On Sat, 2007-08-18 at 14:05 -0600, Karl Larsen wrote:
> Matthew Saltzman wrote:
> > On Sat, 2007-08-18 at 12:30 -0600, Karl Larsen wrote:
> >
> >> David Krings wrote:
> >>
> >>> [...]
> >>>
> >> Hi David, your not lunatic or anything. I just proved that Grub in
> >> F7 is buggy. I did this by doing what I had been trying all morning on
> >> F7 in one sinple thing on FC6. It works fine now :-)
> >>
> >
> > Out of curiosity Karl, what are the results of "rpm -q grub" on your FC6
> > and F7 systems?
> >
> >
> >> If you have an old FC6 on your computer try it there. It will work.
> >>
> >>
> >>
> >>
> Well they are both grub-0.97-13. So it is not the grub software and
> that fits my experiance. When I did the grub work on f7 it looked like
> it had worked but it didn't. I think there is a small problem and have
> no idea what it is. Makes writing a bug hard.
As others have pointed out, this seems like the behavior you'd expect
from the cylinder-1024 issue.
You described your installation at one point as having one large /
partition and no separate /boot. That's fine for most purposes, and
your nine-year-old BIOS may be able to *see* the disk with no problem,
but it may not be able to *read* the entire disk with no problem. So
when grub asks the BIOS to load data from the disk, that operation may
fail.
The usual workaround is to create a small (100MB or so) /boot partition
that holds grub and your kernels and ramdisks. This partition would be
near the front of your disk (below cylinder 1024), so all pieces that
the BIOS needs to be able to read are accessible.
Not doing this may result in unpredictable behavior, where grub works if
everything it tries to read happens to be below cylinder 1024, but fails
if not.
Newer BIOSes don't have this restriction, but BIOSes as old as yours
almost surely do.
>
>
--
Matthew Saltzman
Clemson University Math Sciences
mjs AT clemson DOT edu
http://www.math.clemson.edu/~mjs
