[
https://issues.apache.org/jira/browse/AMBARI-18289?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15453975#comment-15453975
]
Hadoop QA commented on AMBARI-18289:
------------------------------------
{color:red}-1 overall{color}. Here are the results of testing the latest
attachment
http://issues.apache.org/jira/secure/attachment/12826538/AMBARI-18289-2.patch
against trunk revision .
{color:red}-1 patch{color}. Top-level trunk compilation may be broken.
Console output:
https://builds.apache.org/job/Ambari-trunk-test-patch/8546//console
This message is automatically generated.
> Invalid negative values for some AMS metrics
> --------------------------------------------
>
> Key: AMBARI-18289
> URL: https://issues.apache.org/jira/browse/AMBARI-18289
> Project: Ambari
> Issue Type: Bug
> Components: ambari-metrics
> Affects Versions: 2.4.0
> Reporter: Aravindan Vijayan
> Assignee: Aravindan Vijayan
> Priority: Critical
> Fix For: 2.5.0
>
> Attachments: AMBARI-18289-2.patch
>
>
> PROBLEM
> Negative values are being reported for IOPS in the "system servers" dashboard
> BUG
> This was a dockerized environment. Negative rate values were seen because
> read/write counters were dropping below the previous values at random times.
> On further investigation, it revealed that this was due to the docker volume
> groups present on the host. It is expected of docker containers to add/remove
> the volume groups during the lifecyle of a container. So, when a volume group
> goes away, the read/write counters do not contribute to the total counter
> values, thus making the value go below the last seen value.
> FIX
> Have a provision to discard such "special" disk partitions through a skip
> pattern. Individual disk counter metrics can be used to get disk specific
> counter values. If skipped, they will not contribute to the global counter
> metric.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
