You are viewing a plain text version of this content. The canonical link for it is here.

Posted to yarn-issues@hadoop.apache.org by "stefanlee (JIRA)" <ji...@apache.org> on 2018/08/11 03:50:00 UTC

[jira] [Created] (YARN-8651) We must increase min Resource in FairScheduler after increase number of NM

stefanlee created YARN-8651:
-------------------------------

             Summary: We must increase min Resource in FairScheduler after increase number of NM
                 Key: YARN-8651
                 URL: https://issues.apache.org/jira/browse/YARN-8651
             Project: Hadoop YARN
          Issue Type: Bug
          Components: fairscheduler
    Affects Versions: 2.4.0
            Reporter: stefanlee


Nowadays，our cluster has a strange phenomena，before we increase the scale of NodeManager, the resource utilization could be 100%, but we found the resource utilization does not promote as the cluster expansion and many queue's used resource stay at min resource  although they have many demand request.

Then we increase their min resource dynamically，the resource utilization of these queues  goes  up and the resources of the entire cluster are also used after that.

So i doubte if the bug in *FairSharePolicy#compare.*



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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