After vgcfgrestore of primary boot disk still can't activate vg00?

[Steve_615]

The pri. boot disk failed on HPUX 11.i. VG00 is mirrored. Had to reboot system, interupt and boot off the alt. disk. Later, replaced the bad disk. Brought sys. to maint. mode, deactivated VG00 and ran vgcfgrestore for vg00 on to device file for failed disk. Took a very long time (about an hour) but said it completed o.k. When I then tried to act. VG00 I had to override quorum, but then I received the following error:

Can't read VGDA1 header and trailer for /8/4.5.0 (this is the orig. failed disk)
and VG00 won't activate.

Why would this be happening?

To get the system back up I just took the drive out and booted from the alt. I'm now up without mirroring.

Thanks in advance for any of your help.

[Devender Khatana]

Hi,

It seems you have not done a pvcreate on the replaced disk. Note that it has to be done with -B option for boot disks.

HTH,
Devender

[Mohanasundaram_1]

Hi Steve,

Chances are the new disk is faulty. It does not take 1 hour for vgcfgrestore.

check the diskinfo on this disk. It may be showing 0 bytes. If so, replace it.

Otherwise your steps are ok. Did you get any messages of "crosslinked device files" ?

did you use hpux -lm to go to maintenance mode?

With regards,
Mohan.

[Steve_615]

I don't believe there is any reason to pvcreate the replacement disk. This is recovering a failed disk that was already a mirror. Not creating an initial mirrored root disk.

I will have to check the new disk with a pvdisplay as suggested. I will order a new disk today and try again.

Thanks for your suggestions..........

[Trond Haugen]

One hour is a very long time for a vgcfgrestore. It should be done in a metter of seconds.
You should not have tp pvcreate it. But do remember to run mkboot.
After successfully activating it you will have to re-sync (vgsync vg00).
Ad as a roule always reboot after your work is done in maintanance mode.

Regards,
Trond

[Devender Khatana]

Hi Steve,

Apart from replacing disk there are other ways also to check functionality. One is to use pvcreate on it and see if it does give some error. The other can be copying null to whole media using dd command. Another may be using mediainit command.

But before that see the output of "diskinfo -v /dev/rdsk/cxtydz"

You have plenty of options with you. Allthough it seems your current resources can get a new spare very easily that is why you are going to check by replacing it straightsway without testing it.

HTH,
Devender