[jira] [Commented] (SOLR-7836) Possible deadlock when closing refcounted index writers.

["Erick Erickson (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SOLR-7836?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644312#comment-14644312 ] 

Erick Erickson commented on SOLR-7836:
--------------------------------------

OK, the patch doesn't address the real issue. I still think it's bad to leave those dangling pauseWriters around, but the _real_ issue appears to be over in DirectUpdateHandler2. There's a ref counted index writer obtained in addDoc0. But then addDoc0 calls addAndDelete which tries to get a ref counted index writer again. If another thread sets pauseWriter in between, then it's deadlocked.

I think the solution is to just pass the IndexWriter down to addAndDelete, but won't have time to really look until this evening.


> 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
>         Attachments: 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: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org