[
https://issues.apache.org/jira/browse/MAPREDUCE-5561?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13801874#comment-13801874
]
Jason Lowe commented on MAPREDUCE-5561:
---------------------------------------
Yes, the test is definitely racy. There's no guarantee the job will be in the
FAIL_ABORT state while when we look at it asynchronously. A couple of
approaches to fixing this:
# As [~kkambatl] points out, we can skip the FAIL_ABORT check. The real
purpose of this test is to verify we eventually get to the FAILED state without
hanging when tasks fail.
# A more deterministic, explicit test for FAIL_ABORT would be to use an output
committer with a barrier, similar to TestingOutputCommitter but with the
barrier in the abortJob method, so we can guarantee the job will pause in the
FAIL_ABORT state. Then we can release the committer from the barrier and
verify the job proceeds to failed.
> org.apache.hadoop.mapreduce.v2.app.job.impl.TestJobImpl testcase failing on
> trunk
> ---------------------------------------------------------------------------------
>
> Key: MAPREDUCE-5561
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5561
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Affects Versions: 2.2.0
> Reporter: Cindy Li
> Assignee: Karthik Kambatla
> Priority: Critical
> Attachments:
> org.apache.hadoop.mapreduce.v2.app.job.impl.TestJobImpl-output.txt
>
>
> Running org.apache.hadoop.mapreduce.v2.app.job.impl.TestJobImpl
> Tests run: 15, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 9.029 sec
> <<< FAILURE! - in org.apache.hadoop.mapreduce.v2.app.job.impl.TestJobImpl
> testFailAbortDoesntHang(org.apache.hadoop.mapreduce.v2.app.job.impl.TestJobImpl)
> Time elapsed: 5.507 sec <<< FAILURE!
> java.lang.AssertionError: expected:<FAIL_ABORT> but was:<FAILED>
> at org.junit.Assert.fail(Assert.java:93)
> at org.junit.Assert.failNotEquals(Assert.java:647)
> at org.junit.Assert.assertEquals(Assert.java:128)
> at org.junit.Assert.assertEquals(Assert.java:147)
> at
> org.apache.hadoop.mapreduce.v2.app.job.impl.TestJobImpl.assertJobState(TestJobImpl.java:817)
> at
> org.apache.hadoop.mapreduce.v2.app.job.impl.TestJobImpl.testFailAbortDoesntHang(TestJobImpl.java:418)
--
This message was sent by Atlassian JIRA
(v6.1#6144)
