Hi,
welcome to the club.... if you read through recent archives of the
list you will find, that there
are several people on the list (including myself) facing the same.
It seems that the lilo-patch is buggy. It accepts a raid-device for
boot but does not build
the boot block on all devices.
There were suggestions on how to build the boot-block on each drive manually.
I used the following procedure to ship around this bug and keep a
running system:
- every drive on its own ide-bus
- Autodection on for all drives in BIOS (so system starts even when a
disk fails)
- Build a boot disk for the raid and leave it in system
- boot priority for booting ide, than for floppy.
So far, this worked. I manually disabled both ide-drives without a problem.
What do you think you gain by putting swap on a raid-volume?? I set
up two swap-volumes. one on each
drive. This works, even if one disk is missing, giving just a notice
while the system tries to activate swaps.
Regards
Jochen
At 20:57 Uhr +0000 05.12.1999, Lyndon David wrote:
>Hi,
>
>I have a test system on which I am setting up mirrored disks before
>taking the plunge and
>doing it for real on my server. I am having a problem with getting
>lilo to work on both
>mirrred disks hda and hdc. I am using the lilo from redhat6.1 which
>is version 21 with
>raid1 patch
>
>I have /boot on /dev/md0 swap on /dev/md1 and / on /dev/md2 and
>lilo.conf looks like:
...
