hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel Templeton (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-7819) Allow PlacementProcessor to be used with the FairScheduler
Date Thu, 01 Feb 2018 00:49:03 GMT

    [ https://issues.apache.org/jira/browse/YARN-7819?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16347834#comment-16347834
] 

Daniel Templeton commented on YARN-7819:
----------------------------------------

Comments:
* The write lock should be locked outside the {{try}} in {{allocateOnNode()}}
* Can we please move {{setQueueName()}} up to {{RMContainer}}?  Every single use of it involves
a cast from {{RMContainer}} to {{RMContainerImpl}}, and there are no other subclasses of {{RMContainer}}.
* Will calling the assignment node-local in the metrics update confuse the metrics?  What
if it's not actually node local?
* The log messages on FSLeafQueue:360,364 is pretty cryptic.  Either provide a clear explanation
of what happened and what should be done about it or consider making then info or debug messages
(or both!)
* The log message on FairScheduler:1879 definitely shouldn't be a warn.  See story as my previous
point.
* Seems like there should be a cleaner way to do this:{code}        Resource resource =
            schedulingRequest.getResourceSizing().getResources();
        schedulingRequest.getResourceSizing().setResources(
            getNormalizedResource(resource));{code} Like maybe add a normalize method to the
resource sizing and/or move this operation into the {{createRMContainer()}} method.
* {{Resources.greaterThan(none)}} probably isn't what you want.  You probably want {{!Resources.isNone()}}.


On [~haibo.chen]'s synchronization concerns, the only thing that concerns me is the {{FSAppAttempt}}
state.  I'd have to dig to see if it's really an issue, though.

> Allow PlacementProcessor to be used with the FairScheduler
> ----------------------------------------------------------
>
>                 Key: YARN-7819
>                 URL: https://issues.apache.org/jira/browse/YARN-7819
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Arun Suresh
>            Assignee: Arun Suresh
>            Priority: Major
>         Attachments: YARN-7819-YARN-6592.001.patch, YARN-7819-YARN-7812.001.patch
>
>
> The FairScheduler needs to implement the {{ResourceScheduler#attemptAllocationOnNode}}
function for the processor to support the FairScheduler.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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


Mime
View raw message