[
https://issues.apache.org/jira/browse/MAPREDUCE-5848?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13975600#comment-13975600
]
Jason Lowe commented on MAPREDUCE-5848:
---------------------------------------
Yes, FileSystem is doing this. See the ClientFinalizer in FileSystem. One
example where we've seen this race occur is when tasks start running outside
their memory boundaries and the NM kills them. Sometimes the task kills
cleanly and the AM gets the useful container status added by the NM explaining
why the container was killed. Other times the task ends up failing first due
to actions caused by the SIGTERM processing and the useful container status
arrives too late.
> MapReduce counts forcibly preempted containers as FAILED
> --------------------------------------------------------
>
> Key: MAPREDUCE-5848
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5848
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Affects Versions: 2.1.0-beta
> Reporter: Carlo Curino
> Assignee: Subramaniam Krishnan
> Attachments: YARN-1958.patch
>
>
> The MapReduce AM is considering a forcibly preempted container as FAILED,
> while I think it should be considered as KILLED (i.e., not count against the
> maximum number of failures).
--
This message was sent by Atlassian JIRA
(v6.2#6252)
