[
https://issues.apache.org/jira/browse/JCR-3469?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13508652#comment-13508652
]
Marcel Reutegger commented on JCR-3469:
---------------------------------------
Looks like this is solaris specific. More background information is available
here: http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4385444
On Solaris the HotSpot option UseVMInterruptibleIO is enabled per default with
Java 6. This makes I/O behaviour in presence of Thread interrupts platform
specific. I'll try to reproduce the test failure on a Solaris instance...
> Thread interrupt may result in closed index files
> -------------------------------------------------
>
> Key: JCR-3469
> URL: https://issues.apache.org/jira/browse/JCR-3469
> Project: Jackrabbit Content Repository
> Issue Type: Bug
> Components: jackrabbit-core
> Affects Versions: 2.4
> Reporter: Marcel Reutegger
> Assignee: Marcel Reutegger
> Priority: Minor
> Fix For: 2.6
>
> Attachments: JCR-3469.patch, JCR-3469.patch, JCR-3469-test.patch
>
>
> Interrupting a thread, which is currently performing I/O on Lucene index
> files, may result in closed index files. Jackrabbit versions prior to 2.4
> didn't have that problem because only with the update of Lucene to 3.0.3 in
> Jackrabbit 2.4 the underlying default for the FSDirectory in Lucene changed
> from SimpleFSDirectory to NIOFSDirectory. The latter uses a FileChannel to
> perform concurrent read operations on a single channel, but because of the
> way how FileChannel is specified an interrupt will result in the FileChannel
> to be closed.
> This was also discussed on the Lucene mailing list. See:
> http://lucene.markmail.org/thread/ea7mj2kpdvyuesss
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
