Re: NNM 9.20 Application Failover (805 Views)
Reply
Valued Contributor
sumit_kumar
Posts: 219
Registered: ‎07-18-2012
Message 1 of 8 (889 Views)
Accepted Solution

NNM 9.20 Application Failover

Hi Expert,

 

Currently both NNM Server are installed with same Oracle Database. When I start the services only on Primary Server then it's working fine. And if I start the service only on Secondary Server then it's is also working fine.

 

Now I have followed the HP Doc to make the Application failover, all steps are successfully done except the output of nnmcluster . I am getting the only individual cluster member server name.

 

Please help on this

 

Regards,

Sumit Kumar

Regards,
Sumit Kumar
Honored Contributor
dieter boschung
Posts: 236
Registered: ‎12-18-2008
Message 2 of 8 (885 Views)

Re: NNM 9.20 Application Failover

Hi Sumit

 

Did you copy one cluster.keystore file to the other system?

 

Do you see the same cluster name and CRC value on both systems by running the command: nnmcluster -display

 

If the CRC value or the cluster name are not identical, then they cannot see each other.

 

=============================== Current cluster state ============
State ID: 000000004000000022
Date/Time: 26 Mrz 2013 - 08:39:39 (GMT+0100)
Cluster name: myCluster (key CRC:3'041'891'359)

 

BR

Dieter

HP Support
If you find that this or any post resolves your issue, please be sure to mark it as an accepted solution.
Valued Contributor
sumit_kumar
Posts: 219
Registered: ‎07-18-2012
Message 3 of 8 (880 Views)

Re: NNM 9.20 Application Failover

Hi Dieter ,

 

I have follow the HP Document step by step.

 

Yes I have copy one cluster.keystore file to another system and also genrate the self sign cretificate.i have also changed the nms-cluster file on both the system with the following parameter:

 

com.hp.ov.nms.cluster.name =ClusterNNM

com.hp.ov.nms.cluster.member.hostnames = rs11i0702ap152.sbigen.in,rs19ic602ap153.sbigen.in

 

The Output of nnmcluster command is attached of both server.

Regards,
Sumit Kumar
Honored Contributor
dieter boschung
Posts: 236
Registered: ‎12-18-2008
Message 4 of 8 (867 Views)

Re: NNM 9.20 Application Failover

Hi Sumit

 

Is there a firewall between the systems or is the Windows firewall active?

 

Ports tcp:7800-7804 and tcp:7810 must be reachable on either side for the nnmcluster to see each other.

 

Another point to mention is that the hostnames in the nms-cluster.properties must be the real physical hostnames returned by the hostname command, not a secondary hostname pointing to another IP-Address.

 

It is worth checking the jgroups.log and nnmcluster-daemon.0.0.log files under %NNM_LOG%\nnm

 

And of course, both systems must be running the same OS and same NNMi release and same patch level.

 

BR

Dieter

HP Support
If you find that this or any post resolves your issue, please be sure to mark it as an accepted solution.
Valued Contributor
sumit_kumar
Posts: 219
Registered: ‎07-18-2012
Message 5 of 8 (863 Views)

Re: NNM 9.20 Application Failover

Hi Dieter,

 

Thanks for reply !

 

both server is in same subnet, So firewall does not come into the picture.Both system are windows system and patch level is same of NNM 9.20.

 

I have checked the logs also, but not get the userfull info ! Attached the logs

 

Pelase help me on this.

 

Regards,

Sumit Kumar

 

Regards,
Sumit Kumar
Honored Contributor
dieter boschung
Posts: 236
Registered: ‎12-18-2008
Message 6 of 8 (859 Views)

Re: NNM 9.20 Application Failover

Sumit

 

This looks like a communication problem.

 

I think it would be best to have wireshark installed on both systems, then setup wireshark to filter on tcp ports 7800, 7801,7802,7803,7804 and 7810 on both systems. (I use the filter 'tcp portrange 7800-7810')

 

Next, start nnmcluster on one of the nodes only and check wireshark.

 

On the system where nnmcluster runs you should see and outgoing request to <remote addres>:7800 -> 7804 ; the local nnmcluster tries to connect to that range.

 

Then go to the other system and check if you see these requests come in, if they do not arrive, then there is a device blocking them. If they arrive, then start a nnmcluster process and you should see that one sending the same requests which should be seen on the first system too.

 

You may also want to consider opening a case if this doesn't help to find the problem as more investigations need to be done.

 

Kind regards

Dieter

HP Support
If you find that this or any post resolves your issue, please be sure to mark it as an accepted solution.
Valued Contributor
sumit_kumar
Posts: 219
Registered: ‎07-18-2012
Message 7 of 8 (855 Views)

Re: NNM 9.20 Application Failover

Hi,

 

There is no communication problem because there is no firewall between both of servers.

 

HP is not able to find out the problem.

 

Regards,

Sumit kumar

Regards,
Sumit Kumar
Valued Contributor
sumit_kumar
Posts: 219
Registered: ‎07-18-2012
Message 8 of 8 (805 Views)

Re: NNM 9.20 Application Failover

So finally this problem has been resolved,

 

I have changed the LAN card number in nms-cluster-local.properties file.

 

 

Regards,

Sumit Kumar

Regards,
Sumit Kumar
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation.