[ 
https://issues.apache.org/jira/browse/YARN-10010?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16986740#comment-16986740
 ] 

Wilfred Spiegelenburg commented on YARN-10010:
----------------------------------------------

The thread pool size is configurable via 
{{yarn.nodemanager.logaggregation.threadpool-size-max}} if you need more 
threads than you can set it higher.
I would recommend that you set it to a number almost as high as the number of 
simultaneous application you expect in the cluster. I cannot give you an exact 
number but 100 threads especially in larger clusters might not be enough.

Can we also close this as a dupe of YARN-8364 ?

> NM upload log cost too much time
> --------------------------------
>
>                 Key: YARN-10010
>                 URL: https://issues.apache.org/jira/browse/YARN-10010
>             Project: Hadoop YARN
>          Issue Type: Improvement
>            Reporter: zhoukang
>            Assignee: zhoukang
>            Priority: Major
>         Attachments: notfound.png
>
>
> Since thread pool size of log service is 100.
> Some times the log uploading service will delay for some apps.like below
>  !notfound.png! 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org

Reply via email to