[
https://issues.apache.org/jira/browse/AMQ-3301?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Arthur Naseef updated AMQ-3301:
-------------------------------
Attachment: amq-3301.patch
Patch which eliminates the race condition on the in-JVM lock test.
> LockFile in-jvm check is not thread-safe
> ----------------------------------------
>
> Key: AMQ-3301
> URL: https://issues.apache.org/jira/browse/AMQ-3301
> Project: ActiveMQ
> Issue Type: Bug
> Components: Broker
> Affects Versions: 5.4.2
> Reporter: Arthur Naseef
> Priority: Minor
> Attachments: amq-3301.patch
>
>
> kahadb/src/main/java/org/apache/kahadb/util/LockFile.java adds in-JVM locking
> to Channel.tryLock() using system properties. However, there is time between
> the read-and-check for an existing lock and setting a new lock.
> I found this problem while running multiple brokers in a single JVM for
> diagnostic purposes with two brokers using a shared directory to simlute a
> production H/A environment.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
