[
https://issues.apache.org/jira/browse/SOLR-7836?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14695714#comment-14695714
]
Erick Erickson edited comment on SOLR-7836 at 8/13/15 6:20 PM:
---------------------------------------------------------------
bq: It looks fine to me that you moved the ulog.add() outside of the writer
get/release.
That's exactly what makes me nervous. But having it _within_ the get/release is
where the problem was.
Let me know if there's anything you'd like to chat about.
And it wasn't predictably repeatable (of course). Sometimes the script I
attached would run 10-15 iterations OK then barf. Sometimes faster. Never more
than 50. With the changes I ran with 400-500 iterations without a problem.
Since I can get it to fail reasonably predictably though, let me know if you
want me to run any changes locally.
was (Author: erickerickson):
bq: It looks fine to me that you moved the ulog.add() outside of the writer
get/release.
Exactly what makes me nervous. But having it _within_ the get/release is where
the problem was.
Let me know if there's anything you'd like to chat about.
> Possible deadlock when closing refcounted index writers.
> --------------------------------------------------------
>
> Key: SOLR-7836
> URL: https://issues.apache.org/jira/browse/SOLR-7836
> Project: Solr
> Issue Type: Bug
> Reporter: Erick Erickson
> Assignee: Erick Erickson
> Fix For: Trunk, 5.4
>
> Attachments: SOLR-7836-synch.patch, SOLR-7836.patch, SOLR-7836.patch,
> SOLR-7836.patch, deadlock_3.res.zip, deadlock_5_pass_iw.res.zip, deadlock_test
>
>
> Preliminary patch for what looks like a possible race condition between
> writerFree and pauseWriter in DefaultSorlCoreState.
> Looking for comments and/or why I'm completely missing the boat.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
