[
https://issues.apache.org/jira/browse/MAPREDUCE-5115?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13616682#comment-13616682
]
Bikas Saha commented on MAPREDUCE-5115:
---------------------------------------
Simple fix could be to check against a counter of reduces instead of
reduces.isEmpty() in the ScheduledRequests.assign().
> MR app master may try to assign a reduce priority container to a map
> --------------------------------------------------------------------
>
> Key: MAPREDUCE-5115
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5115
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Bikas Saha
>
> ScheduledRequests.assign() checks reduces.isEmpty() and releases REDUCE
> priority containers. But it could have received more REDUCE priority
> containers than reduces.size() in which case it will not release excess
> REDUCE priority containers. Later on in ScheduledRequests.assignToReduce() it
> will not be able to assign them to reduces. These containers will fall
> through to ScheduledRequests.assignMapsWithLocality() where they will get
> assigned to maps or crash depending on the behavior of Java assert.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
