Uploaded image for project: 'Infinispan'
  1. Infinispan
  2. ISPN-3055

Unreleased lock after node restart

This issue belongs to an archived project. You can view it, but you can't modify it. Learn more

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Major Major
    • 5.3.0.Final
    • 5.2.1.Final
    • None
    • Hide

      Step 1. Run SimpleClusterWriter and SimpleClusterRead. Test rollback functionality.
      Step 2. Kill SimpleClusterWriter and start it again.
      Step 3. Make rollback.
      Step 4. Make normal commit which leads to acquired lock from the previous rollback?!?

      Will change old value => 724 to the new value => 513
      Apr 29, 2013 3:26:58 PM org.infinispan.interceptors.InvocationContextInterceptor handleAll
      ERROR: ISPN000136: Execution error
      org.infinispan.util.concurrent.TimeoutException: Unable to acquire lock after [1 seconds] on key [k] for requestor [DldGlobalTransaction

      {coinToss=5887910943873340717, lockIntention=k, affectedKeys=[], locksAtOrigin=[]}

      GlobalTransaction:<node1-15887>:3:remote]! Lock held by [DldGlobalTransaction

      {coinToss=-39459977763357633, lockIntention=null, affectedKeys=[], locksAtOrigin=[]}

      GlobalTransaction:<node1-15887>:2:remote]
      at org.infinispan.util.concurrent.locks.LockManagerImpl.lock(LockManagerImpl.java:213)
      at org.infinispan.util.concurrent.locks.LockManagerImpl.acquireLock(LockManagerImpl.java:186)

      Show
      Step 1. Run SimpleClusterWriter and SimpleClusterRead. Test rollback functionality. Step 2. Kill SimpleClusterWriter and start it again. Step 3. Make rollback. Step 4. Make normal commit which leads to acquired lock from the previous rollback?!? Will change old value => 724 to the new value => 513 Apr 29, 2013 3:26:58 PM org.infinispan.interceptors.InvocationContextInterceptor handleAll ERROR: ISPN000136: Execution error org.infinispan.util.concurrent.TimeoutException: Unable to acquire lock after [1 seconds] on key [k] for requestor [DldGlobalTransaction {coinToss=5887910943873340717, lockIntention=k, affectedKeys=[], locksAtOrigin=[]} GlobalTransaction:<node1-15887>:3:remote]! Lock held by [DldGlobalTransaction {coinToss=-39459977763357633, lockIntention=null, affectedKeys=[], locksAtOrigin=[]} GlobalTransaction:<node1-15887>:2:remote] at org.infinispan.util.concurrent.locks.LockManagerImpl.lock(LockManagerImpl.java:213) at org.infinispan.util.concurrent.locks.LockManagerImpl.acquireLock(LockManagerImpl.java:186)

      Using dummy transaction manager I was able to produce locks that are not released. The same behavior can be reproduced with Glassfish AS 3.1.1 using its own TM.

        1. simple_cluster_node1.xml
          1 kB
        2. simple_cluster_node2.xml
          1 kB
        3. SimpleClusterRead.java
          0.7 kB
        4. SimpleClusterWriter.java
          2 kB

            [ISPN-3055] Unreleased lock after node restart

            Mircea Markus (Inactive) added a comment -

            integrated.

            Mircea Markus (Inactive) added a comment - integrated.

            Pedro Ruivo added a comment -

            I've take a look at it and the good new is it has nothing to do with the node restart

            If you start firs the Reader and then the Writer and abort a transaction, the bug will be revealed.

            the bug is quite simple. we are not sending the RollbackCommand if the lockingMode==PESSIMISTIC.

            Thanks Deyan!

            Pedro Ruivo added a comment - I've take a look at it and the good new is it has nothing to do with the node restart If you start firs the Reader and then the Writer and abort a transaction, the bug will be revealed. the bug is quite simple. we are not sending the RollbackCommand if the lockingMode==PESSIMISTIC. Thanks Deyan!

            Deyan Pandulev (Inactive) added a comment -

            I've tested with the suggested 5.2.6 Final version of Infinispan and there is no change in the behavior described in the bug.

            Deyan Pandulev (Inactive) added a comment - I've tested with the suggested 5.2.6 Final version of Infinispan and there is no change in the behavior described in the bug.

            Mircea Markus (Inactive) added a comment -

            Indeed looks like a bug - thanks for reporting it. Can you give it a try with latest 5.2 release (5.2.6 today)?

            Mircea Markus (Inactive) added a comment - Indeed looks like a bug - thanks for reporting it. Can you give it a try with latest 5.2 release (5.2.6 today)?

              pruivo@redhat.com Pedro Ruivo
              dpandulev Deyan Pandulev (Inactive)
              Archiver:
              rhn-support-adongare Amol Dongare

                Created:
                Updated:
                Resolved:
                Archived: