-
15. Re: javax.transaction.HeuristicMixedException whlie starting Infinispan
tcoates1362 Aug 12, 2010 6:06 PM (in response to manik)I'm starting to think that I'm loosing my mind... since updating to the latest source I can't reproduce this particular issue either. Also, the test that I claimed wasn't working yesterday is now passing. I don't have it in me to check out a revision from yesterday and make sure I'm not crazy - I guess I'll just hope for the best. =)
I was able to create a test demonstrates the issue I'm seeing with the JBossStandaloneJTAManagerLookup. I tried using the system property you suggested to change the default transaction manager but it didn't seem to be working. Instead I just create a new Configuration object and set a couple parameters (including transaction manager lookup class) and I get the exception. The test class is attached.
Thanks for your help getting up and running with the test framework. Now that I'm not seeing as many problem as I was originally, I may try running the latest build in my AS and see if I can make some progress there.
-
16. Re: javax.transaction.HeuristicMixedException whlie starting Infinispan
tcoates1362 Aug 12, 2010 7:04 PM (in response to tcoates1362)Ok, I think I found a repro for my NPE I reported earlier. It seems to happen with a cache loader when doing a put (update) for a key that already exists after a stop/start. Attached is a test that exposes it. My earlier proposed fix I think has some bad side effects though (lock timeouts) so it's likely not a good solution. At least I have a test though, right?
-
TreeCacheWithLoaderDupTest.java.zip 771 bytes
-
-
17. Re: javax.transaction.HeuristicMixedException whlie starting Infinispan
manik Aug 13, 2010 10:14 AM (in response to tcoates1362)I've created a JIRA to track this - ISPN-600. I've also added a more genericised version of the test into our testsuite:
http://viewvc.jboss.org/cgi-bin/viewvc.cgi/infinispan?view=revision&revision=2213
-
18. Re: javax.transaction.HeuristicMixedException whlie starting Infinispan
manik Aug 13, 2010 10:24 AM (in response to tcoates1362)1 of 1 people found this helpfulThanks for this, I've added a test method following the pattern you demonstrated in your test, to an existing test which now exhibits the problem.
I've also created a JIRA to track this - ISPN-601.
-
19. Re: javax.transaction.HeuristicMixedException whlie starting Infinispan
tcoates1362 Aug 13, 2010 12:12 PM (in response to manik)Thanks guys. Glad I could finally provide some usable tests.
My upgrade project is currently blocked by the following 3 issues:
I have a watch on all of them so once they're resolved, I'll take another swing at the upgrade. I really appreciate the help and the responsiveness on the forums.
-Tom
-
20. Re: javax.transaction.HeuristicMixedException whlie starting Infinispan
galder.zamarreno Aug 16, 2010 5:16 AM (in response to tcoates1362)https://jira.jboss.org/browse/ISPN-600 is fixed now. Build 4.1.x branch to get the update or wait for next snapshot to be uploaded.
-
21. Re: javax.transaction.HeuristicMixedException whlie starting Infinispan
mircea.markus Aug 16, 2010 11:42 AM (in response to tcoates1362)This only fails when using a ManagedConnectionFactory. If I swap out a PooledConnectionFactory it works fine. I presume the managed version fails because it returns the same connection that is part of the active transaction for that thread and you can't perform new operations on that connection once the transaction is already in prepare or commit. This feels like a bit of a design flaw to me... I didn't see an obvious fix.
Jdbc cache store do not work with JTA and managed connections, see my last comment on https://jira.jboss.org/browse/ISPN-586. In order to pass this limitation we need to change API, so this can only be approached in 5.x( https://jira.jboss.org/browse/ISPN-604).