[jboss-jira] [JBoss JIRA] (WFLY-11343) Lock is not released when JTS is enabled and a timer is cancelled inside a transaction

[Tom Jenkinson (Jira)]

    [ https://issues.jboss.org/browse/WFLY-11343?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13665238#comment-13665238 ] 

Tom Jenkinson commented on WFLY-11343:
--------------------------------------

[~kabirkhan] the reason it would be different would be because it would be a semphore and not a reentrant lock. As the comment suggests when either a reaper times out the transaction or the transaction is propagated over a transport medium that uses a thread pool then different threads can end up calling the cancel that don't own the lock.

I was just transferring the suggestion from Stuart, EJB timer code is not my specialism either.

> Lock is not released when JTS is enabled and a timer is cancelled inside a transaction
> --------------------------------------------------------------------------------------
>
>                 Key: WFLY-11343
>                 URL: https://issues.jboss.org/browse/WFLY-11343
>             Project: WildFly
>          Issue Type: Bug
>          Components: EJB
>         Environment: EAP 6.4.18
>            Reporter: Shaun Appleton
>            Assignee: Tom Jenkinson
>            Priority: Major
>
> Description of problem:
> Lock is held by EJB timer (TimerServiceImpl)  and in case JTS transaction is cancelled the lock won't be released correctly . This code is the problem:
> https://github.com/wildfly/wildfly/blob/78dee79dd0f49c6cbd2b8db5d8640980faa08892/ejb3/src/main/java/org/jboss/as/ejb3/timerservice/TimerServiceImpl.java#L640
> Basically it holds the timer lock until the transaction completes, and then attempts to release it in afterCompletion. The problem is that when JTS is enabled afterCompletion will be called by a seperate thread, which can't call unlock as it is not the owner.
> A simple fix could be to just change the lock to a semaphore, so that the other thread can release it.
> Version-Release number of selected component (if applicable):
> How reproducible:
> Always
> Steps to Reproduce:
> 1. Start a JTA/JTS tx and call an EJB timer inside it
> 2. Make the transaction timeout
> 3. Capture a thread dump and see a thread like below  (0a0818b80 is locked object)
> ---------------------------------------------------------------------------------------------------------
> "Incoming-2,RouterPoliciesClusterGroup,svc-3-comecimpolicy-59228 (, payload=52 bytes)" #661 prio=5 os_prio=0 tid=0x00007f5214c2d000 nid=0x1e67 waiting on condition [0x00007f520ac97000]
>    java.lang.Thread.State: WAITING (parking)
>     at sun.misc.Unsafe.park(Native Method)
>     - parking to wait for  <0x00000000a0818b80> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
>     at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
>     at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
>     at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442)
>     at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1074)
>     at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1134)
>     at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
>     at java.lang.Thread.run(Thread.java:748)
>    Locked ownable synchronizers:
>     - <0x00000000ea1f4c48> (a java.util.concurrent.locks.ReentrantLock$NonfairSync)
> ---------------------------------------------------------------------------------------------------------
> Actual results:
> Lock being held forever
> Expected results:
> Lock getting released



--
This message was sent by Atlassian Jira
(v7.12.1#712002)