Could you post the "discarded" message? Also post the "Starting Server" output and maybe attach the boot/server logfiles of your nodes.
It looks like that you configure the two nodes wrong or there is another node seen (by accident) and configured for a different cluster.
Looks to me that is really the cause -- 2 clusters on the same network.
A simple test for that is to disconnect your test machine from the network whatsoever and try to create the cluster. The messages will disappear.
Thanks for the information. I think this is correct as the one server with the warning does have another j2ee server running on it (it's a WLS no Jboss). We must have selected the same multicast address (see "set mcAddress = 184.108.40.206" above). I think changing the mcAddress in the script above for both Jboss instances will resolve the problem. I'll try this later today.
All, thanks for your input on this question. I truly appreciated it. I worked on diagnosing the problem last week by turning off the Jboss server on the second server. This stopped the "discarded message from non-member" warning leading me to believe the problem was not associated with another j2ee server on the same multicast address. It turned out that the problem was that I tried to implement a virtual ip address by adding cfg-eth0 (that's eth0 'colon' 0) to /etc/sysconfig. By removing this config file, I was able to remove the warning messages on the other jboss server.