LVM problem
Ben Scott
dragonhawk at gmail.com
Fri Dec 14 16:38:41 EST 2007
On Dec 14, 2007 12:23 PM, Dan Coutu <coutu at snowy-owl.com> wrote:
> Yes, I did them in that order. The reason that I did the mkfs ...
If you ran mkfs on a device that was a PV, after you ran "pvcreate",
you've overwritten that PV's on-disk metadata, and basically wrecked
that PV.
All the subsequent error messages about not finding the signature or
UUID are likely because LVM has copies of the metadata in other
locations. So LVM knows there *should* be a PV there, but can't find
it goes to look. (I know LVM caches metadata under /etc/lvm/, and I
think it also keeps copies of all the metadata for a VG in each member
of the VG.)
> # vgextend VolGroup00 /dev/sdb
> /dev/cdrom: open failed: No medium found
> Volume group "VolGroup00" successfully extended
The /dev/cdrom part is likely a red herring. Some LVM utilities
scan all the possible devices in the system, looking for devices which
might have PVs on them. Since any IDE node (hda, hdb, hdc, etc.)
might have a hard disk on it, it's going to probe all of those. Once
of those was likely your CD-ROM.
The second message, about the VG being successfully extended,
indicates the PV was successfully added to the VG. Then you blew it
away with the mkfs. ;-)
> I've tried a new pvcreate but can't get rid of the uuid error message
> and cannot do any vg* commands at all because of it.
You probably need to first do a vgreduce to remove the (wrecked) PV
from the VG. You may also need to do a pvremove to remove all traces
of the wrecked PV; not sure on that.
-- Ben
More information about the gnhlug-discuss
mailing list