tez-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Bikas Saha (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (TEZ-342) TaskScheduler should cycle through nodes while allocating containers if assigned multiple containers per node
Date Tue, 06 Aug 2013 05:45:49 GMT

    [ https://issues.apache.org/jira/browse/TEZ-342?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13730389#comment-13730389
] 

Bikas Saha commented on TEZ-342:
--------------------------------

When trying to assign a container, the task scheduler first looks for all node local requests
for that container. Only when there are 0 node local requests does it move to rack or ANY.
So even if there are multiple containers on a node there does not seem any benefit of not
trying to allocate them. Specially after TEZ-343 this jira does not seem actionable.
                
> TaskScheduler should cycle through nodes while allocating containers if assigned multiple
containers per node
> -------------------------------------------------------------------------------------------------------------
>
>                 Key: TEZ-342
>                 URL: https://issues.apache.org/jira/browse/TEZ-342
>             Project: Apache Tez
>          Issue Type: Improvement
>            Reporter: Siddharth Seth
>            Assignee: Siddharth Seth
>              Labels: TEZ-0.2.0
>
> Should help with task locality. (To some extent, similar to MAPREDUCE-5352)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message