[
https://issues.apache.org/jira/browse/YARN-3513?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14535026#comment-14535026
]
Naganarasimha G R commented on YARN-3513:
-----------------------------------------
Thanks for the comment [~devaraj.k],
its a good catch but IMHO, i would like to retain the current approach as
* even if values go wrong its only for one run of the monitor and doesn't fail
anything.
* change of log level is rare operation
* better to save unnecessary computations for each run of monitoring than
compared to single invalid log.
Open to remove this check if you still feel necessary :)
> Remove unused variables in ContainersMonitorImpl and add debug log for
> overall resource usage by all containers
> ----------------------------------------------------------------------------------------------------------------
>
> Key: YARN-3513
> URL: https://issues.apache.org/jira/browse/YARN-3513
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: nodemanager
> Reporter: Naganarasimha G R
> Assignee: Naganarasimha G R
> Priority: Trivial
> Labels: BB2015-05-TBR, newbie
> Attachments: YARN-3513.20150421-1.patch, YARN-3513.20150503-1.patch,
> YARN-3513.20150506-1.patch, YARN-3513.20150507-1.patch,
> YARN-3513.20150508-1.patch, YARN-3513.20150508-1.patch
>
>
> Some local variables in MonitoringThread.run() : {{vmemStillInUsage and
> pmemStillInUsage}} are not used and just updated.
> Instead we need to add debug log for overall resource usage by all containers
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
