Skip to contentSkip to conversation

Seem to be in trouble

linux-lvm Nicholas Veeser 1 participant 1 message Feb 15, 2005

lore
Share & tools
lore
#0 lore raw

So I have been converting my Suse 9.1 fileserver to use lvm2. (2.6.4) Seems I have have run into trouble and not sure what to do to get out of it.

All of my physical volumes are Raid 1 mirros. (/dev/md0, /dev/md1)

I was just converting /dev/md1 to a PV.
I was then moving all of the PE's in /dev/md0 to /dev/md1. Just about the point that it finished, the machine hung.

I could log in, but then just about everything I could do with the filessystem would cause a process to hang.
So after about 10 minutes, I rebooted the machine. Then I got a whole bunch of errors and the fsck was being run on all the filesystems. Then I got an oops. So I rebooted without those disks and upgraded the kernel to 2.6.10.

Things seem to be mostly ok now except:
One of my logical volumes is messed up.
I have a new logical volume called /dev/system/pvmove0 which is huge. I don't want to delete it in case I can use it to recover.

When I run pvmove, it detects the move in progress, but says there are problems:

# pvmove
Number of segments in active LV pvmove0 does not match metadata
Number of segments in active LV pvmove0 does not match metadata
ABORTING: Mirror percentage check failed.

Any ideas on what to do now?
Is there anyway to recover?

Nicholas