I've added a test to illustrate the conditions under which the race condition may arise, though it involves a lot of reflection use. In a production environment, this arises once every 3-4 hours with an app that is constantly rollbacking.
https://github.com/apache/activemq/pull/256 Could anyone verify if this is a feasible way to address the out of order retry? -- Sent from: http://activemq.2283324.n4.nabble.com/ActiveMQ-Dev-f2368404.html
