[
https://issues.apache.org/jira/browse/LUCENE-4241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13419584#comment-13419584
]
Mark Miller commented on LUCENE-4241:
-------------------------------------
Off-topic now, but to keep this discussion in one place:
See the discussion here: http://markmail.org/message/txiwakxolegjqafv
I close the IW and tell it to wait for merges to finish. Then I close the
directory. Because this is a cloud test with lost of jetties running, at the
end of the test I have to shut them down - in many cases i want to test non
graceful shutdowns - I want to shut them down quickly. That causes an
interrupt, that causes IW.close to return, that causes the dir to be closed
while merge threads are still running.
>From my tests perspective, I really don't care that this happens.
I played around with always doing graceful shutdowns, but not only do I want to
test none graceful shutdowns, making everything graceful makes the tests a lot
longer. Also, I don't care about that exception even if it was killer, because
either im going to restart jetty and replace that index with recovery or we are
shutting down and the test is over.
I'm using the API's sanely - I can't control that interrupt easily.
> non-reproducible failures from RecoveryZkTest - mostly
> NRTCachingDirectory.deleteFile
> -------------------------------------------------------------------------------------
>
> Key: LUCENE-4241
> URL: https://issues.apache.org/jira/browse/LUCENE-4241
> Project: Lucene - Java
> Issue Type: Bug
> Reporter: Hoss Man
> Attachments:
> RecoveryZkTest.testDistribSearch-100-tests-failures.txt.tgz, just-failures.txt
>
>
> Since getting my new laptop, i've noticed some sporadic failures from
> RecoveryZkTest, so last night tried running 100 iterations againts trunk
> (r1363555), and got 5 errors/failures...
> * 3 asertion failures from NRTCachingDirectory.deleteFile
> * 1 node recovery assertion from
> AbstractDistributedZkTestCase.waitForRecoveriesToFinish caused by OOM
> * 1 searcher leak assertion: opens=1658 closes=1652 (possibly lingering
> affects from OOM?)
> see comments/attachments for details
--
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
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
