So if you could, I can post here the latest changes I have made to the
o.h.c.OptimisticTreeCache class for you to verify. It should fix the vast majority of the
issue you were seeing. There are still some "transactionality mismatch" (for
lack of a better term) issues Hibernate and JBoss Cache need to work through. For
Hibernate users this would effect only two areas I am aware of:
1) query result caching. This will not work at all with JBoss Cache in optimistic mode.
The underlying issue being that Hibernate tries to write some info needed for query result
caching into the underlying cache during after-transaction processing which JBoss Cache
barfs at. Actually this *will* work as long as the queries are not requested less than .6
seconds after updates to any of the "table spaces" that affect that result...
2) an edge case where CMT is utilized relying on Hibernate's flush-on-completion where
Hibernate made no calls into the TreeCache prior to the commit/flush and the flush
actually needs to write state changes...
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3957607#...
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...