Hibernate makes legitimate calls into the cache during afterCompletion(). See http://www.jboss.com/index.html?module=bb&op=viewtopic&t=102202 and related JIRA http://jira.jboss.com/jira/browse/JBCACHE-982.
There's this bit in TxInterceptor.setTransactionInContext() that may be relevant: | ctx.setTransaction( tx );//tx == null || !isValid(tx) ? null : tx); // make sure the tx is valid, otherwise set as null - see JBCACHE-785 | if (ctx.getTransaction() == null) ctx.setGlobalTransaction(null); // nullify gtx as well Not sure how the commented out bit was meant to work or why it's commented, but if it were executed before the ctx.setTransaction( tx ); call, it would have the effect of treating Hibernate's afterCompletion() call as occuring outside the tx scope, which AFAICT is Hibernate's intent. Hmm... that might screw up the if test in invoke() that immediately follows this call. Not sure; out of my depth here. :-) View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4028397#4028397 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4028397 _______________________________________________ jboss-user mailing list jboss-user@lists.jboss.org https://lists.jboss.org/mailman/listinfo/jboss-user