[jira] [Updated] (SOLR-7659) DirectoryFileStream.releaseCommitPointAndExtendReserve -- should it extend first?

["Shawn Heisey (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SOLR-7659?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Shawn Heisey updated SOLR-7659:
-------------------------------
    Attachment: SOLR-7659.patch

Patch (against branch_5x) that reverses the statements in the mentioned method.  It also renames the method.  CHANGES.txt mentions this in the "upgrading" section.

> DirectoryFileStream.releaseCommitPointAndExtendReserve -- should it extend first?
> ---------------------------------------------------------------------------------
>
>                 Key: SOLR-7659
>                 URL: https://issues.apache.org/jira/browse/SOLR-7659
>             Project: Solr
>          Issue Type: Improvement
>          Components: replication (java)
>    Affects Versions: 5.2.1
>            Reporter: Shawn Heisey
>            Priority: Minor
>             Fix For: 5.3, Trunk
>
>         Attachments: SOLR-7659.patch
>
>
> After a misunderstanding on my part, leading to an incorrect documentation edit, [~shalinmangar] helped me understand how the commit reserve duration works.
> The resulting discussion on the dev list brought up a possible problem in ReplicationHandler.  We wondered whether it might be possible for another thread to sneak in a commit point deletion in between the two statements found in the releaseCommitPointAndExtendReserve method on the DirectoryFileStream object.  If it is safe to do so, I propose reversing those two statements so the reserve extension is done before releasing the commit point.



--
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