[
https://issues.apache.org/jira/browse/YARN-3995?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15070690#comment-15070690
]
Sangjin Lee commented on YARN-3995:
-----------------------------------
If I recall, this window of opportunity is going to be quite small because any
non-AM container will be completed before the app can be finished (and the AM
container is completed). For this inversion to occur, there would have to be
writes that originate from a remote NM that had a container (which had already
been completed) but get delayed in reaching the timeline collector for some
reason.
I suspect a simple linger might be sufficient, but do we see a case where we
might miss writes otherwise?
> Some of the NM events are not getting published due race condition when AM
> container finishes in NM
> ----------------------------------------------------------------------------------------------------
>
> Key: YARN-3995
> URL: https://issues.apache.org/jira/browse/YARN-3995
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: nodemanager, timelineserver
> Affects Versions: YARN-2928
> Reporter: Naganarasimha G R
> Assignee: Naganarasimha G R
> Labels: yarn-2928-1st-milestone
>
> As discussed in YARN-3045: While testing in TestDistributedShell found out
> that few of the container metrics events were failing as there will be race
> condition. When the AM container finishes and removes the collector for the
> app, still there is possibility that all the events published for the app by
> the current NM and other NM are still in pipeline,
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
