APA - failover, failback question

[Susan Harris]

I have an APA question. I have a server, with 2 nics which are setup with APA for failover. We lost one of our switches last week. The server connectivity was fine, due to the failover. I just noticed that now I'm still only using 1 nic, the other has not 'failed back' yet. But the switch is up and supposedly configured fine. Is there a way to 'fail back' the second card, online ?
Thanks,

Sue

[Bernhard Mueller]

Sue,

there was are recent thread :

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=264543

where Geoff told us it works like a miracle...

Regards,
Bernhard

[Alzhy]

If you have a "manual" configuration -- ie. you're using APA for failover/redundancy - then once your switch is up, you need not do anything as there is no concept of primary lan and secondary lan in APA configs.. Your system will continue to use whichever NIC it survived on and use the repaired link as failover.

HTH.

[Alzhy]

And if your usage of APA is true aggregation/trunking (ie. bandwidth + redundancy) -- then when the switch is restored to normal health, your bandwidth will resume its aggregate link speed -- ie. if you've 2 x 1 GB NIC's aggregated - then on resumption of the failed NIC connection - you should have your bandwidth restored to 2 GB..

[James Randall]

I've configured APA at some of my client locations. And as part of my testing...I test this very thing.

Yank one of the NIC connections....and see that the system stays live. Replace the connection - And verify that the NIC comes back into the pool.

#1 - I'd check the physical connections. (Most HP cards have a link light - Verify it is lit. Also they usually have some sort of traffic indicator. If its flashing....then at least DATA is physically getting to the card.)

#2 - Check your system configurations. (APA, etc).

James

[Jim Keeble]

It sounds as though you are using the Lan Monitor feature as opposed to a load balancing aggregate.

If so, the config file is usually /etc/lanmon/lanconfig.ascii. The primary interface should have a higher priority assigned. If the priority (the last column in the config file) is the same, then the failover group won't switch back until a failure occurs on the secondary interface.

If you find that the primary interface does have a higher priority, then check the card's "operational status" with "lanadmin -g " , where , for example, ppa = 3 for lan3. If operational status=down, check the card, the cable, and the switch port for electrical problem of some sort.