> > lilo.conf
> > -----------
> > disk=/dev/md0
> > # fdisk of hda
> > bios=0x80
> > sectors=63
> > heads=16
> > cylinders=39770
> > partition=/dev/md1
> > # from fdisk -ul -- starts at block 408
> ^^^^^ you probably mean cylinder :)
> > start=410256
> >
> > boot=/dev/hda
> > delay=50
> > vga=normal
> >
> > image=/boot/bzImage
> > root=/dev/md0
> > label=Linux
> > read-only
> >
> > -----------------------
> > this all looks OK to me
>
> To mee too, but since I've moved to hw raid now, and RedHat started
> shipping a lilo version that can handle sw raid1, I haven't been
> looking at it all that hard lately.
>
> > lilo complains: don't know how to handle device 0x0901
>
> Could there be a problem with lilo not being able to handle that
> many cylinders? If you just lie to it and set the number of
> cylinders to 1024 or so, does that help? Other than that, I cannot
> really come up with any good ideas without digging into the lilo
> source... :)
>
There was a 200 meg ext2 partition I was using for development as the
first partition. md1 was the second (16megs).
When I re-arranged the partitions to put md1 in front (i.e. start-63)
all was well. Me-thinks that lilo is not all that happy with the
workaround for geometry. I'm pretty sure the start partition
parameter is correct since it is reported by fdisk -ul
Oh well.... works slick now.
[EMAIL PROTECTED]
