Skip to contentSkip to conversation

rescue data after hd crash

linux-lvm chris 3 participants 5 messages May 31, 2006

lore
Share & tools
lore

chris

author
#0 lore raw

hi all!

sorry if this is a double post, but i think the first post didn't succeed due to some registration race condition.

so here's my problem:

I have create a volume group with 4 250GByte harddisks with an ext3 file system on the 1TB result.

now the third harddisk failed and the volume group is "dead" - obviously.

lvscan tells me (like any other lv*, pv* command):
Couldn't find device with uuid 'A57CS6-SFlw-1ROR-uIvO-cX2o-OE2g-N3SVZN'. Couldn't find all physical volumes for volume group media_volgrp.

in my understanding many of the files should still be intact (those on the first two disks), and I would like to rescue these files.

i found no hints about such a crash recovery.
(checked http://www.tldp.org/HOWTO/LVM-HOWTO/)

my approach would be to "fake" the third disk, as see what the filesystem makes out of it.

any help is welcome.

thanks, chris

#1 lore raw
Quoted 2 lines · chris

I have create a volume group with 4 250GByte harddisks with an ext3 file system on the 1TB result.

Stripe or concat?

Quoted 1 line

now the third harddisk failed and the volume group is "dead" - obviously.

The way you created the volume had no redundancy - therefore 25% of your data is gone, as well as the volume's "completeness".

If you did a concat, then you (with playing) may be able to get some data back from the first 500GB of the volume - but I wouldn't want to rely on anything that came back from it. I'd suggest getting hold of 2 identical blank disks, and doing a binary (dd) copy of disks 1 and 2 to them - and then work on them. You might be able to use vgcfg{backup,restore} to tell the system that this volume is actually only on 2 disks.... However, anything that references data > 500GB into the volume is almost definitely lost - because although you've got 750G-1TB, it's almost all going to be linked back into the section that you've lost.

If you did a stripe, then you can forget getting any data back off it. You've lost every 4th block of whatever size you striped on. E.g. with a 64k stripe size, you'll have lost from 128k-192k, 384k-448k, etc. - and there's almost definitely going to be nothing salvagable from that.

Graham

#2 lore raw

vgchange -ay --partial

Quoted 29 lines · Graham Wood
Quoted 2 lines · chris

I have create a volume group with 4 250GByte harddisks with an ext3 file system on the 1TB result.

Stripe or concat?

Quoted 1 line

now the third harddisk failed and the volume group is "dead" - obviously.

The way you created the volume had no redundancy - therefore 25% of your data is gone, as well as the volume's "completeness".

If you did a concat, then you (with playing) may be able to get some data back from the first 500GB of the volume - but I wouldn't want to rely on anything that came back from it. I'd suggest getting hold of 2 identical blank disks, and doing a binary (dd) copy of disks 1 and 2 to them - and then work on them. You might be able to use vgcfg{backup,restore} to tell the system that this volume is actually only on 2 disks.... However, anything that references data > 500GB into the volume is almost definitely lost - because although you've got 750G-1TB, it's almost all going to be linked back into the section that you've lost.

If you did a stripe, then you can forget getting any data back off it. You've lost every 4th block of whatever size you striped on. E.g. with a 64k stripe size, you'll have lost from 128k-192k, 384k-448k, etc. - and there's almost definitely going to be nothing salvagable from that.

Graham

_______________________________________________
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/

Signature · 15 lines
Regards,
Heinz    -- The LVM Guy --

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen                                 Red Hat GmbH
Consulting Development Engineer                   Am Sonnenhang 11
Storage Development                               56242 Marienrachdorf
                                                  Germany
[email protected]                            PHONE +49  171 7803392
                                                  FAX   +49 2626 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
author
#3 lore raw

Thanks Heinz,

this looks like what i was serching for. still, it didn't work out.

the first call of this command issued (apart from multiple "couln't find device..") "device-mapper ioctl cmd 9 failed: Das Argument ist ung�ltig"
(i hate german error messages, who invented this? it says "illegal
argument")

but also "1 logic voliume(s) .. now active".

lvscan says "Volume group ... not found"
mount fails.

what should i try next?

regards, Chris

PS: about the stripes, I don't think I have any.

