[ACCUMULO-3892] Problematic ITs

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Cannot Reproduce
• Affects Version/s: 1.7.0
• Component/s: test
• Assignee: Unassigned
• Reporter: Josh Elser
• Created: 06/Jun/15 00:19
• Updated: 19/Mar/18 17:57
• Resolved: 19/Mar/18 17:57

Description

There are a few tests that I've regularly seem not want to pass during nightly automated testing.

E AssertionError: test timed out after 300000 milliseconds
E java.lang.Exception: test timed out after 300000 milliseconds
E at java.lang.Thread.sleep(Native Method)
E at org.apache.accumulo.core.util.UtilWaitThread.sleep(UtilWaitThread.java:27)
E at org.apache.accumulo.core.client.impl.TableOperationsImpl.waitForTableStateTransition(TableOperationsImpl.java:1159)
E at org.apache.accumulo.core.client.impl.TableOperationsImpl.online(TableOperationsImpl.java:1223)
E at org.apache.accumulo.test.AssignmentThreadsIT.testConcurrentAssignmentPerformance(AssignmentThreadsIT.java:77)

E AssertionError: test timed out after 90000 milliseconds
E java.lang.Exception: test timed out after 90000 milliseconds
E at sun.misc.Unsafe.park(Native Method)
E at java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:226)
E at java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedNanos(AbstractQueuedSynchronizer.java:1033)
E at java.util.concurrent.locks.AbstractQueuedSynchronizer.tryAcquireSharedNanos(AbstractQueuedSynchronizer.java:1326)
E at java.util.concurrent.CountDownLatch.await(CountDownLatch.java:282)
E at org.apache.accumulo.core.client.impl.TableOperationsImpl.addSplits(TableOperationsImpl.java:388)
E at org.apache.accumulo.test.BalanceFasterIT.test(BalanceFasterIT.java:60)

E AssertionError: test timed out after 360000 milliseconds
E java.lang.Exception: test timed out after 360000 milliseconds
E at java.lang.Object.wait(Native Method)
E at java.lang.Thread.join(Thread.java:1281)
E at java.lang.Thread.join(Thread.java:1355)
E at org.apache.accumulo.test.InterruptibleScannersIT.test(InterruptibleScannersIT.java:98)

E AssertionError: test timed out after 1440000 milliseconds
E java.lang.Exception: test timed out after 1440000 milliseconds
E at java.lang.Object.wait(Native Method)
E at java.lang.Object.wait(Object.java:503)
E at java.lang.UNIXProcess.waitFor(UNIXProcess.java:263)
E at org.apache.accumulo.test.functional.MetadataMaxFilesIT.test(MetadataMaxFilesIT.java:84)

This is with a quadrupled timeout (if not 8x by now). I know these tests all can pass (as I ensured that doing 1.7.0 testing), but apparently something isn't good enough when running on a EC2 m1.xlarges IIRC. I believe all of these tests are stretching the bounds on what is really suitable for an integration test, IMO. Perhaps there is a better way to write the test and verify the correctness that we want expect to see, but these tests are definitely not reliably testing what we hope they are for me.