Occurences of "ISPN000069: Cache named [default-host/clusterbench] exists but isn't in a state to handle remote invocations" on shutdown

AFAIR related to the order of starting caches and channels.

Lets track here if we remove the logging and/or which build to get the fix to.