2 Weblogic cluster in the same network

Question

we have 2 servers (dev/int), one of this has 3 Weblogic clusters with one managed server and different mutlicast addresses.

server 1 has the multicast addresses 239.192.3.7 and 239.192.3.8 and 239.192.3.9 for ione of his cluster
server 2 has the multicast addresses 239.192.4.7 and 239.192.4.8 and 239.192.4.9 for ione of his cluster

The admin and managed servers starts without errors. The managed servers connect to his clusters and it looks well. Both server are in the same network (a.b.c.d/24) and connected to the same switch.

If I deploy a service to one of this cluster e.g. 239.192.3.7, I received an timeout. With netstat I see connection to the other cluster from server 2 (239.192.4.7). In the log of this cluster (server2), I saw the try of service deployment from server1. So after I stopped the clusters of server2, I can deploy the service on server1 without any trouble.

Where is the problem? To much multicast addresses in one network?

So maybe anybody can help me, thanks!

EDIT (10.05.2013):
Some days ago I take a new install of this server with his 3 cluster configurations. Maybe I had a mistake in my configuration.

In this new installation I had the same error, now I looked again on server2 with netstat -la --numeric-ports and see two connections to the another server1. It look like this:

tcp 0 0 server2:8088 server1:57963 ESTABLISHED
tcp 2 0 server2:7890 server1:34010 ESTABLISHED

Each connection will be created by a start from a managed server. But only this two connections with every same source ports.

Have you tried changing it to unicast? Might be worth it to see what happens. — Display Name is missing, Apr 30 '13 at 15:10
No, because in the future we need to create a cluster with more than one physical server. — luk3, May 10 '13 at 09:48

score 3 · Accepted Answer · answered May 13 '13 at 11:44

I solved the problem by defined a special coherence.clusteraddress in the default start environment.

I add following lines by updating the EXTRA_JAVA_PROPERTIES variable in the setDomainEnv.sh script inside of the bin directory of the soa and osb domain. For server1 and server2 I used different clusteraddresses.

-Dtangosol.coherence.clusteraddress=239.192.4.7 -Dtangosol.coherence.clusterport=31323 -Dtangosol.coherence.ttl=0 -Dtangosol.coherence.log=jdk
-Dtangosol.coherence.clusteraddress=239.192.4.8 -Dtangosol.coherence.clusterport=31324 -Dtangosol.coherence.ttl=0 -Dtangosol.coherence.log=jdk

2 Weblogic cluster in the same network

1 Answers1