"dukehoops" wrote : OK, I wired up DualNodeJtaTransactionManagerImpl. The test now passes if UserCount is below 4 or less, at 5 or above the test fails. However, the failure I am seeing with this test is NOT the same as with the test in our environment. It is this: | | anonymous wrote : ------------------------------------------------------------------------------- | | Test set: org.hibernate.test.cache.jbc2.functional.MVCCConcurrentWriteTest | | ------------------------------------------------------------------------------- | | Tests run: 6, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 182.113 sec <<< FAILURE! | | testManyUsers(org.hibernate.test.cache.jbc2.functional.MVCCConcurrentWriteTest) Time elapsed: 165.831 sec <<< FAILURE! | | junit.framework.AssertionFailedError: Timed out waiting for user threads to finish. Their state at the time of forced shutdown: TEST CONFIG [userCount=5, iterationsPerUser=40, thinkTimeMillis=10] STATE of UserRunners: org.hibernate.test.cache.jbc2.functional.MVCCConcurrentWriteTest$UserRunner@ffc3eb[customerId=4 iterationsCompleted=30 completedAll=false causeOfFailure=org.hibernate.cache.CacheException: org.hibernate.cache.CacheException: org.jboss.cache.lock.TimeoutException: Unable to acquire lock on Fqn [/TS/test/org/hibernate/cache/UpdateTimestampsCache/Contacts] after [15000] milliseconds for requestor [Thread[UserRunnerThread-4,5,main]]! Lock held by [GlobalTransaction::903] | | at ... | | | | I missed this UpdateTimestampsCache detail last week. Another user has reported a similar lock conflict over the UpdateTimestampsCache -- see http://www.jboss.org/index.html?module=bb&op=viewtopic&t=152079. With your test I should be able to reproduce. There shouldn't be long-lived locks held on the UpdateTimestampsCache; if there are it's a bug. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4218790#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...