How to identify defekt disk when server doesn't boot?

[ErikGregg]

Hello,

we have a rp7410 server with 2 disks concatenated in a volume group. The server doesn't boot any more because one of the disk has failed. Fortunately the server has sent an email with an event notification a few days before, I know which disk has to be changed.

My question is: If there would be no email, how could I identify the failed disk if the server doesn't boot anymore? If everything I have is the service processor, is it possible to get a detailed information about the failure like in the event notification mail?

Cheers

[Torsten.]

Usually you have a mirrored boot disk, so you can still boot if one of them has failed. If you don't have a mirror and the boot disk is bad, you can boot a recovery shell to get the status of the disk.

[mvpel]

The boot section of the disk is created when the pvcreate command is run with the "-B" option - this sets aside some space at sector 0 for the LIF volume, which contains the "HPUX" boot file which the ISL (initial system loader) reads and executes while the system is booting.

I've found that it's probably not a bad idea to run a "lifls" on the boot disks (/dev/rdsk/c0t0d0) once in a while, to make sure that the LIF volume is still intact - it's a bad thing to find out that your OS mirror disk has a trashed LIF when the other mirror has already failed. The LIF volume can be recreated using the "mkboot" command if needed.