I've recently had two drives fail in a RAID5 array during an upgrade. They weren't really failed, but had to be re-added to the array. The array has rebuilt itself just fine.
But, my lvm config appears to be in bad shape.
I have the VG archive file and it appears correct.
My problem is that I keep receiving "Incorrect metadata area header checksum" when I try to do nearly any operation on the PV, VG, or LVs.
I tried to recreate the PV on the array (/dev/md1) using the VG archive file.
Code:
# pvcreate --uuid "<long_UUID>" --restorefile /etc/lvm/archive/raid5a_00007.vg pv0
Incorrect metadata area header checksum
Incorrect metadata area header checksum
uuid <long_UUID> already in use on "/dev/md1"
OK, so it's there, but has some header problems. So, I try to do something similar with the volume group, but 'vgscan' only returns "Incorrect metadata area header checksum" as does nearly every VG and PV operation.
'pvscan' returns that the physical volume exists but that the one volume is not assigned to a volume group.
Is there a way to just "force" the PV to accept the data from the VG archive file and apply it?
Thanks,
--dave;
