[
https://issues.apache.org/jira/browse/MAPREDUCE-4938?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13551561#comment-13551561
]
Jason Lowe commented on MAPREDUCE-4938:
---------------------------------------
This is complicated by the fact that the RM returned success for the
application submission even though the submission was really invalid. It just
took it a longer time to figure it out when the problem occurs. Maybe this
should be fixed on the YARN side instead?
> Job submission to unknown queue can leave staging directory behind
> ------------------------------------------------------------------
>
> Key: MAPREDUCE-4938
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-4938
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: client
> Affects Versions: 2.0.3-alpha, 0.23.5
> Reporter: Jason Lowe
>
> There is a race where submitting a job to an unknown queue can appear to
> succeed to the client and then subsequently fail later. Since there was no
> AM ever launched, there was nothing left to cleanup the staging directory.
> At that point the client is the only thing that can cleanup the staging
> directory.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
