[
https://issues.apache.org/jira/browse/HBASE-20100?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16379222#comment-16379222
]
stack commented on HBASE-20100:
-------------------------------
.001 Allow that region may already be in the OPEN state.
Allow OPEN as a possible state when we update region transition state.
Usually state is OPENING but if crash before finish step is completed,
on replay, master may have read that the state is OPEN from meta table
and so will think it open....
Small fix.
> TestEnableTableProcedure flakey
> -------------------------------
>
> Key: HBASE-20100
> URL: https://issues.apache.org/jira/browse/HBASE-20100
> Project: HBase
> Issue Type: Sub-task
> Components: amv2, flakey
> Reporter: stack
> Priority: Major
> Attachments: HBASE-20100.branch-2.001.patch
>
>
> Failed in the nightly in interesting way.
> https://builds.apache.org/job/HBase%20Nightly/job/branch-2/398/
> A subprocedure of enable table is assigning regions. The test is doing a kill
> of the procedure store to ensure we can handle all state transtions even in
> face of failure. In this case, the kill of PE framework came in the finish of
> the assign procedure AFTER we'd updated hbase:meta and the internal AM state
> moving Region from OPENING to OPEN. Rerunning this step on restoration of the
> AMv2 WAL Store results in....
> {code}
> 2018-02-27 16:28:02,290 WARN [PEWorker-1]
> assignment.RegionTransitionProcedure(331): Retryable error trying to
> transition: pid=64, ppid=52, state=RUNNABLE:REGION_TRANSITION_FINISH;
> AssignProcedure table=testRecoveryAndDoubleExecution,
> region=cdc8d2202f4aa4daf381bf1bf2e2fe0c; rit=OPEN,
> location=104080720bc8,38818,1519748808967
> org.apache.hadoop.hbase.exceptions.UnexpectedStateException: Expected
> [OFFLINE, CLOSED, SPLITTING, SPLIT, OPENING, FAILED_OPEN] so could move to
> OPEN but current state=OPEN
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
