[
https://issues.apache.org/jira/browse/SOLR-7836?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14682096#comment-14682096
]
Erick Erickson commented on SOLR-7836:
--------------------------------------
Thanks guys, getting it all in my head is...interesting.
[[email protected]] Yeah, I saw that comment. In that case, removing the two
synchronizations in the refactored methods _other_ than addAndDelete is
probably indicated. The one _in_ addAndDelete was there originally, just within
the IndexWriter try/finally which is where the issues was since it'd go out and
get a new searcher eventually.
[[email protected]] I had to write a shell script to re-submit that
individual test repeatedly, it'd pretty much always fail for me by 50 runs.
I'll back those changes out and run it on my machine where it fails reliably
and post the results when I get a deadlock.
> 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
>
>
> 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]
