[
https://issues.apache.org/jira/browse/MAPREDUCE-2756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13085872#comment-13085872
]
Robert Joseph Evans commented on MAPREDUCE-2756:
------------------------------------------------
I did not provide a patch for trunk because this mostly impacts MR-279. It is
rather difficult to cause an Error instead of an Exception to happen except
with the change in binary compatibility caused by MR-279. It think it would be
simpler to just wait for MR-279 to merge to trunk before applying the patch.
As far as manual testing. We ran about 100 different pig scripts with custom
Storage functions in them. Before the change one of them would consistently
deadlock, and all the others would fail, but with no indication of why. After
the change they all still fail. No deadlocks, and there is a log message
indicating why it failed.
> JobControl can drop jobs if an error occurs
> -------------------------------------------
>
> Key: MAPREDUCE-2756
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-2756
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: client, mrv2
> Affects Versions: 0.23.0
> Reporter: Robert Joseph Evans
> Assignee: Robert Joseph Evans
> Priority: Minor
> Fix For: 0.23.0
>
> Attachments: MR-2756-V1.patch
>
>
> If you run a pig job with UDFs that has not been recompiled for MRV2. There
> are situations where pig will fail with an error message stating that Hadoop
> failed and did not give a reason. There is even the possibility of deadlock
> if an Error is thrown and the JobControl thread dies.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
