[
https://issues.apache.org/jira/browse/MAPREDUCE-5155?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13634784#comment-13634784
]
nemon lou commented on MAPREDUCE-5155:
--------------------------------------
Another test case failed with the same reason:
Failed tests: testUpdatedNodes(org.apache.hadoop.mapreduce.v2.app.TestMRApp):
Expecting 2 more completion events for killed expected:<4> but was:<2>
These lines of code has race condition issue:
{quote}
app.waitForState(task1Attempt, TaskAttemptState.KILLED);
app.waitForState(task2Attempt, TaskAttemptState.KILLED);
events = job.getTaskAttemptCompletionEvents(0, 100);
Assert.assertEquals("Expecting 2 more completion events for killed", 4,
events.length);
{quote}
> Race condition in test case TestFetchFailure cause it to fail
> -------------------------------------------------------------
>
> Key: MAPREDUCE-5155
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5155
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: test
> Affects Versions: 2.0.3-alpha
> Environment: Suse x86_64 GNU/Linux
> Java(TM) SE Runtime Environment (build 1.6.0_32-b05
> Reporter: nemon lou
> Priority: Minor
> Attachments:
> org.apache.hadoop.mapreduce.v2.app.TestFetchFailure-output.txt,
> org.apache.hadoop.mapreduce.v2.app.TestFetchFailure.txt
>
>
> I run into this once:
> testFetchFailureWithRecovery(org.apache.hadoop.mapreduce.v2.app.TestFetchFailure):
> Num completion events not correct expected:<1> but was:<0>
> There is a race condition between job.getTaskAttemptCompletionEvents and
> dealing with JOB_TASK_ATTEMPT_COMPLETED event.
> If job.getTaskAttemptCompletionEvents invoked because of task in SUCCEEDED
> state ,but before JOB_TASK_ATTEMPT_COMPLETED event scheduled,the test case
> will fail.
--
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