hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Suma Shivaprasad (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-7643) Handle recovery of applications on auto-created leaf queues
Date Wed, 13 Dec 2017 02:36:01 GMT

    [ https://issues.apache.org/jira/browse/YARN-7643?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16288627#comment-16288627
] 

Suma Shivaprasad commented on YARN-7643:
----------------------------------------

{quote}
1) Why AbstractCSQueue changes required? (And it doesn't look correct for queue refresh cases.)
{quote}
Earlier to this fix , the global scheduler configuration for enabling user metrics "yarn.scheduler.capacity.user-metrics.enable"
was taken from queue specific configuration instead of scheduler CapacitySchedulerContext.configuration
which was incorrect. 

TestWorkPreservingRMRestart.checkCSQueue has validations to check userMetrics and this step
fails without this change
 // ************ check user metrics ***********
    QueueMetrics userMetrics =
        queueMetrics.getUserMetrics(app.getUser());
    assertMetrics(userMetrics, 1, 0, 1, 0, 2, availableResources.getMemorySize(),
        availableResources.getVirtualCores(), usedResource.getMemorySize(),
        usedResource.getVirtualCores());
{quote} 2) Several places need to reformat in CapacityScheduler#addApplicationOnRecovery,
like too long line, blank lines, etc. Could you take care of them? {quote}
Will fix



> Handle recovery of applications on auto-created leaf queues
> -----------------------------------------------------------
>
>                 Key: YARN-7643
>                 URL: https://issues.apache.org/jira/browse/YARN-7643
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: capacity scheduler
>            Reporter: Suma Shivaprasad
>            Assignee: Suma Shivaprasad
>         Attachments: YARN-7643.1.patch
>
>
> CapacityScheduler application recovery should auto-create leaf queue if it doesnt exist.
Also RMAppManager needs to set the queue-mapping placement context so that scheduler has necessary
placement context to recreate the queue



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

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