors are not
> cleared up properly.
> See the attachment for log
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
