Re: Etch won't load after eight attempts
On 06/03/2008 06:47 PM, Ken Heard wrote:
On 26 March 2008 I purchased a computer with an Intel dual core CPU and
a SATA hard drive. Since then I have off and on spent in the aggregate
77.94 hours trying to install and configure Etch on it using the third
revision of net-install disk.
Whoa!
I have so far made eight attempts to get a working system using the
Debian installer. Each time something happened to make the installation
unusable, and I could not find any way to fix each problem without
starting the installation all over again.
What happened earlier today is more or less typical of the what I have
been up against. I was able to complete the eighth installation with
the xfs file system and KDE. On the recommendation of the installer I
used the LILO boot loader instead of GRUB.
It is absolutely necessary that you use XFS? AFAIK, XFS and Grub don't
mix very well. That's probably the reason why the installer recommended
LILO; however, I doubt that installation with LILO is as well debugged
as installation using Grub.
One of the first things I did by way of configuration today was to run
aptitude update and then to run aptitude upgrade. On of the packages to
be upgraded was the kernel (2.6.18-6-686). As part of the configuration
of this package by aptitude, I was told that this was the same kernel
already installed and that consequently I would have to reboot as soon
as the upgrade was finished.
I would suggest rebooting as soon as the base system is installed. You
need to know that booting works as soon as possible, and the base system
should be a bootable system.
I did so, but the reboot ended when the machine hung after returning
the following the following message:
"LILO 22.6.1 Loading EBDA is big; kernel setup stack overlaps LILO
second stage"
A wild guess is that you only need to do "/sbin/lilo" after you've
managed to boot the system. Use a rescue disk that will let you chroot
into your partition. Etch CD#1 allows you to do this (it's saved me a
couple of times).
I also recommend getting a Knoppix disk for situations like these.
What does this message mean? What if anything can I do to get the
machine working again? (Subsquent boots returned the same message after
booting?)
Or is my only option to attempt a ninth installation and hope for the best?
Ken Heard
Re-read the LILO documentation--especially the part where it says
"Always do /sbin/lilo."
LILO is not like Grub. LILO can't read paritions--only disk sectors.
When LILO is installed (/sbin/lilo), it creates list of sectors (a
sector-map) to read upon bootup. Any changes to any of the files used by
lilo require that the sector-map be recreated (/sbin/lilo).
Whenever you change any aspect of LILO's configuration, or whenever you
make any changes to the kernel, "always do /sbin/lilo."
Good luck.
Reply to: