After power outtage my vg is gone

linux-lvm Klaus Pedersen 2 participants 6 messages Oct 31, 2001 – Nov 11, 2001 1 message not in the archive

Share & tools

Klaus Pedersen

author

2001-10-31 00:35 UTC #0 lore raw

Hi again,
I really need help on this - sorry to send another request for help. Does anybody have a clue ? anything that can help me remount my vg without dataloss if possible ?
Can anybody point me to documentation in saving a vg ? or a newsgroup on usenet I can try ?

To say it another way - _any_ help would be appreciated !

Thanks in advance

/Klaus Pedersen

----- Original Message -----
From: "Klaus Pedersen" <[email protected]>
To: <[email protected]>
Sent: Sunday, October 28, 2001 7:09 PM
Subject: After power outtage my vg is gone

Quoted 1 line

I'm using Debian Linux on 2.4.6 (and whatever version of LVM that comes

with

Quoted 16 lines

it).
After accidentaly pushing the power switch on my Linux server and starting it up again, I have problems mounting my vg again.

This is what happens :

Vault:/home/ultra# mount /glftpd/site
mount: /dev/site_vg/lvol1 is not a valid block device
Vault:/home/ultra# vgscan
vgscan -- reading all physical volumes (this may take a while...)
vgscan -- found inactive volume group "site_vg"
vgscan -- ERROR "vg_read_with_pv_and_lv(): allocated LE of LV" can't get
data of volume group "site_vg" from physical volume(s)
vgscan -- ERROR "vg_read_with_pv_and_lv(): allocated LE of LV" creating
"/etc/lvmtab" and "/etc/lvmtab.d"

Quoted 4 lines

Vault:/home/ultra#
I read through the archive of the list because I am absolutely sure I am

not

Quoted 1 line

the first who had this problem. I read some stuff about using the

uuid_fixer

Quoted 1 line

tool, bit as far as I was able to read, the problem that fixes only shows

Quoted 1 line

after extending the VG (Which I did a week or so, ago), and only for

kernels

Quoted 19 lines

up to 2.4.5.
Is this true ?
Vault:~/uuid_fixer# ./uuid_fixer /dev/hda8 /dev/hdb1
UUIDs discovered:
W9tkR7eq026WBMCBrUCdebI32vRyFuLn
u5vBzCHwu5mjAt5yN0lS4Ecrto1jI9NZ
I tried running it anyway, and after that I ran vgscan - and the same problems showed up.
What do I do to recoved my data ?
Any help is greatly appreciated
/Klaus

Andreas Dilger

2001-10-31 00:51 UTC #1 lore raw

Quoted 5 lines · Klaus Pedersen

I really need help on this - sorry to send another request for help. Does anybody have a clue ? anything that can help me remount my vg without dataloss if possible ?
Can anybody point me to documentation in saving a vg ? or a newsgroup on usenet I can try ?

Try "vgcfgrestore".

Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/

Klaus Pedersen

author

2001-10-31 01:15 UTC #2 lore raw

I think that gave a little breakthrough :-)
First of all - as murphys law suggests, I don't have a backup to use with vgcfgrestore - but with the tool I did manage to recreate both /etc/lvmtab and /etc/lvmtab.d :

Vault:/home/ultra# vgcfgrestore -n site_vg /dev/hda8
vgcfgrestore -- VGDA for "site_vg" successfully restored to physical volume
"/dev/hda8"

Vault:/home/ultra# vgcfgrestore -n site_vg /dev/hdb1
vgcfgrestore -- VGDA for "site_vg" successfully restored to physical volume
"/dev/hdb1"

the only problem is that they are both really empty, so with vgscan I find nothing :

Vault:/etc# vgscan
vgscan -- reading all physical volumes (this may take a while...)
vgscan -- no volume groups found

But to my big relief, the data is still present on the physical drives :

Vault:/etc# pvdisplay /dev/hda8
--- Physical volume ---
PV Name               /dev/hda8
VG Name               site_vg
PV Size               14.1 GB / NOT usable 1.21 MB [LVM: 135 KB]
PV#                   2
PV Status             available
Allocatable           yes
Cur LV                1
PE Size (KByte)       4096
Total PE              3610
Free PE               108
Allocated PE          3502
PV UUID               W9tkR7-eq02-6WBM-CBrU-Cdeb-I32v-RyFuLn

Vault:/etc# pvdisplay /dev/hdb1
--- Physical volume ---
PV Name               /dev/hdb1
VG Name               site_vg
PV Size               57.25 GB / NOT usable 3.96 MB [LVM: 178 KB]
PV#                   1
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       4096
Total PE              14654
Free PE               0
Allocated PE          14654
PV UUID               u5vBzC-Hwu5-mjAt-5yN0-lS4E-crto-1jI9NZ

Thanks for the help so far Andreas, but where do I go from here ?

