The boot process proceeds normally up to "Your system seems to have shut down uncleanly...". I then press Y at the prompt to "...force a system file check". The system file check finds no problems but does not solve the Logical Volume Management stall problem that follows.

The next line in the boot readout is "Checking root file system". It checks [OK], then "Remounting root filesystem in read-write mode" is [OK], and the process proceeds to "Setting up Logical Volume Management". It then stalls and never completes the boot process. This problem only occurs after the system has experienced a power failure. I have tried "linux rescue" from the CentOS install disks without being able to find the problem. I have also used Knoppix to access the disk file system and run fsck.ext3 on the hard drives. fsck does not find any file system problems. That is why I think there is a LVM problem which stalls the system.

I have recovered this server from this condition once before. At that time I copied the server partitions using Acronis Disk Director Suite to a USB drive, formatted the server and reinstalled CentOS 4.4. After that I had to reload all the server apps and then copy back all the relevant database, Apache and database access apps. This is a laborious process and leaves our client off line for a couple of days. There must be a more efficient way to recover Logical Volume Management. Does anyone have a process for rebuilding it?

Is there a real reason you're using LVMs? 9 times out of 10, they only add an extra layer of complexity & are not utilized *at all* for doing what they are intended to do. LVM provides a neat way to manage expanding file systems, but if you're not adding/removing/replacing disks, it's really of no use. (I'm sure someone will come in after me a sing LVM's praises hehehe)

The reason I ask is that you can simply remove the LVM from the equation if you don't use it. Simply edit your fstab (you'll need to boot from a cd distro or drop to a prompt when the fsck comes up) so that instead of it reading LVM001 (or however it's in there), change it to the standard /dev/hda1 (or whatever you have). You probably don't even have to rebuild grub.

What I did was this (also with cent).

Boot off of CD.
fdisk -l to find the drive(s)
mount each one & make sure the data is there
edit fstab on the mounted drive to point to standard /dev/ file system
edit menu.lst (I think this was necessary - not totally sure)
umount drives
reboot.

After I had the same problem described here several times, I discovered another, perhaps easier, way to fix the issue.

Boot linux, using , <esc> to access the grub menu.
Press "a" to add a numeral "3" to the normal boot line (go to runlevel 3, I believe)
When the boot process complains that you shut down uncleanly, press "Y" to clean things up.
When the cleanup is complete, the boot should now complete and end up in a request to log in. Log in as root (so you can shut down later).
Type "startx" to the # prompt.
To a terminal session type "shutdown -h now".
When you restart, the boot should proceed normally.

@draichle

I guess you did not have the LVM problem.

And to others that might want to reply, this thread is 5 years old

Not so. I had exactly the symptoms that were originally reported by creata.

OK, apologies