[
https://issues.apache.org/jira/browse/MAPREDUCE-5481?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13773033#comment-13773033
]
Jason Lowe commented on MAPREDUCE-5481:
---------------------------------------
bq. The more weird thing is that (the reducer task will never start) happens
when we set number of reducer is 1, but if i increase the number of reducer to
2, the reducers can start to work...
That is quite odd. It definitely sounds like something changed recently that
might be bad.
Unfortunately I'm not able to reproduce the problem on my machine and don't
have access to the build machines. TestUberAM never times out for me but
rather fails when it tries to verify counter metrics at the end of the tests.
Can someone post the test output along with the AM log(s) from the minicluster
logs directory? Hopefully those can shed some light as to why the reducer
isn't getting scheduled by the AM or if it is scheduled why the container isn't
being allocated to run it.
> TestUberAM timeout
> ------------------
>
> Key: MAPREDUCE-5481
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5481
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2, test
> Affects Versions: 3.0.0
> Reporter: Jason Lowe
> Assignee: Xuan Gong
> Priority: Blocker
>
> TestUberAM has been timing out on trunk for some time now and surefire then
> fails the build. I'm not able to reproduce it locally, but the Jenkins
> builds have been seeing it fairly consistently. See
> https://builds.apache.org/job/Hadoop-Mapreduce-trunk/1529/console
--
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
