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-4712) CPU Usage Metric is not captured properly in YARN-2928
Date Thu, 03 Mar 2016 05:02:18 GMT

    [ https://issues.apache.org/jira/browse/YARN-4712?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15177177#comment-15177177

Naganarasimha G R commented on YARN-4712:

Thanks for the comments [~djp] & [~varun_saxena],\
bq. Regarding checkstyle, you can fix them for now.
As you can note in the latest patch line length issues are already taken care of.

bq. We shouldn't let Eclipse's bug affect our code convention.
Well its not that i dont want to do it, but i presume Eclipse optimizes it in some ways and
does only when required, Anyway have taken care of it but it would more easy to rely on the
editors formatter if accepted  :) 

bq.  it seems more things need to be fixed for UNAVAILABLE case,
agree , milliVcoresUsed can be set to 0 in UNAVAILABLE case, right ?

bq. It sounds weird if cpuUsageTotalCoresPercentage is -1 in UNAVAILABLE case.
we have set it -1 to indicate not to store this  value in the ATS. if its *unavaiable do we
need to store it as 0 or not store at all* ?

bq. it make cpu metric to be either 0 or 1 which is not expected here?
As [~varun_saxena] explained it directly gives as percent values (no need to multiply with
100) and we round of to only remove the decimals values

[~djp], if you can confirm on these queries, i can finish the patch

> CPU Usage Metric is not captured properly in YARN-2928
> ------------------------------------------------------
>                 Key: YARN-4712
>                 URL: https://issues.apache.org/jira/browse/YARN-4712
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelineserver
>            Reporter: Naganarasimha G R
>            Assignee: Naganarasimha G R
>              Labels: yarn-2928-1st-milestone
>         Attachments: YARN-4712-YARN-2928.v1.001.patch, YARN-4712-YARN-2928.v1.002.patch
> There are 2 issues with CPU usage collection 
> * I was able to observe that that many times CPU usage got from {{pTree.getCpuUsagePercent()}}
is ResourceCalculatorProcessTree.UNAVAILABLE(i.e. -1) but ContainersMonitor do the calculation
 i.e. {{cpuUsageTotalCoresPercentage = cpuUsagePercentPerCore /resourceCalculatorPlugin.getNumProcessors()}}
because of which UNAVAILABLE check in {{NMTimelinePublisher.reportContainerResourceUsage}}
is not encountered. so proper checks needs to be handled
> * {{EntityColumnPrefix.METRIC}} uses always LongConverter but ContainerMonitor is publishing
decimal values for the CPU usage.

This message was sent by Atlassian JIRA

View raw message