You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-issues@hadoop.apache.org by "Amol Kekre (Created) (JIRA)" <ji...@apache.org> on 2012/01/13 19:58:40 UTC
[jira] [Created] (MAPREDUCE-3667) Gridmix jobs are failing with OOM
in reduce shuffle phase.
Gridmix jobs are failing with OOM in reduce shuffle phase.
----------------------------------------------------------
Key: MAPREDUCE-3667
URL: https://issues.apache.org/jira/browse/MAPREDUCE-3667
Project: Hadoop Map/Reduce
Issue Type: Bug
Components: mrv2
Affects Versions: 0.23.0
Reporter: Amol Kekre
Priority: Blocker
Fix For: 0.23.1
Roll up bug for gridmix3 benchmark
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3667) Gridmix jobs are failing with
OOM in reduce shuffle phase.
Posted by "Vinod Kumar Vavilapalli (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/MAPREDUCE-3667?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13198079#comment-13198079 ]
Vinod Kumar Vavilapalli commented on MAPREDUCE-3667:
----------------------------------------------------
bq. Hadoop 0.20.204 reducers does not goes OOM, I have seen this in Hadoop.Next only.
That is right, Karam. We would ideally want YARN+MR to be able to run with the same configuration. Unfortunately mrv2 reduce runtime has changed a lot, and it performs better in so many other situations but worse in this particular case. The implementation has changed in such a way that supporting 0.7 is not possible with the current set of jobs in gridmix without running into OOM.
I agree that it is a regression but for now this should be fine. We'd ideally want to fix this, will open a separate ticket, but the timelines for that fix are uncertain at this point.
> Gridmix jobs are failing with OOM in reduce shuffle phase.
> ----------------------------------------------------------
>
> Key: MAPREDUCE-3667
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3667
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 0.23.0
> Reporter: Amol Kekre
> Priority: Blocker
> Fix For: 0.23.1
>
>
> Roll up bug for gridmix3 benchmark
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (MAPREDUCE-3667) Gridmix jobs are failing with
OOM in reduce shuffle phase.
Posted by "Arun C Murthy (Resolved) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/MAPREDUCE-3667?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy resolved MAPREDUCE-3667.
--------------------------------------
Resolution: Not A Problem
This was caused by over-aggressive configs in the reduce.
> Gridmix jobs are failing with OOM in reduce shuffle phase.
> ----------------------------------------------------------
>
> Key: MAPREDUCE-3667
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3667
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 0.23.0
> Reporter: Amol Kekre
> Priority: Blocker
> Fix For: 0.23.1
>
>
> Roll up bug for gridmix3 benchmark
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3667) Gridmix jobs are failing with
OOM in reduce shuffle phase.
Posted by "Karam Singh (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/MAPREDUCE-3667?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13194597#comment-13194597 ]
Karam Singh commented on MAPREDUCE-3667:
----------------------------------------
Hi Arun
I have similar setting for 0.23 and Hadoop-0.20.204 with GridMix and same trace
mapred.job.shuffle.input.buffer.percent=0.7 is Hadoop 0.20.204
And
mapreduce.reduce.shuffle.input.buffer.percent=0.7
Hadoop 0.20.204 reducers does not goes OOM, I have seen this in Hadoop.Next only.
And even in Hadoop 0.23 the behavior is not consistent, random number of job used fail due to this.
W.r.t to Hadoop-0.20.204 is not a regression
> Gridmix jobs are failing with OOM in reduce shuffle phase.
> ----------------------------------------------------------
>
> Key: MAPREDUCE-3667
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3667
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 0.23.0
> Reporter: Amol Kekre
> Priority: Blocker
> Fix For: 0.23.1
>
>
> Roll up bug for gridmix3 benchmark
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira