[
https://issues.apache.org/jira/browse/YARN-8537?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16544113#comment-16544113
]
Yeliang Cang edited comment on YARN-8537 at 7/14/18 8:26 AM:
-------------------------------------------------------------
Sorry, it seems that branch-3.1.0 does not have the problem any more.
I discovered this problem in branch-2.7.3 after apply patch YARN-4090,
YARN-7382.
Restart ResourceManager when there are jobs still running, and the first
container that recovered for each job is not the am. And therefore queue's
AMResourceUsage is not correct.
But I can not find this problem in branch-3.1.0, will close this as invalid.
Sorry to disturb...
was (Author: cyl):
Sorry, it seems that branch-3.1.0 does not have the problem any more.
I discovered this problem in branch-2.7.3 after apply patch YARN-4090,
YARN-7382.
Restart ResourceManager when there are jobs still running, and the first
container that recovered for each job is not the am. And therefore queue's
AMResourceUsage is not correct.
But I can not find this problem in branch-3.1.0, will close this as invalid.
Sorry to disturb...
> Misleading note in FSAppAttempt class
> -------------------------------------
>
> Key: YARN-8537
> URL: https://issues.apache.org/jira/browse/YARN-8537
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 3.1.0
> Reporter: Yeliang Cang
> Assignee: Yeliang Cang
> Priority: Minor
>
> In o.a.h.y.s.r.scheduler.fair.FSAppAttempt class, the note below:
> {code}
> // If not running unmanaged, the first container we recover is always
> // the AM. Set the amResource for this app and update the leaf queue's AM
> // usage
> {code}
> In fact, the first container that recovered on my machine mostly is not the
> AM!
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
