[
https://issues.apache.org/jira/browse/YARN-6158?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15860301#comment-15860301
]
Miklos Szegedi commented on YARN-6158:
--------------------------------------
I just noticed, think it is the same issue as YARN-3933. This patch copies,
what CapacityScheduler does, so it helps any refactoring in the future to merge
some codebase.
> FairScheduler: app usage can go to negative
> -------------------------------------------
>
> Key: YARN-6158
> URL: https://issues.apache.org/jira/browse/YARN-6158
> Project: Hadoop YARN
> Issue Type: Bug
> Components: fairscheduler, resourcemanager
> Affects Versions: 3.0.0-alpha2
> Reporter: Miklos Szegedi
> Assignee: Miklos Szegedi
> Attachments: YARN-6158.000.patch
>
>
> FiCaSchedulerApp.containerCompleted checks, if the container being completed
> is in the active list:
> {code}
> // Remove from the list of containers
> if (null == liveContainers.remove(containerId)) {
> return false;
> }
> {code}
> Fair scheduler should do the same, otherwise multiple different container
> close events leave the application with negative resource usage in
> {{queue.getMetrics().releaseResources}} and {{attemptResourceUsage.decUsed}}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
