After removing the faulty hard disk that I was using to boot from, I decided to try installing using just the two 40GBs that were in RAID. It seems you can’t use GRUB when booting from an LVM partition, so the Debian installer gives you LILO, and insists that the Master Boot Record is located on /dev/hdb. The problem of course being that this motherboard only allows you to boot from the first HDD in the system, which is /dev/hda.

So I tried forcing LILO to install to hda, ahah, success. I’m no longer getting INSERT BOOT DISK, but after decompressing the kernel, the system just reboots. No warnings, no errors, just reboots. Bugger. Then I read that it was possible to have an LVM partition along side normal disk partitions, so I created two normal partitions on the first disk (one for the system, one for swap), and left the rest as LVM. Because it’s only RAID 0, it makes no difference that the whole second hard disk is one LVM partition. So, I’ve got three partitions now, one for the system (active flag, bootable), one swap, and one big data partition. Base system copied over fine, GRUB installed OK, time to reboot in to the new system.

GRUB gives me a boot menu, decompressing the kernel, and bam! Crash, same error I got when the old boot HDD failed. Have I lost another hard disk? Is it the mainboard controller? Does Etch simply not like this hardware configuration? Do I care enough to continue trying to fix this damn thing?