Re: new Etch install fails to boot
On 08/01/2007 07:23 AM, Steve Kleene wrote:
The motherboard (Abit BX133 440BX) has four IDE connectors allowing 8 drives
in total. hda-hdd have a speed of 33 MB/s max. hde-hdh have a speed of 66
MB/s max. So I have always had the two hard drives connected to hde and hdf.
There are no other hard drives. [...]
I think I smell a bug in the installer. Your HD configuration may be
"throwing it for a loop."
You probably will have to install Lilo manually--with minimal help from
the Debian installation and configuration system.
I used to have fun with this when I used Lilo in conjunction with
another Linux distro :-)
I forget which device the CD drive is, but
most of the IDE connectors are not connected to anything. Since I brought
this up 6 years ago, the MBR has been on hde and supported both Win98 and Red
Hat. This all worked until I tried to build Etch on hde (with no Windows).
It is a good idea that maybe grub is trying to read the MBR on the wrong disk
(hdf). I may have to break down and disconnect hdf before I resume. I may
also try to read the MBR with "od" if it's available just to see where this
string "GRUB Loading stage1.5: is coming from.
I'd suspect that the BIOS is trying to load the MBR from /dev/hde, so
Lilo or Grub needs to write its MBR there. Here are some WRONG places to
which the "boot-loader" might be writing:
/dev/hde1
/dev/hde2
/dev/hde3
/dev/hde5
/dev/hdf
/dev/hdf1
/dev/hdf2
...(anything on hdf)...
I did not try to edit the lilo.conf that was installed. It was really a
bare-bones file.
Thanks.
After you change it, always remember to do /sbin/lilo. And after you
change any file referenced by /etc/lilo.conf, always remember to do
/sbin/lilo.
Reply to: