hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Naganarasimha G R (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-6467) CSQueueMetrics needs to update the current metrics for default partition only
Date Fri, 02 Jun 2017 02:23:04 GMT

    [ https://issues.apache.org/jira/browse/YARN-6467?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16034044#comment-16034044

Naganarasimha G R commented on YARN-6467:

Thanks [~manirajv06@gmail.com] for incorporating the changes mentioned offline. This way it
will make simpler to support metrics for multiple partitions YARN-6492 .
Overall approach looks good but few nits in the patch is as follows

# QueueMetrics ln no 353 - 356: setAvailableResourcesToQueue(Resource) needs to call setAvailableResourcesToQueue(String
label, Resource limit). Similarly take care if any other methods are overloaded.
# QueueMetrics : please modify label to partition throughout the class where ever added
# QueueMetrics ln no 512 - 521: {{releaseResources(String, Resource)}} eihter need to consider
label, or may be removed if not used anywhere (check the history who added it)
# QueueMetrics : we need to capture what all we will capture at partition/queue/user level,
if not it will be confusing. for example {{allocatedContainers and aggregateContainersAllocated}},
we need to relook whether its at partition level only. we can discuss more on it.
# QueueMetrics ln 341 -345 : overloaded method setAvailableResourcesToQueue can be called
# FiCaSchedulerApp ln no 573 : unnecessary new line
# FiCaSchedulerApp ln no 574 : it should be node.getPartition() though container has the same
partition. in some narrow cases it might be wrong (partition sharing).
# FairScheduler ln no 1076 -1077 : changes not required

QueueMetrics : {{_decrPendingResources(int, Resource)}}, {{_incrPendingResources(int, Resource)}}
&  {{_incrPendingResources}} needs to consider partition too when updating pendingMB abd
pendingVcore. (add a note so that it can be handled in next jira)

P.S. please check and confirm whether the checkstyle issues and test case failure is not related
to your patch modifciations.

> CSQueueMetrics needs to update the current metrics for default partition only
> -----------------------------------------------------------------------------
>                 Key: YARN-6467
>                 URL: https://issues.apache.org/jira/browse/YARN-6467
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: capacity scheduler
>    Affects Versions: 2.8.0, 2.7.3, 3.0.0-alpha2
>            Reporter: Naganarasimha G R
>            Assignee: Manikandan R
>         Attachments: YARN-6467.001.patch, YARN-6467.001.patch, YARN-6467.002.patch
> As a followup to YARN-6195, we need to update existing metrics to only default Partition.

This message was sent by Atlassian JIRA

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

View raw message