[
https://issues.apache.org/jira/browse/MAPREDUCE-2961?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13113486#comment-13113486
]
Vinod Kumar Vavilapalli commented on MAPREDUCE-2961:
----------------------------------------------------
Oh, BTW, I took Karam's help to see this patch's effect. Before this patch, we
had 200 core-pool size for a 100,000 map-only jobs on ~350 nodes. With this
patch, which automatically takes care of the thread-pool size(and sets to ~350
for this job in particular), we saw ~30% improvement on wall-clock time.
> Increase the default threadpool size for container launching in the
> application master.
> ---------------------------------------------------------------------------------------
>
> Key: MAPREDUCE-2961
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-2961
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: mrv2
> Affects Versions: 0.23.0
> Reporter: Mahadev konar
> Assignee: Vinod Kumar Vavilapalli
> Priority: Blocker
> Fix For: 0.23.0
>
> Attachments: MAPREDUCE-2961-20110923.1.txt,
> MAPREDUCE-2961-20110923.txt, MAPREDUCE-2961.patch, MAPREDUCE-2961.patch
>
>
> Currently the default threadpool size is 10 for launching containers in
> ContainerLauncherImpl. We should increase that to 100 for a reasonable
> default, so that container launching is not backed up by a small thread pool
> size.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
