[
https://issues.apache.org/jira/browse/MAPREDUCE-5505?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13772291#comment-13772291
]
Jian He commented on MAPREDUCE-5505:
------------------------------------
{code}
case REBOOT:
if (isUnregistered && appContext.isLastAMRetry()) {
return JobState.ERROR;
{code}
We probably don't need "appContext.isLastAMRetry()" this check here, since if
this is the last retry, the app will fail on RM side. After this AM exits,
JobClient is able to query RM for final status, in which case JobClient will be
told FAILED. This is good for all transitions to follow the same logic.
If that's the case we can create a common function to handle the logic that for
every final state, if registered return final state, otherwise return the
previous state
> Clients should be notified job finished only after job successfully
> unregistered
> ---------------------------------------------------------------------------------
>
> Key: MAPREDUCE-5505
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5505
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Jian He
> Assignee: Zhijie Shen
> Attachments: MAPREDUCE-5505.1.patch, MAPREDUCE-5505.1.patch
>
>
> This is to make sure user is notified job finished after job is really done.
> This does increase client latency but can reduce some races during unregister
> like YARN-540
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira