JBoss instances don't find each other

[elko]

I have following problem: wenn I start my jboss-instances, they see each other, but don't recognize each other to be members of the same cluster.
Any suggestions? thnx

ServerLog (host: pavullo, other host: sacuzzo):
...
-------------------------------------------------------
GMS: address is 172.30.5.82:32828
-------------------------------------------------------
20:55:33,216 INFO [TreeCache] TreeCache local address is 172.30.5.82:32828
20:55:37,678 INFO [TreeCache] viewAccepted(): [sacuzzo:32808|1] [sacuzzo:32808, 172.30.5.82:32828]
20:55:47,778 INFO [TreeCache] received the state (size=1024 bytes)
20:55:48,091 INFO [TreeCache] state was retrieved successfully (in 14874 milliseconds)
20:55:48,092 INFO [TreeCache] parseConfig(): PojoCacheConfig is empty
20:55:50,668 INFO [DefaultPartition] Initializing
20:55:56,212 INFO [STDOUT]
-------------------------------------------------------
GMS: address is 172.30.5.82:32831 (additional data: 16 bytes)
-------------------------------------------------------
20:55:58,252 INFO [DefaultPartition] Number of cluster members: 1
20:55:58,252 INFO [DefaultPartition] Other members: 0
20:55:58,252 INFO [DefaultPartition] Fetching state (will wait for 30000 milliseconds):
20:55:58,261 INFO [DefaultPartition] New cluster view for partition DefaultPartition (id: 0, delta: 0) : [172.30.5.82:1099]
20:55:58,274 INFO [DefaultPartition] State could not be retrieved (we are the first member in group)
20:55:58,412 INFO [DefaultPartition] I am (172.30.5.82:1099) received membershipChanged event:
20:55:58,412 INFO [DefaultPartition] Dead members: 0 ([])
20:55:58,413 INFO [DefaultPartition] New Members : 0 ([])
20:55:58,413 INFO [DefaultPartition] All Members : 1 ([172.30.5.82:1099])
20:55:58,961 INFO [HANamingService] Started ha-jndi bootstrap jnpPort=1100, backlog=50, bindAddress=/172.30.5.82
20:55:59,449 INFO [DetachedHANamingService$AutomaticDiscovery] Listening on /172.30.5.82:1102, group=230.0.0.4, HA-JNDI address=172.30.5.82:1100

...

[elko]

Addition:

what i have done:

1) in cluster-service.xml: (mbean HANamingService) BindAddress from
${jboss.bind.adress} to ip-address changed

2) route add -net 224.0.0.0 netmask 240.0.0.0 dev eth0

3) started with:
run.sh -c al -b [ip-addr]

[canty28]

Elko,

I have found couple of things in the cluster. JBOSS by default it uses UDP protocol for clustering and UDP protocol has a restriction on send and receive. I initially had problems because the war file which i use is around 100MB and i had problems sending it over to the other clusters. I used TCP protocol and it was click..boom ...way too fast. As part of clustering, I did 2 things modified the following 2 files

1. tc5-cluster-service.xml
2. cluster-service.xml

I commented out the UDP protocol and uncommented the TCP protocol information, gave the ip's of the boxes which i needed in the cluster and started them back and each node was able to recognize the other.

[elko]

Hi canty28,

thank you for your answer, I'll try it.

best reguards, elko

[elko]

Hi canty28,

thank you for your help, It has worked!!!!

Regards, elko