> this brings me back to my original post in this thread. I get an
> error when I try to specify the "real" device and the MBR does not
> get written. do I have to stop the raid for the boot partition and
> mount the partitions as individual sdx ? I would think that if I
> tell lilo to write the MBR to a raid 1 that it would end up on both
> disks in the array.
real lilo does not know about or understand raid devices. I don't
know exactly what the redhat lilo patch does, but I bet it makes some
assumptions about geometry and what you wish to do when a disk fails
and must be rebooted -- i.e. whether or not you want the disk to boot
from the existing bios position or if you will move it -- very mbrd
dependent.
using stock lilo, you can write it anyway you want.
lilo does not care if the disk is mounted or not, it will write to
the mbr of the specified drive "period" whether it is in a raid array
or whether it is mounted is not an issue.
look carefully at the examples in the boot_root...HOWTO. They all
come from real systems, not ones that were dreamed up.
I have about 8 raid systems here + several at customer sites using
various disks IDE, scsi, raid1, 5, 10, 10+1.
The boot requirements on most require that the boot drive be bios 80
which implies that the "D:" (bios 81) drive must be move to achieve
boot if the "C:" (bios 80) drive fails -- a few motherboards will
boot from the alternate drive -- usually I don't guess, just provide
a boot floppy that mounts the md device and does not care if C or D
is present (automount) so in the event of drive failure, if a reboot
is necessary then the diskette can be used.
>
> my original post:
>
> > Hi, I have a raid level 1 on 2 scsi drives and an IDE drive.
used to be you could not mix scsi and ide in raid -- don't know about
the current code -- might ask the list guru's
> > I'm trying to set MBR on all my drives so if the first one goes down the
> > system will still boot.
see above about bios device
> >
> > I followed advice from this list's archive and set up a lilo.conf file for
> > each drive:
> > lilo.conf.hda
> > lilo.conf.sda
> > lilo.conf.sdb
> >
> > when I try
> > lilo -C /etc/lilo.conf.sdb
> > I get this error:
> > [root@gold /root]# lilo -C /etc/lilo.conf.sdb
> > Warning: /dev/sdb is not on the first disk
> > Sorry, don't know how to handle device 0x0905
> >
> >
> >
> > here is my lilo.conf.sdb
> > the other two are identical except for the boot= they are boot=/dev/sda and
> > boot=/dev/hda
> > I am also not sure about the bios=0x80 param (from the boot on raid how-to)
> > it also mentions bios=81
> > which do I use? I have an adaptec 3940U2W
> >
> > ---------lilo.conf.sdb
> > disk=/dev/md0
> > bios=0x80
> > sectors=63
> > heads=255
> > cylinders=2231
> > partition=/dev/md5
> > boot=/dev/sdb
> > map=/boot/map
> > install=/boot/boot.b
> > vga=normal
> > default=linux-grc
> > keytable=/boot/us.klt
> > prompt
> > timeout=50
> > message=/boot/message
> > image=/boot/vmlinuz-2.2.14-grc
> > label=linux-raid
> > root=/dev/md0
> > initrd=/boot/initrd.img
> > append=""
> > read-only
> > image=/boot/vmlinuz-2.2.14-grc
> > label=linux-grc
> > root=/dev/hda5
> > initrd=/boot/initrd.img
> > append=""
> > read-only
> > image=/boot/vmlinuz-2.2.14-15mdk
> > label=old_linux-up
> > root=/dev/hda5
> > initrd=/boot/initrd.img
> > append=""
> > read-only
> > image=/boot/vmlinuz-secure
> > label=linux
> > root=/dev/hda5
> > initrd=/boot/initrd-secure.img
> > append=""
> > read-only
> > image=/boot/vmlinuz
> > label=linux-up
> > root=/dev/hda5
> > initrd=/boot/initrd.img
> > append=""
> > read-only
> > other=/dev/fd0
> > label=floppy
> > unsafe
> >
> >
> >
>