[
https://issues.apache.org/jira/browse/HUDI-2073?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17369525#comment-17369525
]
Yue Zhang commented on HUDI-2073:
---------------------------------
I just make a pr to fix this bug. linked here
https://github.com/apache/hudi/pull/3157
> The sparkJob of hoodieClusteringJob running through sparkSubmit will not quit
> even it is finished or failed.
> ------------------------------------------------------------------------------------------------------------
>
> Key: HUDI-2073
> URL: https://issues.apache.org/jira/browse/HUDI-2073
> Project: Apache Hudi
> Issue Type: Bug
> Reporter: Yue Zhang
> Priority: Major
> Labels: pull-request-available
>
> Users can launch HoodieClusteringJob through sparkSubmit to
> # Scheduling clustering
> # Execute clustering
> But these spark jobs will never finished and SparkSubmit never quit even jobs
> are finished or failed.
> This is because clustering job will init a SparkRDDWriteClient to doSchedule
> or do cluster But did not close this client after that. It will cause that
> `jsc.stop();` can't kill this sparkJob and hang forever.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
