hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Owen O'Malley (JIRA)" <j...@apache.org>
Subject [jira] Resolved: (HADOOP-945) Heavy map tasks should be assigned first
Date Fri, 20 Mar 2009 15:36:51 GMT

     [ https://issues.apache.org/jira/browse/HADOOP-945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Owen O'Malley resolved HADOOP-945.
----------------------------------

    Resolution: Cannot Reproduce

It already assigns them based on the size of the input splits.

> Heavy map tasks should be assigned first
> ----------------------------------------
>
>                 Key: HADOOP-945
>                 URL: https://issues.apache.org/jira/browse/HADOOP-945
>             Project: Hadoop Core
>          Issue Type: Improvement
>          Components: mapred
>            Reporter: Runping Qi
>
> In many cases I noticed that a few  heavy map tasks finished really late since they were
started late too. When the jobtracker chooses map tasks for task trackers, it should takes
the input data size of the tasks into account, and try to assign the ones with large input
first.

-- 
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