[jakajaksic]

By the way, if there are some gereral no-no's in our code, feel free to correct us (Manik or anyone else reading this). This is our first application in java and may be less than optimal.

[sebersole]

FYI, after some discussions with Manik we believe we have identified a source (hopefully *the* source) of these issues...

http://opensource.atlassian.com/projects/hibernate/browse/HHH-1796

Try the latest code from Hibernate SVN, or wait till we release 3.2.0.ga in a few days and try that.

[bigm25]

I did follow this thread 'cause we're facing the same problems.

I did c/o hibernate trunk this morning, but i still get (With jboss-cache 1.3.0.SP2) these strange don't kno
w how to handle messages (when using OPTIMISTIC cache locking)

2006-05-31 09:23:44,254 INFO [org.jboss.cache.interceptors.OptimisticNodeInterceptor] CRUD Method _put(GlobalTransaction:<null>:311, /de/eyetea/hfax/entity/auktion/Car2Auktion/de.eyetea.hfax.entity.auktion.Car2Auktion#80339, item, CacheEntry(de.eyetea.hfax.entity.auktion.Car2Auktion)[EYETEA,null,null,2006-05-30 21:48:08.132,0,null,BID,null,null,80827,1000,null,null,false,80338,56384,80339,80339], true, 0) called - don't know how to handle!

With PESIMISTIC cache locking we still get Exceptions like

2006-05-31 09:39:06,580 ERROR [org.jboss.cache.lock.IdentityLock] write lock for /de/eyetea/hfax/entity/auktion/Gebot/de.eyetea.hfax.entity.auktion.Gebot#80821 could not be acquired after 0 ms. Locks: Read lock owners: {}
Write lock owner: GlobalTransaction:<null>:175
 (caller=GlobalTransaction:<null>:168, lock info: write owner=GlobalTransaction:<null>:175 (org.jboss.cache.lock.LockStrategyReadCommitted@3729c5))

Note that the we have

<attribute name="LockAcquisitionTimeout">5000</attribute>

in jboss-service.xml, but the exceptions says "0 ms".

We're using jboss cache as a "standalone" .sar / mbean, could this be the reason for this behaviour?

[manik]

Re: optimistic locking, sounds like you're doing a putFailFast(). Make sure you're using the OptimisticTreeCacheProvider and not the TreeCacheProvider that comes with Hibernate 3.2.

Re: pess. locking, those errors should not be considered as errors. They should be INFO or DEBUG level messages in the current use-case. This has been fixed in the 1.4.0 branch.

Cheers,
Manik