hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sunil G (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (YARN-5889) Improve user-limit calculation in capacity scheduler
Date Sat, 26 Nov 2016 16:19:58 GMT

     [ https://issues.apache.org/jira/browse/YARN-5889?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Sunil G updated YARN-5889:
    Attachment: YARN-5889.v2.patch

Thanks [~eepayne]. Addressed most of the comments.

bq.It seems like this should be longer than 1 ms. Isn't the default scheduling interval 5
seconds? That may be too long, but I think it should be at least a second.
I could slightly explain here. I think 1msec is also smaller. It could be possible that containers
are released and created very fast in a big cluster. Each such change in resource will have
an impact on user-limit calculation. Earlier it was not affecting this problem since we were
calculating user-limit in every allocation cycle. So we could try with this small timeout,
and can do some SLS tests to see the trend. I will post some results very soon.

> Improve user-limit calculation in capacity scheduler
> ----------------------------------------------------
>                 Key: YARN-5889
>                 URL: https://issues.apache.org/jira/browse/YARN-5889
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: capacity scheduler
>            Reporter: Sunil G
>            Assignee: Sunil G
>         Attachments: YARN-5889.v0.patch, YARN-5889.v1.patch, YARN-5889.v2.patch
> Currently user-limit is computed during every heartbeat allocation cycle with a write
lock. To improve performance, this tickets is focussing on moving user-limit calculation out
of heartbeat allocation flow.

This message was sent by Atlassian JIRA

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

View raw message