Skip to contentSkip to conversation

LVM problem - how to recover?

linux-lvm Richard Kampmann 3 participants 5 messages Nov 28, 2006 – Dec 8, 2006

lore
Share & tools
lore
#0 lore raw

Hi everybody ...

I recently had a lvm crash from which I now have to recover. So now I need your help ...

The system is a SUSE 10.1 system with actual patches.

# uname -rio
2.6.16.21-0.25-default i386 GNU/Linux
# lvm version
  LVM version:     2.02.02 (2006-02-07)
  Library version: 1.02.03 (2006-02-08)
  Driver version:  4.5.0
#

There was the unenlightened crash and before there was a volume goup spanning two physical partitions (sda2, sdb2), each on one physical disk (sda, sdb). The system froze (maybe in coincidence with some snapshot-handling) and I had to press the reset button. When rebooting lvm told me he could not open the volume group because one pv was missing. The partition looked OK when checking the partition table. I tried several things - and to do a vgcfgrestore but it did not help.

The volume group (system_vg) contained the root partition, so I installed a new system on a third disc and tried to recover the old volume group. Unfortunately when doing the pvcreate for the two physical volumes I did NOT choose the "--zero n" switch.

Fact is: I now have the volume group containing two physical discs and two physical volumes with the correct PV UUIDs. But the first sectors are zeroed, aren't they?

Any way to recover the LVs of the old volume group? A vgcfgrestore (YES, there was a vgcfgbackup before!) does not work ... bad luck.

Ah - and all logical volumes contain Reiser Filesystems V 3.6 - they MUST still be there ...

So - what could be my next steps?

Bye

Richard Kampmann

#1 lore raw

Hi everybody,

I'm a bit disappointed, because there was really no feedback to my mail.

The situation is a real problem for me - is there anything missing? Or is there no way to recover (can't believe that ...)?
Wrong way or place to ask? Then please tell me where to go ...

I tried to find a solution myself (also with help of Google ...), but it did not help me.

Bye

Richard Kampmann

Am Dienstag, 28. November 2006 21:15 schrieb Richard Kampmann:

Quoted 42 lines

Hi everybody ...

I recently had a lvm crash from which I now have to recover. So now I need your help ...

The system is a SUSE 10.1 system with actual patches.

# uname -rio
2.6.16.21-0.25-default i386 GNU/Linux
# lvm version
  LVM version:     2.02.02 (2006-02-07)
  Library version: 1.02.03 (2006-02-08)
  Driver version:  4.5.0
#

There was the unenlightened crash and before there was a volume goup spanning two physical partitions (sda2, sdb2), each on one physical disk (sda, sdb). The system froze (maybe in coincidence with some snapshot-handling) and I had to press the reset button. When rebooting lvm told me he could not open the volume group because one pv was missing. The partition looked OK when checking the partition table. I tried several things - and to do a vgcfgrestore but it did not help.

The volume group (system_vg) contained the root partition, so I installed a new system on a third disc and tried to recover the old volume group. Unfortunately when doing the pvcreate for the two physical volumes I did NOT choose the "--zero n" switch.

Fact is: I now have the volume group containing two physical discs and two physical volumes with the correct PV UUIDs. But the first sectors are zeroed, aren't they?

Any way to recover the LVs of the old volume group? A vgcfgrestore (YES, there was a vgcfgbackup before!) does not work ... bad luck.

Ah - and all logical volumes contain Reiser Filesystems V 3.6 - they MUST still be there ...

So - what could be my next steps?

Bye

Richard Kampmann

#2 lore raw

Hi Richard,
1. What does/did vgcfgrestore tell you when you tried to use it? You said it did not work....

2. Does debugreiserfs /dev/mapper/VGgroup-LVlist say that there is a filesystem there? (assuming you have not created a new filesystem there && also assuming you can perform "vgchange -a y VGgroup" )

_thanks

Richard

--- Richard Kampmann <[email protected]>

Quoted 77 lines · wrote:

Hi everybody,

I'm a bit disappointed, because there was really no feedback to my mail.

