I have an old AMD box that I use mostly for experimenting stuff, and it runs Debian unstable. Yesterday I had to turn it off, and after that it wouldn't boot again.
I have a couple of kernels installed, all compiled locally (with
3.11.0+ (from git, very recent);
When I boot, it will complain that it did not find the LVM volume group (there is only one). From the initramfs shell, I ran
lvm lvmdiskscan, which tells me that there is not even a physical volume available (the data below was manually copied, I can't copy/paste):
read_urandom: /dev/urandom: openfailed: No such file or directory 0 disks 0 partitions 0 LVM physical volume whole disks 0 LVM physical volumes
/dev/mapper does not contain the files for the logical volumes.
Now, the strange part is that if I boot from a rescue USB stick, then LVM works just fine. Booting from the USB stick and starting a shell on the HD root partition, I could mount all LVM volumes. I was even able to run
upgrade, and I have already ran
grub-install, and I even compiled a new kernel (running from the USB stick), and installed it.
But when I boot again from the HD, LVM stops working again.
Here are the versions o some relevant packages:
grub-pc 1.99-27.1 grob-pc-bin 1.99-27.1 grub-common 1.99-27.1 grub-pc-common 1.99-27.1 initramfs-tools 0.113 lvm2 2.02.98-6+b1
(I have been holding the update of grub because I know there are issues with the new versions; all the rest is the most recent version).
So -- what else could I try besides reinstalling the OS?
edit: the rescue disk runs kernel 3.2.0-4-amd64; not sure if this makes any difference.