DumpTaskTest causes intermittent failures with Clover

[Peter Beaman]

Periodically the persistit-coverage build fails with console output that looks like this:

Tests run: 4, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 104.061 sec
Running com.persistit.DumpTaskTest
Page 1 in volume persistit(/tmp/persistit_test_data/persistit) at index 1 timestamp=953 status=vd type=Data
Page 2 in volume persistit(/tmp/persistit_test_data/persistit) at index 2 timestamp=950 status=vd type=Data
PersistitDump_201211281645.000000000000
Tests run: 2, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 3.547 sec
Running com.persistit.JournalManagerTest
[main] WARNING Failure while registering MXBean: javax.management.InstanceAlreadyExistsException: com.persistit:type=Persistit
[main] WARNING Failure while registering MXBean: javax.management.InstanceAlreadyExistsException: com.persistit:type=Persistit,class=BufferPool.16384
[JOURNAL_FLUSHER] WARNING Journal flush operation took 4,920ms last 8 cycles average is 708ms
[JOURNAL_FLUSHER] WARNING Journal flush operation took 2,046ms last 8 cycles average is 539ms
[main] WARNING Failure while registering MXBean: javax.management.InstanceAlreadyExistsException: com.persistit:type=Persistit
[main] WARNING Failure while registering MXBean: javax.management.InstanceAlreadyExistsException: com.persistit:type=Persistit,class=BufferPool.16384

In each occurrence, it appears that the shutdown after DumpTaskTest failed to remove the ManagementMXBean, suggesting that the close() method did not complete correctly. This phenomenon has not been observed on normal builds.

At this point we think the failure is merely inconvenient and are therefore giving it LOW importance.