On startup, verify that all nodes in the cluster have identical cache configuration, to avoid issues like the one presented in the user forum post.

A hash (that's plattform/jvm independent) could be calculated on startup and coordinator could verify that all cache configurations have the same hash all across the cluster.