The 1.2.0.Beta1 release builds in top of the clustering functionality already existing in 1.2.0.Alpha2 by adding the last major piece of functionality, high availability with transparent failover.

The release ships with an example (which also doubles as smoke test) that demonstrates the transparent failover capabilities. A vanilla JMS client creates a connection to the JMS provider and sends message over that connection. Separately, we kill the cluster node managing the connection (using System.halt(), to simulate an abrupt VM failure), but the event goes unnoticed by the client, which continues to send and receive messages over the same connection.

Obviously, the example requires a cluster of at least two nodes in order to work.

Other additions: fully tested HTTP transport support, improved testing framework to support automated clustering testing, and overall strengthening of the clustering code.

Referenced by: