[ 
https://issues.apache.org/jira/browse/YARN-10559?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17269420#comment-17269420
 ] 

Eric Payne commented on YARN-10559:
-----------------------------------

Thanks [~ananyo_rao] for bringing up this issue and for providing a patch. 
Sorry for my delay in responding.

I think the overall requirements seem reasonable. Preemption is supposed to 
mimic what the capacity scheduler would do when a queue has FairOrderingPolicy 
set if all jobs were launched simultaneously. So preempting from a user's first 
app to give to the same user's second app makes sense.

The only caveat I would add is that (as is the case with all preempted 
containers) just because the preemption monitor decides to preempt a container 
because of a particular request, that doesn't mean that the capacity scheduler 
will then assign the container as expected. The states of the queue and cluster 
are constantly in flux and so a container preempted for one app could easily go 
to a different app in the same queue or a different queue.

I will try to look at the changes next week.

> Fair sharing intra-queue preemption support in Capacity Scheduler
> -----------------------------------------------------------------
>
>                 Key: YARN-10559
>                 URL: https://issues.apache.org/jira/browse/YARN-10559
>             Project: Hadoop YARN
>          Issue Type: Improvement
>          Components: capacityscheduler
>    Affects Versions: 3.1.4
>            Reporter: VADAGA ANANYO RAO
>            Assignee: VADAGA ANANYO RAO
>            Priority: Major
>         Attachments: FairOP_preemption-design_doc_v1.pdf, 
> FairOP_preemption-design_doc_v2.pdf, YARN-10559.0001.patch, 
> YARN-10559.0002.patch, YARN-10559.0003.patch, YARN-10559.0004.patch, 
> YARN-10559.0005.patch, YARN-10559.0006.patch, YARN-10559.0007.patch, 
> YARN-10559.0008.patch
>
>   Original Estimate: 168h
>  Remaining Estimate: 168h
>
> Usecase:
> Due to the way Capacity Scheduler preemption works, If a single user submits 
> a large application to a queue (using 100% of resources), that job will not 
> be preempted by future applications from the same user within the same queue. 
> This implies that the later applications will be forced to wait for 
> completion of the long running application. This prevents multiple long 
> running, large, applications from running concurrently.
> Support fair sharing among apps while preempting applications from same queue.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
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