[ 
https://issues.apache.org/jira/browse/YARN-2113?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Eric Payne updated YARN-2113:
-----------------------------
    Attachment: YARN-2113.apply.onto.0012.ericp.patch

Thanks [~sunilg] for the all of your work and for the updated patch.

I think we are making the USERLIMIT_FIRST code to complicated. [The original 
problem I was hoping to 
solve|https://issues.apache.org/jira/browse/YARN-2113?focusedCommentId=15951222&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15951222]
 was that no preemption would occur for a user that is #users > (100/MULP). So, 
if you have a MULP of 50%, 2 active users can be in the queue. So, if a third 
user comes in AND userlimit_first is true, no preemption should occur for the 
third user.

I'm still thinking this through, but I am attaching a patch that I think 
addresses my concern. This patch would be applied on top of 
YARN-2113.0012.patch.

> Add cross-user preemption within CapacityScheduler's leaf-queue
> ---------------------------------------------------------------
>
>                 Key: YARN-2113
>                 URL: https://issues.apache.org/jira/browse/YARN-2113
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: scheduler
>            Reporter: Vinod Kumar Vavilapalli
>            Assignee: Sunil G
>         Attachments: IntraQueue Preemption-Impact Analysis.pdf, 
> TestNoIntraQueuePreemptionIfBelowUserLimitAndDifferentPrioritiesWithExtraUsers.txt,
>  YARN-2113.0001.patch, YARN-2113.0002.patch, YARN-2113.0003.patch, 
> YARN-2113.0004.patch, YARN-2113.0005.patch, YARN-2113.0006.patch, 
> YARN-2113.0007.patch, YARN-2113.0008.patch, YARN-2113.0009.patch, 
> YARN-2113.0010.patch, YARN-2113.0011.patch, YARN-2113.0012.patch, 
> YARN-2113.apply.onto.0012.ericp.patch, YARN-2113.v0.patch
>
>
> Preemption today only works across queues and moves around resources across 
> queues per demand and usage. We should also have user-level preemption within 
> a queue, to balance capacity across users in a predictable manner.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org

Reply via email to