Replication queues cannot be used with REPL_SYNC - only REPL_ASYNC. I haven't heard of this regression, would like to see your test. :-)
Cleaned up, testng'd and uploaded.
The test includes a config file, it is documented and... fails. You should be able to drop this straight into the 2.1.1.GA tag and run.
Thanks for this - I have created a JIRA for this JBCACHE-1348. See comments on JIRA.
Also, I have included a (slightly modified) version of your test in our test suite. :-)
Our real scenario is this: We have two caches, one for hard binary data (dataCache) and one for meta data (metaCache) consisting mainly of strings. The server processing updates the dataCache within a transaction (configurable with JTA manager) which default to a jboss dummy transaction. But as a part of that processing the metaCache is also updated. The metaCache uses a replication queue and asynchronous replication.
Can I work around this by using another transaction manager? (Thinking that the metaCache is only involved in the transaction because it shares the same, singleton, instance of cache manager).
You could suspend the tx for the call to the metaCache, and then resume it afterwards...
This is not a TM specific issue so changing TMs won't really help.