[
https://issues.apache.org/jira/browse/HBASE-11641?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14084197#comment-14084197
]
Lars Hofhansl commented on HBASE-11641:
---------------------------------------
Here's at least something that's weird. In startCluster we call
TEST_UTIL.shutdownMiniHBaseCluster() even though that happens on teardown of
each test anyway. shutdownMiniHBaseCluster() happens to reset
WAIT_ON_REGIONSERVERS_MINTOSTART, which testMasterStartsUpWithLogSplittingWork
sets. I don't think that causes the issue, unless somehow leads to timing
issues upon shutdown (i.e. the RSs haven't even started, yet).
I assume this was a leftover from when the test was changed to shutdown the
cluster in the teardown method (and it is probably something I did).
> TestDistributedLogSplitting.testMasterStartsUpWithLogSplittingWork fails
> frequently
> -----------------------------------------------------------------------------------
>
> Key: HBASE-11641
> URL: https://issues.apache.org/jira/browse/HBASE-11641
> Project: HBase
> Issue Type: Bug
> Reporter: Lars Hofhansl
> Attachments: 11641-logs.txt
>
>
> 3 out of last four run of my RC jenkins build failed in this test.
> Might be related to HBASE-11041 too.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
