[
https://issues.apache.org/jira/browse/HBASE-6636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13439950#comment-13439950
]
stack commented on HBASE-6636:
------------------------------
Trying to backport HBASE-5549. It removes the test just before the failing one
(because its useless). This test kills the master. The failing test is
failing because it can't find newly launched master... fails with connection
refused. HBASE-5549 also has improvements around finding new active master in
cluster and does other clean up in tests around zk usage.
> [0.92 UNIT TESTS]
> TestMasterZKSessionRecovery.testRegionAssignmentAfterMasterRecoveryDueToZKExpiry
> failed in 0.92 #513 and #496
> -------------------------------------------------------------------------------------------------------------------------------
>
> Key: HBASE-6636
> URL: https://issues.apache.org/jira/browse/HBASE-6636
> Project: HBase
> Issue Type: Bug
> Affects Versions: 0.92.1
> Reporter: stack
> Assignee: stack
> Priority: Critical
>
> The test class TestMasterZKSessionRecovery has been removed in trunk. Its
> master tests were moved elsewhere or removed because useless (See nkeywal
> reasoning over in HBASE-5572 "KeeperException.SessionExpiredException
> management could be improved in Master"; it was actually removed by
> HBASE-5549 "Master can fail if ZooKeeper session expires").
> TestMasterZKSessionRecovery in 0.92 and 0.94 has an extra test that was not
> in trunk, the sporadically failing
> testRegionAssignmentAfterMasterRecoveryDueToZKExpiry. This was added by
> "HBASE-6046
> Master retry on ZK session expiry causes inconsistent region assignments".
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
