hadoop-mapreduce-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Arun C Murthy (JIRA)" <j...@apache.org>
Subject [jira] Created: (MAPREDUCE-1046) Remove mapred.{map|reduce}.child.ulimit and use mapred.job.{map|reduce}.memory.mb as the task's ulimit value
Date Mon, 28 Sep 2009 02:49:16 GMT
Remove mapred.{map|reduce}.child.ulimit and use mapred.job.{map|reduce}.memory.mb as the task's
ulimit value
------------------------------------------------------------------------------------------------------------

                 Key: MAPREDUCE-1046
                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1046
             Project: Hadoop Map/Reduce
          Issue Type: Improvement
          Components: client, task
    Affects Versions: 0.21.0
            Reporter: Arun C Murthy


Currently mapred.{map|reduce}.child.ulimit is the ulimit value set for the child map/reduce
task. We also have mapred.job.{map|reduce}.memory.mb in the capacity-scheduler which specifies
the 'slot-size' in terms of memory for the child task. I propose we deprecate mapred.{map|reduce}.child.ulimit
and use the value of mapred.job.{map|reduce}.memory.mb - this will help reduce proliferation
of config knobs.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message