Re: What are the two roots in grub?

[hendrik@topoi.pooq.com]

On Mon, Feb 19, 2007 at 01:15:58PM -0500, Greg Folkert wrote:
> On Mon, 2007-02-19 at 12:46 -0500, hendrik@topoi.pooq.com wrote:
> > Do I understand this correctly?
> > 
> > In my /boot/grub/menu.lst there is a stanza
> > 
> > title		other: Debian GNU/Linux, kernel 2.6.18-3-486
> > root		(hd0,7)
>   ^^^^
> This particular "root" is the "grub installation root" IOW it will load
> the following line like this
> 
> (hd0,7)/vmlinuz-2.6.18-3-486 root=/dev/mapper/lovesong-other ro
> 
> > kernel		/vmlinuz-2.6.18-3-486 root=/dev/mapper/lovesong-other ro 
>                                         ^^^^
> That root is the "root" of your machine. and it is passed onto your
> kernel as a command line option.

I see... so (hd0,7) is the partition where grub looks up all the 
unadorned filenames bentioned in the stanza.  And the other root is the 
root of the filesystem in the to-be-booted system.
> 
> > initrd		/initrd.img-2.6.18-3-486
> > savedefault
> 
> 
> Your re-write is bad. 2.6.8 will not work after you have installed etch
> on the same "logical volume". The udevd you have installed now requires
> a kernel version of 2.6.14 or newer.

But the etchtobe system is at present just a copy of the sarge system, 
with the necessary changes to fstab.  So it presumably has the same udev 
as the sarge system, and the same kernel.  There may be trouble upgrading 
to etch, but that's what I'm setting up to do.

> 
> Hendrik, I have to say you are making things TOOO complicated for your
> self. I have seen your posts over that past few months and not
> understanding why you even put yourself through all of this.

I'm discovering it's partly because of an undiagnosed failing hard disk, 
which made things that should be easy hard.  I've swapped drives and am 
now trying to copy the functioning system from the failing to a good 
hard drive.

I also had trouble with the sarge-to-etch upgrade some months ago.  I 
had copied the sarge system when sarge went stable, and was keeping 
one copy stable and the other testing.  It survied the xfree->xorg 
transition, but did not survive the transition to a modular xorg.
 My users were quite happy to still have a sarge to fall back on, 
although up to then they had been using etch.

So having multiple systems is mostly a matter of having something to 
fall back on when necessary -- something possible with Debian but not 
with Windows, by the way.

> 
> grub is not that hard to understand. If you have it installed in
> multiple locations *AND* the MBR then you have created yourself a myriad
> of problems.

grub just on the MBR of /dev/hda.

It's lilo that was on a variety of floppy disks -- as a place to try out 
lilo configurations without damaging the MBR, and as a an emergency boot 
device in case I installed Windows or something.

I haven't digured out how th install grub in a floppy disk yet, by the 
way.

> You seem to want to try everything on a single machine...
> not a good idea. Get a cheap machine and do all your blatherskite on
> that machine. Then after "getting" it... do in on your primary machine
> you use for doing stuff.
> 
> Good luck pulling yourself out of this.

Thanks.  But the sarge system my users are using is still running -- 
from the failing hard drive.  And they're happy.

-- hendrik