I am running JBoss 3.2.1 on Linux with javagroups synchronous cache invalidation in my cluster. After my instances have been up for several days, I see one of the members in the cluster spit out:
-------------------------------------------------------
GMS: address is www7:32784
-------------------------------------------------------
The port in the address is different than the initial port that the server was using on startup. This causes a chain reaction since the other members in the cluster are not aware of the port change and the synchronous cache invalidation calls get timeouts. Unfortunately, this grinds my site to a halt since every page load involves some entity bean interaction which tries to make the synchronous cache invalidation call.
Does anyone know why this happens and how to prevent it? Or at least a workaround to keep my site from crashing.
Thanks.
