How to start jboss 5.0 in cluster on Linux

[vpatchva]

I copied all profile into node1 and node2 on MachineA and same on MachineB. Now Im trying to start node1 from A and node1 on  B as cluster.

When I used the command ./run.sh -c node1 -g jboss1 -u 239.255.100.100 -b 192.x.x.x. -Djboss.messaing.ServerPeerID=1. getting the port 8083 is already in use. Did I missed anything here?

Can someone help me in setting up a cluster of two nodes in different machines?

[dandare100]

You have 2 choices :

Either bind each server instance to a different ip address (Assuming 2 network cards)

or specify a different port set for the second server with

-Djboss.service.binding.set=ports-02

See http://community.jboss.org/wiki/ConfigurePorts

also

http://community.jboss.org/wiki/AS5ServiceBindingManager

[mschwery]

I think I may be having a related problem with multicasting.

If I run ./run.sh -c node2 -u 199.134.220.30 -Djboss.service.binding.set=ports-01 -Djboss.messaging.ServerPeerID=1 I have port 8083 in use.

But if I run ./run.sh -c node2 -Djboss.service.binding.set-ports-01 -Djboss.messaging.ServerPeerID=1 I don't get the port conflict.

The only differance between the two ways of starting it up is the use of the multicast port.

I think that is where the problem is but I'm not sure what the next step should be.

[dandare100]

I think the multicast port is included in the port set and you are overiding it with the -u parameter.

ie You are forcing the 2 instances on the box to use the same multicast address.

Dont specify the -u multicast address and your cluster should be fine providing the 2 nodes use the same port set.

Is there a reason you are specifying the multicast address ?

PS : Are you sure the port conflict is on port 8083 ? Because this has nothing to do with the multicast address.

If it is definitely on port 8083 then do a netstat -a before starting the server to see what is running on that port.

[dinsceg]

Hi,

    i think problem in multicasting.. better you try in without multicasting(-u option)...

Run both the server instances by specifying the below command.

For Node1:-
# ./run.sh -c node1 -g DefaultPartition  -b <IP>

For Node2:-
# ./run.sh -c node2 -g DefaultPartition  -b <IP>

[vpatchva]

Dinesh, thanks for your response.

Im getting the following error message when I run the  command ./run.sh -c node1 -g DefaultPartition -b 199.*.*.*

10:34 Using RMI servercodebase: http://199.*.*.*:8083/

10:34 Error [AbstractKernelController] Error installing to Start name=jboss;service=WebServise state=Create mode=Manual requiredState=Installed

java.lang.Exception: port 8083 already in use.

But if I use the command ./run.sh -c node1 -g Default Partition, it is working and i was able to start the Jboss node1.

It is using RMI servercodebase: http://127.0.0.1:8083 and everything will start fine.

[saradakrishna]

Hi,

I had faced also the same issue as that of you.

But didn't get direct solution to it.I just restarted the system & it worked.

Probably the error "8083 port no already in use" is appearing as you would have started the JBoss AS instance which didn't closed it properly or

that particular port is already in use with some other service.

Moreover, in ./run.sh ,

    -g option is for partition name or your cluster name that you have given to your cluster to differentiate from other cluster present in your network.

"DefaultPartition" is the default partition name given by jboss.You can change it.

Hopefully it will work for you.

Please let me know if your problem gets solved.