No join standby event HACMP

[andy61]

Hi,
We are running HACMP 4.5 on AIX 5.1. We had a a report of a malfunction on the service address and it ran the swap_adapter event, moving the service (en1) onto the standby (en2) . It then reported fail standby event on the new standby (en1). All fair enough I guess. However, netstat/entstat show packets going in and out of the interface en1 with no errors, so it now looks ok, but we haven't had a join standby event. It is a production server, so I'm a bit reluctant to take down the interface to run diag. Is there anything else I can try or anywhere I can look that might help ?

 

[khalidaaa]

Try discovering HACMP-related Info.

smitty hacmp --> Extended Configuration --> Discover HACMP-related Information from Configured Nodes

I'm not sure if this will help detecting the en1 but it won't harm trying it any way.

Regards,
Khalid

 

[andy61]

Thanks for the post Khalidaaa. Unfortunately, we don't have that smitty option on our version. It certainly looks like it could do with a refresh of its information. I don't think this will happen with cl_verify and I don't want to risk stopping and restarting cluster services while keeping the resource running.

 

[ogniemi]

I think it sometimes happens.

I remember once IBM did a cluster for us and during tests, when they unplugged network cable from service ent, HACMP recognized it, did swap adapter and stby lost event occured. Next, when the cable back again, join stby event never appeared (!!!). IBM engeener was waiting ca. a half an hour but join stby never occured although the network was up on the ent quite quick after it was cabled again.

cldump showed that the stby ent state is up.


Feature of bug? - I don't know.

 

[andy61]

Again thanks, but it looks like cldump isn't available either. Which is a shame as it looks like just what I need. I guess we're paying the price of running software designed for the Babbage Differential Engine.