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 "Zhengbo Li (Jira)" <ji...@apache.org> on 2021/05/13 17:53:00 UTC

[jira] [Created] (YARN-10768) Preserve Opportunistic Containers when "findLostContainer"

Zhengbo Li created YARN-10768:
---------------------------------

             Summary: Preserve Opportunistic Containers when "findLostContainer" 
                 Key: YARN-10768
                 URL: https://issues.apache.org/jira/browse/YARN-10768
             Project: Hadoop YARN
          Issue Type: Improvement
            Reporter: Zhengbo Li
            Assignee: Zhengbo Li


The `findLostContainers` method in `RMNodeImpl` is designed to detect any containers that are not running but failed to report its complete status to RM. However the current implementation *always considers O containers as lost*. This has result performance issues for us because we use Opportunistic Containers heavily in our feature.



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

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