hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sandy Ryza (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-5236) JobConf should not support disabling memory limits
Date Sun, 12 May 2013 03:39:16 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-5236?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13655455#comment-13655455

Sandy Ryza commented on MAPREDUCE-5236:

Sorry if my title/description were not very clear -  my understanding of the MR1 behavior
around this isn't super strong.  Currently, JobConf#getMemoryForMapTask returns a default
of DISABLED_MEMORY_LIMIT, which is -1.  I am basically proposing removing all references to
that constant the code, as
1) after MAPREDUCE-5130, mapred-default.xml will set mapreduce.*.memory.mb to 1024, so getMemoryFor*Task
should be consistent with this.
2) correct me if I'm wrong, but DISABLED_MEMORY_LIMIT doesn't make sense in the context of
MR2, as there is no disabled memory limit mode.
> JobConf should not support disabling memory limits
> --------------------------------------------------
>                 Key: MAPREDUCE-5236
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5236
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>    Affects Versions: 2.0.4-alpha
>            Reporter: Sandy Ryza
>            Assignee: Sandy Ryza
> In MR1, a special value of -1 could be given for mapreduce.job.map|reduce.memory.mb to
disable memory limits.  In MR2, this makes no sense, as with slots gone, this value is used
for requesting resources and scheduling.

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

View raw message