[jira] [Commented] (MAPREDUCE-5892) Derive MR-AM container size and Xmx based on the job size: number of splits and reduces

["BELUGA BEHR (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/MAPREDUCE-5892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16758574#comment-16758574 ] 

BELUGA BEHR commented on MAPREDUCE-5892:
----------------------------------------

I would like to propose that we tackle this from a different angle: [YARN-9260]

Instead of trying to come up with a clever heuristic based on splits and reduces and such, simply re-try the AM with an arbitrarily larger container.  Do this 2 or 3 times with the expectation that the job finally passes.  Users can then change the AM memory configurations higher to avoid the re-tries in the futures instead of having to experience a failure and then immediately, and manually, set a higher number just to get it to run.

 

> Derive MR-AM container size and Xmx based on the job size: number of splits and reduces
> ---------------------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-5892
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5892
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>          Components: mr-am
>            Reporter: Gera Shegalov
>            Priority: Major
>
> https://issues.apache.org/jira/browse/MAPREDUCE-207?focusedCommentId=13997964&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13997964



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

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