stack commented on HBASE-5833:

Looking at this more, it seems like the actual TMF tests run pretty fast but 
then the test gets stuck a long time finishing up on the end... like 10x more 
time finishing than was spent running the test.  This long cleanup seems to be 
our 'timeout'.  In TMF case, digging, there are orphan logsyncer threads 
outstanding trying to sync a fileystem that has been since closed.  This seems 
to be part of the issue (There are other orphan threads -- executors).  TMF 
does HRegion.createRegions but it doesn't pass in a WAL to use in creation.   
In this case, this method makes an individual WAL for the create use.  In TMF, 
and in a bunch of other tests, this hlog is never closed leaving logsyncer 
threads running (and some executors).  Will attach a patch that goes through 
all tests and does close and cleanup of wals that use this special 
createRegions method (there are a bunch).
> 0.92 build has been failing pretty consistently on TestMasterFailover....
> -------------------------------------------------------------------------
>                 Key: HBASE-5833
>                 URL: https://issues.apache.org/jira/browse/HBASE-5833
>             Project: HBase
>          Issue Type: Bug
>            Reporter: stack
>            Assignee: stack
>             Fix For: 0.92.2
>         Attachments: 5833.txt
> Trunk seems fine but 0.92 fails on this test pretty regularly.  Running it 
> local it seems to hang for me.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
For more information on JIRA, see: http://www.atlassian.com/software/jira


Reply via email to