hadoop-mapreduce-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Allen Wittenauer (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (MAPREDUCE-289) JobTracker should not expand jobs if its running low on memory
Date Mon, 21 Jul 2014 17:58:42 GMT

     [ https://issues.apache.org/jira/browse/MAPREDUCE-289?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Allen Wittenauer resolved MAPREDUCE-289.

    Resolution: Fixed

Job submission was changed significantly in 0.20.20x timeline that does the input split calculation
on the client side.... which I'm assuming is the expansion being referred to here.  Closing
this as fixed.

> JobTracker should not expand jobs if its running low on memory
> --------------------------------------------------------------
>                 Key: MAPREDUCE-289
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-289
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>            Reporter: Amar Kamat
> When the JobTracker detects that its running low on memory it should not expand new jobs
if the job has the potential to bring it down. Consider and example where the JobTracker runs
on 60% of the max memory and a new job is submitted which can take upto 40% of the max memory.
 Ideally the JobTracker should _queue_ the job for expansion and expand when sufficient memory
is available.

This message was sent by Atlassian JIRA

View raw message