[MAPREDUCE-4924] flakey test: org.apache.hadoop.mapred.TestClusterMRNotification.testMR

• Type: Bug
• Status: Closed
• Priority: Trivial
• Resolution: Fixed
• Affects Version/s: 2.0.2-alpha
• Fix Version/s: 1.2.0, 2.0.3-alpha
• Component/s: mrv1
• Hadoop Flags: Reviewed
• Assignee: Robert Kanter
• Reporter: Robert Kanter
• Created: 08/Jan/13 01:37
• Updated: 15/Feb/13 13:10
• Resolved: 16/Jan/13 01:08

Description

I occasionally get a failure like this on org.apache.hadoop.mapred.TestClusterMRNotification.testMR

junit.framework.AssertionFailedError: expected:<6> but was:<4>
	at junit.framework.Assert.fail(Assert.java:47)
	at junit.framework.Assert.failNotEquals(Assert.java:283)
	at junit.framework.Assert.assertEquals(Assert.java:64)
	at junit.framework.Assert.assertEquals(Assert.java:195)
	at junit.framework.Assert.assertEquals(Assert.java:201)
	at org.apache.hadoop.mapred.NotificationTestCase.testMR(NotificationTestCase.java:184)
	...

It looks like a race condition:

    // run a job with FAILED status
    System.out.println(UtilsForTests.runJobFail(this.createJobConf(), inDir,
                                                outDir).getID());
    Thread.sleep(2000);
    assertEquals(6, NotificationServlet.counter);
    assertEquals(0, NotificationServlet.failureCounter);

Instead of sleeping for 2 seconds, we should keep checking the counter and fail after a timeout. There's a couple of similar places in the test that should be fixed too.