[linux-lvm] vgscan errors, segmentation faults
Heinz J . Mauelshagen
mauelshagen at sistina.com
Fri Mar 22 03:52:01 UTC 2002
please put together all your vg01.conf* metadata archives *and*
the actual metadata on all devices retrieved with
dd if=/dev/PV of=PV.vgda bs=1k count=4k
(replace PV by hda2 etc. running that command on all PVs)
in a "tar Icf" created archive for further investigation.
Send that to me attached via email ASAP because I'll be leaving for a
vacation trip tonight and can't help you before after Easter again.
Heinz -- The LVM Guy --
On Thu, Mar 21, 2002 at 12:49:27PM +0100, Florian Hinzmann wrote:
> I am desperate. I can't access my LVs at all currently.
> vgscan and vgcfgrestore say something about
> "only found 0 of x LEs for LV foo" and then they die
> with a segmentation fault.
> "vgchange -ay" does the same of course ..
> The box with that problem is an ordinary PC running
> Debian Linux (unstable/sid).
> I am trying to solve this right now. I will check for replies
> to this mail (list and private) regularily. I am eager to
> help debugging this; giving more information or doing
> things someone wants me to try.
> Short: Please help me, I will do whatever I can to help. ;)
> I will now try to remember as much as possible of
> what happened and what I did:
> - I started to play around with LVM (v1.0.1, Debian package) with
> a stock 2.4.17 kernel (compiled by myself).
> - I had a volume group spread across several partitions
> (one per physical harddisk).
> - I added and removed PVs to my VG and resized some LVs.
> - Later I had two VGs and I was shifting them around quite
> - Later I merged that two groups.
> A short time after that action the problems began. I don't remember
> my actions exactly. I think I had LVs with equal names inside both
> VGs. I did some moving of PEs, I renamed at least one LV and at some
> point I tried to merge both VGs.
> If I remember that correctly vgmerge worked fine, but I had some
> problems afterwards. I am not able to recollect all the steps I perfomed.
> I am sure activating did not work at some point and vgscan did not
> display my new VG correctly. The only way I found to mount my LVs
> was calling
> "vgcfgrestore -f foo/vg01.conf /dev/hda2" and
> "vgcfgrestore -f foo/vg01.conf /dev/hdb1" and
> _not_ calling any scan programs afterwards.
> Especially vgscan messed up my setup. I remember getting contradictory
> messages about to which VG /dev/hdb1 belongs while trying to
> solve that.
> Whatever, I was able to use my LVs if I wrote the metadata using
> vgcfgrestore. And calling vgscan was a sure way to make my LVs not
> I thought the new version might solve this for me and that box
> is running with lvm-1.1rc1 (again Debian package) + kernel 2.4.18 +
> linux-2.4.18-VFS-lock.patch + lvm-1.1-rc1-2.4.18.patch.
> But now vgscan and vgcfgrestore segfault like I did
> write above.
> What can I try? What information may I provide?
> Please help!
> Florian Hinzmann
> Florian Hinzmann private: f.hinzmann at hamburg.de
> Debian: fh at debian.org
> PGP Key / ID: 1024D/B4071A65
> Fingerprint : F9AB 00C1 3E3A 8125 DD3F DF1C DF79 A374 B407 1A65
> linux-lvm mailing list
> linux-lvm at sistina.com
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html
*** Software bugs are stupid.
Nevertheless it needs not so stupid people to solve them ***
Heinz Mauelshagen Sistina Software Inc.
Senior Consultant/Developer Am Sonnenhang 11
Mauelshagen at Sistina.com +49 2626 141200
More information about the linux-lvm