public inbox for gentoo-user@lists.gentoo.org
 help / color / mirror / Atom feed
* [gentoo-user] LVM Recovery....
@ 2008-10-18 16:41 BRM
  2008-10-18 19:14 ` Albert Hopkins
  0 siblings, 1 reply; 3+ messages in thread
From: BRM @ 2008-10-18 16:41 UTC (permalink / raw
  To: Users Gentoo

Okay, so I have a desktop running Gentoo 2008.0 (pre-2008.0 release) that has been a little problematic for a while. The power supply died last May/early June - or rather the system wouldn't boot with that power supply - and it hasn't really been booted since early june at the latest. I just got a new power supply for it, got everything connected and it started up without any issues.

So I decided to sync portage and update (emerge --sync && emerge world -vuDN). After a little bit of playing around with package masks (since I had a few packages unmasked and it was being a problem) I got it straightened away, and let is sit for the night last night to continue on its merry way of updating.

Well..I checked up on it a little while ago, and there was a kernel panic in the middle of a build - reporting a hardware error and recommending looking at mcelog --ascii. Okay. So I restart the system and hear a bit of clicking - and the system won't go past POST. Argh.

Well, I turns out my SATA drive was making the clicking sounds; so I disconnected the power and it booted, only it can't find all the Volumes for the Volume Group configuration since one of the three hard drives is disconnected. This is a first for me - I only started using LVM this last spring, and only on this system.

So now I need to get the system backup and running, and start assessing if I need to do any data recovery. I think the two kind of go hand-in-hand at this step since I need to know the mapping to get the system up and running, and I will also need to know that to assess the situation for data recovery...

I looked at the files in /etc/lvm/backup and /etc/lvm/archive to get an idea of what volumes in the group were on that drive.
Question #1: Is there any where else I should look as well?

From the files there I can see I only had data 4 volumes on the drive in question - and they were all only assigned to two single logical partitions - /opt and /usr/local.  I have only installed stuff via emerge; and a search through the root partition using "ls -lR | grep ^l" revealed only symlinks for a few things - netscape plugins, blackdown jdk, sun jdk, eselect linux java, and skye - being on /opt. So...

Question #2: How do I easily recover the installation so that when I replace /opt and /usr/local with a new mapping they will be replenished with whatever was installed?

(I am assuming any important settings were stored to either /etc or the user's home directory. So I am not worried about recovery on at least /opt; and I doubt I should be worried about /usr/local too.)


Question #3: What do I need to do to get the LVM backup so I run the system again?
I found an article on line [1] but am not sure if that is the wisest manner - partly due to that I have physical volume<->logical volumes spreading disks. I am also not 100% sure if the backup is accurate.

Question #4: Is there a way to get the mapping that LVM is searching for? I try running "lvm vgcfgbackup", however, it fails b/c it can't find all the physical volumes in the volume group.

Note: I think the only thing I changed in /etc/lvm.conf was the filter so it would find the hard drives appropriately. Otherwise, I believe it is the base on from the initial emerge of LVM2; however, it's been quite a while since I last touched anything with LVM.

This is a first for me. Right now, I'm more concerned with getting the system back up and running - I can worry about data recovery later.

TIA,

Ben


[1]http://www.linuxjournal.com/article/8874



^ permalink raw reply	[flat|nested] 3+ messages in thread
* Re: [gentoo-user] LVM Recovery....
@ 2008-10-18 21:31 BRM
  0 siblings, 0 replies; 3+ messages in thread
From: BRM @ 2008-10-18 21:31 UTC (permalink / raw
  To: gentoo-user

Well...I'm fairly certain that data recovery might not be very easy - or cheap for the matter.
The system gets stuck during POST while trying to detect the SATA drive.

Using "vgreduce --removemissing" will be okay - once I verify the current state of the VG.
Is there a way to do so _with_ it trying to detect the existence of the partitions or drives? i.e. skip an integrity check and just print out what it thinks the VG is comprised of - that's really what I want at the moment.

Ben



----- Original Message ----
From: Albert Hopkins <marduk@letterboxes.org>
To: gentoo-user@lists.gentoo.org
Sent: Saturday, October 18, 2008 7:14:30 PM
Subject: Re: [gentoo-user] LVM Recovery....

vgreduce --removemissing

However there's no gaurantee you'll be able to recover your data (LVM is
not redundancy).

-a



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2008-10-18 21:31 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-10-18 16:41 [gentoo-user] LVM Recovery BRM
2008-10-18 19:14 ` Albert Hopkins
  -- strict thread matches above, loose matches on Subject: below --
2008-10-18 21:31 BRM

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox