[
https://issues.apache.org/jira/browse/YARN-585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13635454#comment-13635454
]
Bikas Saha commented on YARN-585:
---------------------------------
It looks like the test is completely bypassing the state machine of the RMApp
and so will break every time the state machine is changed in a way the test
does not handle. It directly sends scheduler.addApplication() without waiting
for the app to reach the state in which this can happen. I wonder if the
addapplication is being called twice in here? The test doesnt even call
resourcemanager.start() even though its using the resource manager.
{code}
application.handle(new RMAppEvent(applicationId, RMAppEventType.START));
ApplicationAttemptId attId =
recordFactory.newRecordInstance(ApplicationAttemptId.class);
attId.setAttemptId(this.ATTEMPT_ID++);
attId.setApplicationId(applicationId);
scheduler.addApplication(attId, queue, user);
{code}
> TestFairScheduler#testNotAllowSubmitApplication is broken due to YARN-514
> -------------------------------------------------------------------------
>
> Key: YARN-585
> URL: https://issues.apache.org/jira/browse/YARN-585
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Zhijie Shen
> Assignee: Zhijie Shen
> Fix For: 2.0.5-beta
>
> Attachments: YARN-585.1.patch, YARN-585.2.patch
>
>
> TestFairScheduler#testNotAllowSubmitApplication is broken due to YARN-514.
> See the discussions in YARN-514.
--
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
