You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-issues@hadoop.apache.org by "Varene Olivier (JIRA)" <ji...@apache.org> on 2012/12/11 14:23:21 UTC

[jira] [Commented] (MAPREDUCE-4866) ShuffleRamManager is limited to 2Gb of memory - we should increase that

    [ https://issues.apache.org/jira/browse/MAPREDUCE-4866?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13528953#comment-13528953 ] 

Varene Olivier commented on MAPREDUCE-4866:
-------------------------------------------

Sorry Chris,

but I think your patch is incomplete, only changing maxSize is not enough,

please have a look to the proposed patch MAPREDUCE-4866-INCOMPLETE.patch with all the changes proposed
the patch is incomplete because of array size limitation to around Integer.MAX_VALUE

ReduceTask.java > Class MapOutputCopier > method ShuffleInMemory
line 1507:        byte[] shuffleData = new byte[mapOutputLength];

we need to change this allocation with some kind of big array (whose size can be a long) 

regards
Olivier
                
> ShuffleRamManager is limited to 2Gb of memory - we should increase that
> -----------------------------------------------------------------------
>
>                 Key: MAPREDUCE-4866
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-4866
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>          Components: mrv1
>    Affects Versions: 0.20.2
>         Environment: linux, 64bits cpu, more than 2Gb of memory for each reducer tasks
>            Reporter: Varene Olivier
>            Priority: Minor
>              Labels: patch
>         Attachments: M4866-0.patch, MAPREDUCE-4866-INCOMPLETE.patch
>
>
> Inside the org.apache.hadoop.mapred.ReduceTask.java, the *ShuffleRamManager* is limited to allocate up to 2Gb of memory during the shuffle phase. 
> We should be able to allocate more, to take advantage of the full memory we have on servers.

--
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