[
https://issues.apache.org/jira/browse/LUCENE-3673?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Adrien Grand resolved LUCENE-3673.
----------------------------------
Resolution: Fixed
This seems to have been fixed by other changes. Testing also improved via
LUCENE-5624.
> NativeFSLockFactory Race Condition
> ----------------------------------
>
> Key: LUCENE-3673
> URL: https://issues.apache.org/jira/browse/LUCENE-3673
> Project: Lucene - Core
> Issue Type: Bug
> Components: core/store
> Affects Versions: 3.4, 3.5
> Environment: Mac OS X 10.7, JDK 1.6
> CentOS 5, JDK 1.7u1
> Reporter: Yaokai Jiang
> Priority: Critical
> Labels: concurrency, locking
> Attachments: BreakIt.java, LUCENE-3673.patch
>
>
> When the NativeFSLock releases, it deletes the lock file after lock release.
> In a concurrent situation, it is possible for another thread or process to
> acquire the lock before the delete happens. Any other third thread or process
> will then be able to recreate the lock file and acquire the lock while a lock
> is already held on the deleted file.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
