Kernel Update nuked GRUB & initrd on an LVM2 setup

English Install/Boot/Login
1

For some arcane reason !! ] the most recent kernel update seems to have destroyed my old initrd [obsoleting the vmlinuz-2.6.25.16-0.1-default and initrd[…] symbolic links in /boot]

It wiped out my old grub config, and though it is fortunate I have a backup, the old initird-default and vmlinuz-default images are -gone-!

The new kernel didn’t ever seem to compile. Now I can only access my partitions with a LiveCD.

When I first booted, it told me the “volume group hive” could not be found.

I know I must now create a new initrd for my chosen kernel. However, I am having some trouble doing it from the live CD… here is my output:

linux:/boot1/grub # mkinitrd -f lvm2 -d /dev/hive/root/ -b /boot1

Kernel image: /boot/vmlinuz-2.6.25.5-1.1-default
Initrd image: /boot/initrd-2.6.25.5-1.1-default
Root device (/dev/hive/root/) not found
2008-11-10 18:46:54 ERROR: Bootloader::Library::SetLoaderType: Initializing for unknown bootloader
2008-11-10 18:46:54 ERROR: Bootloader::Core::ListFiles: Running generic function, it should never be called
2008-11-10 18:46:54 ERROR: Bootloader::Core::ParseLines: Running generic function, it should never be called

Please help!! My work is frozen on an unbootable volume :cry:

2

hello?

openSUSE updates should never completely nuke the OS. This isn’t the first time this has happened, either.

3

So I got a little farther, but keep running into walls.

Using chroot to resolve path issues with mkinitrd, though it does not solve them all.

I must still, apparently, mount /sys and /proc and mount the relative root with a --bind switch. [Why?]

However, every time I attempt a command-line mount, it whines about the destination not being a mount point.

e.x. mount --bind /dev/sda4 /core/boot
“bla bla mount point /core/boot does not exist”

What am I doing wrong?