Well, I see that at the start of the execution, even on the next retry, the isValid is true for the statement. Something comes along and changes the state of isValid. I am having trouble finding out what with the debugger because there are quite a few threads running (statistics daemon for example) and the makeInvalid method is being invoked from various places and threads. So catching the correct one is being difficult to I am going to drop back to the old "printf" method and try to track the thread that sets the isValid to true and the thread that calls makeInvalid and try to match them up.
The fix definitely fixed my problem. There is something in my code that is triggering the deadlock and it is that statement that you mention below that is the one that is causing problems. But I want to understand why the fix is needed and how it gets into this state in the first place. The I will look at my code to figure out why this statement causes a problem. -----Original Message----- From: Kathey Marsden (Commented) (JIRA) [mailto:j...@apache.org] Sent: Thursday, December 22, 2011 8:20 PM To: derby-dev@db.apache.org Subject: [jira] [Commented] (DERBY-5552) Derby threads hanging when using ClientXADataSource and a deadlock or lock timeout occurs [ https://issues.apache.org/jira/browse/DERBY-5552?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13175216#comment-13175216 ] Kathey Marsden commented on DERBY-5552: --------------------------------------- Yes, definitely you want your test to be the new Junit style so that is the place to look. I think I have solved the mystery of the missing XA client traces and filed DERBY-5553 for that issue. I am trying to think of a scenario where we we execute a valid statement would get a lock timeout and then at the time of processing the exception the statement would have become invalid. One reason for a statement becoming invalid would be for instance if a table had been altered, for example if we had a prepared statement "SELECT * FROM TAB" and then a column was added or dropped. In that scenario the statement would be recompiled long before the lock timeout occurred. If the problem statement is the one with the lock timeout it would seem to be: SELECT DISTINCT t3.ID, t3.OPLOCK FROM PKG_9145E10G.PLPM_PAIR t3, PKG_9145E10G.PLPM_CHASSIS_9145E10G_USER t2, CORE_V1.DEVICE_ENTITY_USER t1, CORE_V1.DEVICE_ENTITY t0 WHERE ((t0.ID = CAST (? AS INTEGER )) AND (((t2.PLPM_PAIR_ID = t3.ID) AND ((t2.ID = t1.ID) AND (t1.DTYPE = 'PlpmChassis9145E10GUser'))) AND (t0.ID = t1.DEVICE_ENTITY_ID))) with 1 parameters begin parameter #1: 249360 :end parameter ERROR 40XL2: A lock could not be obtained within the time requested. The lockTable dump is: Thu Dec 22 07:03:07 PST 2011 I am not sure how that could have gone invalid between the initial execution and the lock time out. > Derby threads hanging when using ClientXADataSource and a deadlock or lock > timeout occurs > ----------------------------------------------------------------------------------------- > > Key: DERBY-5552 > URL: https://issues.apache.org/jira/browse/DERBY-5552 > Project: Derby > Issue Type: Bug > Components: Network Server > Affects Versions: 10.8.1.2 > Environment: Solaris 10, Glassfish V2.1.1, > Reporter: Brett Bergquist > Priority: Blocker > Attachments: appserverstack.txt, client.tar.Z, derby.log, > derbystackatshutdown.txt, execute.patch, transactionsleft.txt > > > The issue arrives when multiple XA transactions are done in parallel and > there is either a lock timeout or a lock deadlock detected. When this > happens the connection is leaked in the Glassfish connection pool and the > client thread hangs in > "org.apache.derby.client.netReply.fill(Reply.java:172)". > Shutting down the app server fails because the thread has a lock in > "org.apache.derby.client.net.NetConnection40" and another task is calling > "org.apache.derby.client.ClientPooledConnection.close(ClientPooledConnection.java:214)" > which is waiting for the lock. > Killing the appsever using "kill" and then attempting to shutdown Derby > network server causes the Network Server to hang. One of the threads hangs > waiting for a lock at > "org.apache.derby.impl.drda.NeworkServerControlImpl.removeFromSessionTable(NetworkServerControlImpl.java:1525)" > and the "main" thread has this locked at > "org.apache.derby.impl.drda.NetworkServerControlImpl.executeWork(NetworkServerControlImpl.java:2242)" > and it itself is waiting for a lock which belongs to a thread that is stuck > at > "org.apache.derby.impl.services.locks.ActiveLock.waitForGrant(ActiveLock.java:118) > which is in the TIMED_WAITING state. > Only by killing the Network Server using "kill" is possible at this point. > There are transactions left even though all clients have been removed. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira