cmscancl network failure detection

[Vincent Farrugia]

Hello,

We have a 2-node serviceguard cluster version 11.14 running under HPUX 11.00 version. We upgraded 1 node for a rolling upgrade to HPUX 11i and version 11.16 successfully.

We tried to enter the new 11i node in the cluster. Unfortunately, the network probing failed, but the node still entered the cluster successfully, and confirmed so through syslog. I can't find anything awry with the network though.

I attached the cmscancl output of this cluster, clearly showing "network failure detection" under "Comparing binary configuration". Can you try to find out what is actually wrong please? I watched all linkloops and all are ok.

Regards,
Vince

[Sridhar Bhaskarla]

Hi Vince,

If you are applying the configuration again, then I suggest you take out the IP addresses 0.0.0.0 from the interfaces on both the systems.

-Sri

[Vincent Farrugia]

Hello,

Thanks for your answer. lan2, lan5 and lan6 are used as spare network interfaces in case either lan0 (primary and heartbeat network) or lan1 (heartbeat network) fail. Hence I guess I cannot remove them?

Vince

[Sridhar Bhaskarla · Accepted solution]

Hi (again),

If you already have the cluster up and running, then not. But if you are trying to apply (not re-apply) the configuration, you will need too. Otherwise, cmapply/checkconf will return errors.

-Sri

[Vincent Farrugia]

The cluster is up and running now, so I'll leave it as it is. We will see what will happen when we upgrade the other node to 11i and version 11.16 as well. It could be that the errors are gone when it is fully upgraded.

Many thanks and cheers,
Vince