The situation is a real problem for me - is there anything missing? Or is
there no way to recover (can't believe that ...)? Wrong way or place to ask? Then please tell me where to go ...

I tried to find a solution myself (also with help of Google ...), but it did
not help me.

Bye

Richard Kampmann

Am Dienstag, 28. November 2006 21:15 schrieb Richard Kampmann:

Quoted 3 lines

Hi everybody ...

I recently had a lvm crash from which I now have

to recover. So now I need

Quoted 3 lines

your help ...

The system is a SUSE 10.1 system with actual

patches.

Quoted 9 lines
# uname -rio
2.6.16.21-0.25-default i386 GNU/Linux
# lvm version
  LVM version:     2.02.02 (2006-02-07)
  Library version: 1.02.03 (2006-02-08)
  Driver version:  4.5.0
#

There was the unenlightened crash and before there

was a volume goup

Quoted 1 line

spanning two physical partitions (sda2, sdb2),

each on one physical disk

Quoted 1 line

(sda, sdb). The system froze (maybe in coincidence

with some

Quoted 1 line

snapshot-handling) and I had to press the reset

button. When rebooting lvm

Quoted 1 line

told me he could not open the volume group because

one pv was missing. The

Quoted 1 line

partition looked OK when checking the partition

table. I tried several

Quoted 1 line

things - and to do a vgcfgrestore but it did not

help.

Quoted 2 lines

The volume group (system_vg) contained the root

partition, so I installed a

Quoted 1 line

new system on a third disc and tried to recover

the old volume group.

Quoted 1 line

Unfortunately when doing the pvcreate for the two

physical volumes I did

Quoted 3 lines

NOT choose the "--zero n" switch.

Fact is: I now have the volume group containing

two physical discs and two

Quoted 1 line

physical volumes with the correct PV UUIDs. But

the first sectors are

Quoted 3 lines

zeroed, aren't they?

Any way to recover the LVs of the old volume

group? A vgcfgrestore (YES,

Quoted 1 line

there was a vgcfgbackup before!) does not work ...

bad luck.

Quoted 2 lines

Ah - and all logical volumes contain Reiser

Filesystems V 3.6 - they MUST

Quoted 4 lines

still be there ...

So - what could be my next steps?

Quoted 8 lines
Quoted 4 lines

Bye

Richard Kampmann
_______________________________________________

linux-lvm mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at

http://tldp.org/HOWTO/LVM-HOWTO/

____________________________________________________________________________________
Do you Yahoo!?
Everyone is raving about the all-new Yahoo! Mail beta.
http://new.mail.yahoo.com

#3 lore raw
In reply to Richard Kampmann
Quoted 2 lines · Richard Kampmann

Any way to recover the LVs of the old volume group? A vgcfgrestore (YES, there was a vgcfgbackup before!) does not work ... bad luck.

What errors do you get using vgcfgrestore?

Signature · 3 lines
Zac Slade
[email protected]
ICQ:1415282 YM:krakrjak AIM:ttyp99
#4 lore raw

Re: [linux-lvm] LVM problem - how to recover? (solved)

Hi Zac,

Quoted 4 lines
Quoted 2 lines

Any way to recover the LVs of the old volume group? A vgcfgrestore (YES, there was a vgcfgbackup before!) does not work ... bad luck.

What errors do you get using vgcfgrestore?

thank you for your answer. I got various errors ...

The reason for this was that all hints and todos I found had the wrong syntax (maybe old lvm-version ...). The manpage of vgcfgrestore does not tell anything helpful.

In the meantime I found the right syntax and with a little trying and stumbling I managed to get access to the important LVs on both disks. Most of my important data is available - only one filesystem was badly corrupted. Maybe a 'reiserfsck --rebuild-tree' helps (runs at the moment).

So my problem seems to be solved - thanks.

Bye

Richard