Thanks in advance

Klaus Pedersen

----- Original Message -----
From: "Andreas Dilger" <[email protected]>
To: <[email protected]>
Sent: Wednesday, October 31, 2001 7:52 AM
Subject: Re: [linux-lvm] Re: After power outtage my vg is gone

Quoted 3 lines · Klaus Pedersen

Quoted 2 lines
I really need help on this - sorry to send another request for help. Does anybody have a clue ? anything that can help me remount my vg

without

Quoted 18 lines

Quoted 3 lines
dataloss if possible ?
Can anybody point me to documentation in saving a vg ? or a newsgroup on usenet I can try ?
Try "vgcfgrestore".
Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/
_______________________________________________
linux-lvm mailing list
[email protected]
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

Klaus Pedersen

author

2001-11-01 12:41 UTC #3 lore raw

Hi again

I finally got my problem solved, so I write this message to tell others how I solved it, so it's in the mail archives - and also because I want to point out that this should be in the HOW-TO, since my problem is probably not so uncommonm and very easy to solve - IF you know how.

Basically my problem is that after a power outtage I was unable to mount my vgs - I got this message :

Vault:/home/ultra# mount /glftpd/site
mount: /dev/site_vg/lvol1 is not a valid block device

I found out that the problem is that the file /etc/lvmtab and the directory /etc/lvmtab.d is gone for some reason.

Basically I solved this with a big help from Andreas to dropped me a short message to use vgcfgrestore, so I did.
In the how-to I find it very vagely documented exactly what this tool does, and when it's being used.
Also, I was unable to read in the how-to that after this is the sequence of commands that should be used :
vgcfgrestore -n site_vg /dev/hda8
vgcfgrestore -n site_vg /dev/hdb1
(my vg consists of those two partitions)

vgscan -a y (to activate the vg)

and then do the actual mounting :

mount /glftpd/site

Yes, I know that to some very experienced LVM developer and such, this is VERY simple - but to a guy like me who really doesn't know much about LVM, it was very advanced and hard to find out.

So please - put an advanced troubleshooting guide in the LVM how-to on sistina.com - I find it very hard to believe that I am the only one who could use it.

/Klaus

----- Original Message -----
From: "Klaus Pedersen" <[email protected]>
To: <[email protected]>
Sent: Wednesday, October 31, 2001 8:17 AM
Subject: Re: [linux-lvm] Re: After power outtage my vg is gone

Quoted 7 lines

I think that gave a little breakthrough :-)
First of all - as murphys law suggests, I don't have a backup to use with vgcfgrestore - but with the tool I did manage to recreate both /etc/lvmtab and /etc/lvmtab.d :
Vault:/home/ultra# vgcfgrestore -n site_vg /dev/hda8
vgcfgrestore -- VGDA for "site_vg" successfully restored to physical

volume

Quoted 4 lines

"/dev/hda8"

Vault:/home/ultra# vgcfgrestore -n site_vg /dev/hdb1
vgcfgrestore -- VGDA for "site_vg" successfully restored to physical

volume

Quoted 49 lines

"/dev/hdb1"

the only problem is that they are both really empty, so with vgscan I find nothing :

Vault:/etc# vgscan
vgscan -- reading all physical volumes (this may take a while...)
vgscan -- no volume groups found

But to my big relief, the data is still present on the physical drives :

Vault:/etc# pvdisplay /dev/hda8
--- Physical volume ---
PV Name               /dev/hda8
VG Name               site_vg
PV Size               14.1 GB / NOT usable 1.21 MB [LVM: 135 KB]
PV#                   2
PV Status             available
Allocatable           yes
Cur LV                1
PE Size (KByte)       4096
Total PE              3610
Free PE               108
Allocated PE          3502
PV UUID               W9tkR7-eq02-6WBM-CBrU-Cdeb-I32v-RyFuLn

Vault:/etc# pvdisplay /dev/hdb1
--- Physical volume ---
PV Name               /dev/hdb1
VG Name               site_vg
PV Size               57.25 GB / NOT usable 3.96 MB [LVM: 178 KB]
PV#                   1
PV Status             NOT available
Allocatable           yes (but full)
Cur LV                1
PE Size (KByte)       4096
Total PE              14654
Free PE               0
Allocated PE          14654
PV UUID               u5vBzC-Hwu5-mjAt-5yN0-lS4E-crto-1jI9NZ

Thanks for the help so far Andreas, but where do I go from here ?

Thanks in advance

Klaus Pedersen

A message that is not in the archive (sent off-list or not yet received): [email protected]

Andreas Dilger

2001-11-01 15:18 UTC #5 lore raw

Quoted 3 lines · Magosányi Árpád

Another item which needs inclusion is the bit about restoring the uuid. I think of that little perverse 'dd' commands which should be done with pw_uuidlist_on_disk.base.

This is now handled by the "uuid_fixer" commands, which are included in the LVM CVS, I think (I don't know if it is installed).

Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/

Klaus Pedersen

author

2001-11-11 13:47 UTC #6 lore raw

In reply to Klaus Pedersen

Quoted 2 lines

vgscan -a y (to activate the vg)

Duh - this is an error (I know cause the same problems occured - again :P) .. the command _should_ have said :

vgchange -a y

so the sequence of commands is :

vgcfgrestore -n site_vg /dev/hdb1
vgcfgrestore -n site_vg /dev/hda8
vgchange -a y
mount /glftpd/site

That's it - I hope others can benefit from this too.

/Klaus