4

How do I recover from boot error when my root is on RAID and LVM?

view full story
linux-howto

http://serverfault.com – I run an Archlinux server and it has happened to me three times now that I did a pacman -Syu, forgot about it and then a powerfailure happened and my box wouldn't boot anymore. (yes I know I'm not a very thoughtful sysadmin) Ofcourse I had forgotten to run mkinitcpio or some other configuration that pacman did not resolve automatically. Since I made the mistake of having my root partition on both RAID and LVM the recovery process is a bit more complex, so how to do it? edit: I have the answer to my question but I have to wait 7 hours to post it. (HowTos)

xskl's picture
Submitted by xskl on 08/04/2012 – Made popular on 08/04/2012

Stories similar to How do I recover from boot error when my root is on RAID and LVM?

I have fake (nvidia) raid 1 setup using dmraid. All worked pretty well for almost a month and every update went fine. Except the last one to linux 3.1.9-2. There was some output lines from pacman saying something like "hook dmraid is not available" during linux package installation.

Problem after update 1 year 9 weeks ago

pacman log wrote:[2012-03-10 23:28] Running 'pacman -Syuf'Recipe for breakage...It gets worse:pacman log wrote:[2012-03-10 23:30] /tmp/alpm_beBGjj/.INSTALL: line 47: mountpoint: command not found[2012-03-10 23:30] WARNING: /boot appears to be a separate partition but is not mounted.[2012-03-10 23:30] >>> Updating module dependencies.

I'm trying to install /boot in a RAID 1 using 3 TB disks. As the RAID partition do not let you manage partitions than more than 2 TB, CentOS automatically configures the partitions using GPT.

My current machine (HP Proliant ML110 G4) does not support EFI/GPT so it uses MBR to boot the system.

Hoping someone can help...

Here is my setup:

4x Western Digital 1TB drives - 200GB RAID 0 (Windows 7 and Apps / Games) - 1.8TB RAID 10 (Data)

Both arrays use all 4 disks.

Ubuntu has its own 80GB drive (/dev/sde).

In F17 and earlier I have been able to boot from install disk and then ultimately access a raid 1 and then do a custom install. I select a root partition and home partition, etc., and the size of each, always reserving some free space in the volume group that is on top of my raid 1.

In F18 install I can't even see any raid.

on a test machine pacman -Syu offers linux-3.6.2-1.After a successfull update the system won't boot anymore with the following error:[ 0.000000] tsc: Fast TSC calibration failed :: running early hook [udev] :: running hook [udev] :: triggering uevents... Waiting 10 seconds for device /dev/md1 ... ERROR: device '/dev/md1' not found.

Move a Whole Ubuntu Server 6 weeks 9 hours ago

I have a machine with Ubuntu 12.04.1 LTS 64 Bits installed on a LVM except for the boot partition

the current settings is

/dev/sda1 /boot 1GB /dev/sda2 lvm

on LVM

/dev/lvm/root 45GB /dev/lvm/swap 5GB

my plan is move the whole system to a new RAID (while is online) now i have a few questions:

Which files I need to change besides /etc/fstab how should I set this new partition on the RAID?

Well, you can see your RAID partitions and you're getting GRUB to load, and you are even being dropped to the recovery shell which means that the /boot partition is being found and is accessible. All vary, vary good things.I would boot into the Arch live CD/USB again. Then check the UUID of the / Root partition.

I have spend the last few hours looking at posts from people with similar problems.