You are viewing a plain text version of this content. The canonical link for it is here.

Posted to mapreduce-dev@hadoop.apache.org by "Amar Kamat (JIRA)" <ji...@apache.org> on 2011/05/30 11:23:47 UTC

[jira] [Created] (MAPREDUCE-2545) [Rumen] Rumen incorrectly emits memory usages of map/reduce tasks

[Rumen] Rumen incorrectly emits memory usages of map/reduce tasks
-----------------------------------------------------------------

                 Key: MAPREDUCE-2545
                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-2545
             Project: Hadoop Map/Reduce
          Issue Type: Bug
          Components: tools/rumen
            Reporter: Amar Kamat
             Fix For: 0.23.0


Rumen emits job and task (map/reduce) level heap settings (i.e values of mapred.child.java.opts, mapreduce.map.java.opts and mapreduce.reduce.java.opts). It also emits cluster and task (map/reduce) level memory configurations (e.g mapreduce.cluster.mapmemory.mb, mapreduce.cluster.reducememory.mb, mapreduce.map.memory.mb and mapreduce.reduce.memory.mb). Currently, it incorrectly sets map/reduce task level memory usage to the max heap values obtained from mapreduce.map.java.opts and mapreduce.reduce.java.opts. Note that HadoopLogsAnalyzer correctly sets it.

Also the cluster level task level (map/reduce) memory configurations are missing.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira