On Tue, Nov 13, 2001 at 03:34:28PM +0100,
Marcus Hast wrote:
> This is on a LVM volume with 220G of data.
Hmmm... 220, divided by 3... you wouldn't happen to be using IBM
75G drives, would you?
> So I have a few questions:
> Is there any way of getting the data on the other disks back?
> From what I've seen of the logs it's hdg that's bad.
You have reason for a very, very small amount of hope.
Pull the bad drive out immediately. Whatever you do, until you
have the data safely and completely off the devices, don't do
anything that will cause a write to any of the drives - either
log replays when mounting the XFS filesystem, or whatever LVM
might do when initializing.
Try as hard as you can to clone the failing/failed drive to a
new drive; "dd" is your friend. (You might find "conv=noerror"
useful as an option to dd.)
If you succeed in cloning the drive, put the clone into the
original drive's original location. Make sure that LVM won't
write anything to the disks when it initializes the volume.
Boot up with an XFS install CDROM into text install mode - if
you can access LVM volumes from there, anyway. Use the Alt-F2
console to bring up the volume, then to mount the filesystem
(ro,norecovery). If you can see your data, congratulations;
pull it off onto a tape or another server immediately. (You
should be able to bring up your NIC manually with ifconfig if
you need to.)
Andrew Klaassen
|