[
https://issues.apache.org/jira/browse/YARN-514?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13634841#comment-13634841
]
Zhijie Shen commented on YARN-514:
----------------------------------
I've run through the yarn test cases locally, and found no more failures.
I also find problem why the final assert still fails. As RMApp have one more
state in its life cycle, the direct call:
{code}
application.handle(new RMAppEvent(applicationId, RMAppEventType.START));
{code}
moves application from NEW to NEW_SAVING. We still need to give the application
some time to transit from NEW_SAVING to SUMBITTED, where the attempt is created
and RMAppAttemptEventType.START is scheduled. The current code in
TestFairScheduler#testNotAllowSubmitApplication results in that
RMAppAttemptEventType.APP_REJECTED comes earlier than
RMAppAttemptEventType.START. Therefore, the attempt gets APP_REJECTED event
when it is still in NEW. This transition happens not to be defined.
So, the quick test case fix is that add "Thread.sleep(100);" immediately after
"application.handle(new RMAppEvent(applicationId, RMAppEventType.START));" to
yield the process for asyncDispatcher to push application forward.
> Delayed store operations should not result in RM unavailability for app
> submission
> ----------------------------------------------------------------------------------
>
> Key: YARN-514
> URL: https://issues.apache.org/jira/browse/YARN-514
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: resourcemanager
> Reporter: Bikas Saha
> Assignee: Zhijie Shen
> Attachments: YARN-514.1.patch, YARN-514.2.patch, YARN-514.3.patch,
> YARN-514.4.patch, YARN-514.5.patch, YARN-514.6.patch, YARN-514.7.patch,
> YARN-514.8.patch
>
>
> Currently, app submission is the only store operation performed synchronously
> because the app must be stored before the request returns with success. This
> makes the RM susceptible to blocking all client threads on slow store
> operations, resulting in RM being perceived as unavailable by clients.
--
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
