[
https://issues.apache.org/jira/browse/MAPREDUCE-4228?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13269673#comment-13269673
]
Robert Joseph Evans commented on MAPREDUCE-4228:
------------------------------------------------
This is mostly about the utilization of the cluster. In many cases you have a
map/reduce job where the mappers take a very long time. If you don't wait
until all of the mappers have finished there are many many reducers taking up
resources doing nothing waiting for the mappers to finish. In some cases we
have seen this last from several hours up to a day or two.
> mapreduce.job.reduce.slowstart.completedmaps is not working properly to delay
> the scheduling of the reduce tasks
> ----------------------------------------------------------------------------------------------------------------
>
> Key: MAPREDUCE-4228
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-4228
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: applicationmaster, mrv2
> Affects Versions: 0.23.1
> Reporter: Jason Lowe
> Assignee: Jason Lowe
>
> If no more map tasks need to be scheduled but not all have completed, the
> ApplicationMaster will start scheduling reducers even if the number of
> completed maps has not met the mapreduce.job.reduce.slowstart.completedmaps
> threshold. For example, if the property is set to 1.0 all maps should
> complete before any reducers are scheduled. However the reducers are
> scheduled as soon as the last map task is assigned to a container. For a job
> with very long-running maps, a cluster with enough capacity to launch all map
> tasks could cause reducers to launch prematurely and waste cluster resources.
> Thanks to Phil Su for discovering this issue.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
