Andy Poling <[EMAIL PROTECTED]> writes:
> LILO is just not going to let me boot my RAID 1 root.
>
> This is not a slam on LILO. Heck, I've happily used it for years to boot my
> Linux systems, and probably will continue to on my non-root-raid systems
> just because it works out of the box.
How come I've been running this for about a year and a half, then?
First with an initrd setting up the raid, and lately using
autodetect...
To quote from my lilo.conf:
disk=/dev/md1
bios=0x80
sectors=32
heads=64
cylinders=4341
partition = /dev/md0
start = 32
boot=/dev/sda
map=/boot/map
install=/boot/boot.b
image=/boot/vmlinuz-2.2.5-22smp
label=linux
root=/dev/md0
initrd=/boot/initrd-2.2.5-22smp.img
read-only
The root partition is 5-way mirrored on /dev/md0, with each partition
at the start of the disk. The /dev/md1 is never used, but needs to
refer to a device with the same major number as /dev/md0 for lilo to
be happy.
The numbers for sec/head/cyl are simply taken from fdisk -l /dev/sda,
while you have to think a bit about the start parameter. If the
partition is at the start of the disk, the first partition will start
on cyl 0, head 1, sec 0, which means that start = sectors. Other
offsets I haven't looked into.
After running lilo to make the boot-block on the first disk of the
mirror set, I can just copy the boot-block of this disk onto each of
the others, and I can boot my system with any of these mapped by the
SCSI bios as C: (aka bios 0x80).
At the moment, this is running on a stock redhat 6.0 kernel, with an
initrd built with 'mkinitrd --preload raid0 --preload raid1 ...'. If
you roll your own kernel with raid 0.90 and autodetect compiled in,
there is no need for initrd at all.
-Harald
--
Harald Nordg�rd-Hansen, <>< http://bukharin.hiof.no/~hnh/ <>< Phone/Fax:
�stfold College, School of Computer Sciences, Norway <>< +47 6910 4033/4002
