Restore Volume Group

[Carlos Munoz Lopez]

Hi guys, I'm having problems with a volume group. This VG is created in a RAID 5 using four disks, two of them were damaged because of an electrical shutdown. The RAID is created in a IBM FastT Storage. IBM gave us some support in restoring the status of the RAID by replacing one of the two damaged disk, by doing this the RAID status returned to optimal and the remaining damaged disk was marked as clean. At this point I don't know if the data was completely lost or it can be restored in some way. When I try to activate the VG using "vgchange -a y vg_name" I get the following message:

vgchange: Warning: Couldn't attach to the volume group physical volume "/dev/dsk/c15t0d0":
Cross-device link
vgchange: Warning: couldn't query physical volume "/dev/dsk/c15t0d0":
The specified path does not correspond to physical volume attached to
this volume group
vgchange: Warning: couldn't query all of the physical volumes.
vgchange: Couldn't activate volume group "vg_des2":
Quorum not present, or some physical volume(s) are missing.

How can I fix this problem? Is there any way to restore the VG configuration? Do I have to recreate the entire volume group? I would like some help on this. Thanks.

[Mel Burslan]

Similar thing happened to me using HPs own raid array and the result was to cross fingers and pray that you have a really good and recent backup. Because the path you are on right now, will most probably lead to re creation of the whole volume group and logical volumes from the scratch and restoring data from the backup.

Once you lost the redundancy factor number of disk drives out of your raid array, it is basically a losing game. Cutting your losses and starting to rebuild the structure while you are time-wise ahead may be the best you can do.

[KapilRaj]

RAID-5 Protects from 1 disk failure at a time not 2 ..

Hope u have good backups. Time to go for RAID-1 if u can buy few more disks.

Kaps

[Victor Fridyev]

Hi,

After a good pray you can try to do the following:
ll /dev/vgdes2/group
and remember it's minor, let it be 0x030000
strings /etc/lvmtab
and remember all the disk, belonnging to vg_des2
vgexport vg_des2
mkdir /dev/vgdes2
mknod /dev/vgdes2/group c 64 0x030000
vgimport /dev/vgdes2 /dev/dsk/disk1 /dev/dsk/disk2 etc..

You can change the list of disks in a hope to restore data on the vg.

Good luck

[Carlos Munoz Lopez]

This is my /etv/lvmtab:

/dev/vg00
/dev/dsk/c1t2d0
/dev/dsk/c1t0d0
/dev/vg02
/dev/dsk/c2t0d0
/dev/dsk/c2t2d0
/dev/vg_prod2
/dev/dsk/c15t0d1
/dev/vg_des2
/dev/dsk/c15t0d0

When I execute "pvdisplay -v /dev/dsk/c15t0d0" I get the following:

pvdisplay: Warning: couldn't query physical volume "/dev/dsk/c15t0d0":
The specified path does not correspond to physical volume attached to
this volume group
pvdisplay: Warning: couldn't query all of the physical volumes.
pvdisplay: Couldn't retrieve the names of the physical volumes
belonging to volume group "/dev/vg_des2".
pvdisplay: Cannot display physical volume "/dev/dsk/c15t0d0".

How can I do to reuse the PV?

[KapilRaj]

It looks like the PVRA is destroyed, which is logical as you had a RAID-5 + 2 disks crash.

You will need to recreate the PV or do a vgcfg restore onto the device. I hope ioscan is showing the device. see the man pages for vgcfgrestore and give it a try

Kaps

[Leon Allen]

Been there - done that...

Here is what I do to recover....

mv /etc/lvmtab /etc/lvmtab.old
vgscan
vgchange -a y /dev/vg0n
fsck -F vxfs /dev/vg0n/lvolnn
mount .....

Explaination:
lvmtab does not reflect new config - rename it.
Create a new lvmtab - it will scan the new config.
THEN active the volume group
fsck - make sure all is well - you may have to do a full check
Finally, mount your disk -

Cheers

Leon