[
https://issues.apache.org/jira/browse/YARN-4329?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15542134#comment-15542134
]
Naganarasimha G R commented on YARN-4329:
-----------------------------------------
Thanks for working on the patch [~yufeigu] and sorry for the delayed response,
One major comment would be in CS when a node got skipped we were capturing that
info too refer @
{{FiCaSchedulerApp.updateNodeInfoForAMDiagnostics(FiCaSchedulerNode)}}, this
would give some information about why a given node missed to schedule the app
too. Since the approach is different may be we need to handle it slightly
differently here but it would be good to capture
And other minor nits :
* SchedulerApplicationAttempt lno 1030 : Is this comment req ?
* FSAppAttempt lno 820..833 : line indentation is wrong, needs formatting
* FSAppAttempt lno 838..844 & 995..1000: looks similar may be can be optimized
by having a private method ?
* TestMaxRunningAppsEnforcer : test cases seems to not to cover all the
scenarios
> Allow fetching exact reason as to why a submitted app is in ACCEPTED state in
> Fair Scheduler
> --------------------------------------------------------------------------------------------
>
> Key: YARN-4329
> URL: https://issues.apache.org/jira/browse/YARN-4329
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: fairscheduler, resourcemanager
> Reporter: Naganarasimha G R
> Assignee: Yufei Gu
> Attachments: YARN-4329.001.patch, YARN-4329.002.patch
>
>
> Similar to YARN-3946, it would be useful to capture possible reason why the
> Application is in accepted state in FairScheduler
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