Quoted 36 lines · Heinz Mauelshagen

vgchange -ay --partial

Quoted 31 lines · Graham Wood
Quoted 1 line · chris

I have create a volume group with 4 250GByte harddisks with an ext3

file

Quoted 1 line

system on the 1TB result.

Stripe or concat?

Quoted 1 line

now the third harddisk failed and the volume group is "dead" -

obviously.
The way you created the volume had no redundancy - therefore 25% of your data is gone, as well as the volume's "completeness".

If you did a concat, then you (with playing) may be able to get some data back from the first 500GB of the volume - but I wouldn't want to rely on anything that came back from it. I'd suggest getting hold of 2 identical blank disks, and doing a binary (dd) copy of disks 1 and 2 to them - and then work on them. You might be able to use vgcfg{backup,restore} to tell the system that this volume is actually only on 2 disks.... However, anything that references data > 500GB into the volume is almost definitely lost - because although you've got 750G-1TB, it's almost all going to be linked back into the section that you've lost.

If you did a stripe, then you can forget getting any data back off it. You've lost every 4th block of whatever size you striped on. E.g. with a 64k stripe size, you'll have lost from 128k-192k, 384k-448k, etc. - and there's almost definitely going to be nothing salvagable from that.

Graham

_______________________________________________
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/

#4 lore raw
Quoted 11 lines · chris

Thanks Heinz,

this looks like what i was serching for. still, it didn't work out.

the first call of this command issued (apart from multiple "couln't find device..")
"device-mapper ioctl cmd 9 failed: Das Argument ist ung�ltig"
(i hate german error messages, who invented this? it says "illegal
argument")

but also "1 logic voliume(s) .. now active".

Well, looks like your Logical Volume got activated. Of course with a hole punched out (the missing drive).

Try to mount it manually.

Heinz

Quoted 62 lines

lvscan says "Volume group ... not found"
mount fails.

what should i try next?

regards, Chris

PS: about the stripes, I don't think I have any.

Quoted 47 lines · Heinz Mauelshagen

vgchange -ay --partial

Quoted 42 lines · Graham Wood
Quoted 1 line · chris

I have create a volume group with 4 250GByte harddisks with an ext3

file

Quoted 1 line

system on the 1TB result.

Stripe or concat?

Quoted 1 line

now the third harddisk failed and the volume group is "dead" -

obviously.
The way you created the volume had no redundancy - therefore 25% of your data is gone, as
well as the volume's "completeness".

If you did a concat, then you (with playing) may be able to get some data back from the
first 500GB of the volume - but I wouldn't want to rely on anything that came back from it.
I'd suggest getting hold of 2 identical blank disks, and doing a binary (dd) copy of disks 1
and 2 to them - and then work on them. You might be able to use vgcfg{backup,restore} to
tell the system that this volume is actually only on 2 disks.... However, anything that
references data > 500GB into the volume is almost definitely lost - because although you've
got 750G-1TB, it's almost all going to be linked back into the section that you've lost.

If you did a stripe, then you can forget getting any data back off it. You've lost every
4th block of whatever size you striped on. E.g. with a 64k stripe size, you'll have lost
from 128k-192k, 384k-448k, etc. - and there's almost definitely going to be nothing
salvagable from that.

Graham

_______________________________________________
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/

Quoted 5 lines

_______________________________________________
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/

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen                                 Red Hat GmbH
Consulting Development Engineer                   Am Sonnenhang 11
Storage Development                               56242 Marienrachdorf
                                                  Germany
[email protected]                            PHONE +49  171 7803392
                                                  FAX   +49 2626 924446

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-