[ 
https://issues.apache.org/jira/browse/YARN-7681?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16314340#comment-16314340
 ] 

Weiwei Yang commented on YARN-7681:
-----------------------------------

Hi [~asuresh], [~leftnoteasy]

I think [~leftnoteasy] has done the check in YARN-6599, via 
{{AppPlacementAllocator#canAllocate}}, I saw this is already implemented by 
{{SingleConstraintAppPlacementAllocator#checkCardinalityAndPending}}.

[~leftnoteasy], I am not sure what you mean by 2 approaches to allocate 
SchedulingRequest. The 1st approach {{PlacementProcessor}} will eventually call

{code}
 scheduler.attemptAllocationOnNode(
                  applicationAttempt, sReqClone, node);
{code}

this calls scheduler's {{tryCommit}} method and in which 
{{AppPlacementAllocator}} makes some validation before actual allocation. So as 
long as your patch has done the check, we should be fine. Am I missing anything 
here?

Thanks

> Scheduler should double-check placement constraint before actual allocation 
> is made
> -----------------------------------------------------------------------------------
>
>                 Key: YARN-7681
>                 URL: https://issues.apache.org/jira/browse/YARN-7681
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: RM, scheduler
>            Reporter: Weiwei Yang
>            Assignee: Weiwei Yang
>
> This JIRA is created based on the discussions under YARN-7612, see comments 
> after [this 
> comment|https://issues.apache.org/jira/browse/YARN-7612?focusedCommentId=16303051&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16303051].
>  AllocationTagsManager maintains tags info that helps to make placement 
> decisions at placement phase, however tags are changing along with 
> container's lifecycle, so it is possible that the placement violates the 
> constraints at the scheduling phase. Propose to add an extra check in the 
> scheduler to make sure constraints are still satisfied during the actual 
> allocation.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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