all worked fine untill I rebooted the machine and...ooops! The boot loader LILO failed with the "LI" code error. No way to boot the machine. I solved the problem with a knoppix live cd, I mounted the root partition, chrooted into it, mounted /boot partition and run "lilo". After reboot the server was up&running.
But what did it happen? I think there is a bug somewhere.
# /etc/lilo.conf - See: `lilo(8)' and `lilo.conf(5)', # --------------- `install-mbr(8)', `/usr/share/doc/lilo/', # and `/usr/share/doc/mbr/'.
# +---------------------------------------------------------------+ # | !! Reminder !! | # | | # | Don't forget to run `lilo' after you make changes to this | # | conffile, `/boot/bootmess.txt', or install a new kernel. The | # | computer will most likely fail to boot if a kernel-image | # | post-install script or you don't remember to run `lilo'. | # | | # +---------------------------------------------------------------+
# Support LBA for large hard disks. # lba32
# Overrides the default mapping between harddisk names and the BIOS' # harddisk order. Use with caution. #disk=/dev/hde # bios=0x81
#disk=/dev/sda # bios=0x80
# Specifies the boot device. This is where Lilo installs its boot # block. It can be either a partition, or the raw device, in which # case it installs in the MBR, and will overwrite the current MBR. # boot=/dev/sda
# Specifies the device that should be mounted as root. (`/') # root=/dev/sda4
# Enable map compaction: # Tries to merge read requests for adjacent sectors into a single # read request. This drastically reduces load time and keeps the # map smaller. Using `compact' is especially recommended when # booting from a floppy disk. It is disabled here by default # because it doesn't always work. # # compact
# Installs the specified file as the new boot sector # You have the choice between: bmp, compat, menu and text # Look in /boot/ and in lilo.conf(5) manpage for details # install=/boot/boot-menu.b
# Specifies the location of the map file # map=/boot/map
# You can set a password here, and uncomment the `restricted' lines # in the image definitions below to make it so that a password must # be typed to boot anything but a default configuration. If a # command line is given, other than one specified by an `append' # statement in `lilo.conf', the password will be required, but a # standard default boot will not require one. # # This will, for instance, prevent anyone with access to the # console from booting with something like `Linux init=/bin/sh', # and thus becoming `root' without proper authorization. # # Note that if you really need this type of security, you will # likely also want to use `install-mbr' to reconfigure the MBR # program, as well as set up your BIOS to disallow booting from # removable disk or CD-ROM, then put a password on getting into the # BIOS configuration as well. Please RTFM `install-mbr(8)'. # # password=tatercounter2000
# Specifies the number of deciseconds (0.1 seconds) LILO should # wait before booting the first image. # delay=20
# You can put a customized boot message up if you like. If you use # `prompt', and this computer may need to reboot unattended, you # must specify a `timeout', or it will sit there forever waiting # for a keypress. `single-key' goes with the `alias' lines in the # `image' configurations below. eg: You can press `1' to boot # `Linux', `2' to boot `LinuxOLD', if you uncomment the `alias'. # # message=/boot/bootmess.txt prompt timeout=150 # prompt # single-key # delay=100 # timeout=100
# Specifies the VGA text mode at boot time. (normal, extended, ask, <mode>) # # vga=ask # vga=9 # vga=normal
# Kernel command line options that apply to all installed images go # here. See: The `boot-prompt-HOWO' and `kernel-parameters.txt' in # the Linux kernel `Documentation' directory. # # append=""
# If you have another OS on this machine to boot, you can uncomment the # following lines, changing the device name on the `other' line to # where your other OS' partition is. # # other=/dev/hda4 # label=HURD # restricted # alias=3 other=/dev/sda1 label="Other(sda1)"
I made a typescript's upgrade too but I don't think it's notable, of course you'll can ask me it.
I would know what packet is involved in the accident so I can fill a bug ticket. Thank you.
Michele
-- To UNSUBSCRIBE, email to debian-user-REQU...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
> all worked fine untill I rebooted the machine and...ooops! The boot > loader LILO failed with the "LI" code error. > No way to boot the machine. I solved the problem with a knoppix live > cd, I mounted the root partition, chrooted into it, mounted /boot > partition and run "lilo". > After reboot the server was up&running.
> But what did it happen? I think there is a bug somewhere. > [...] > I made a typescript's upgrade too but I don't think it's notable, of > course you'll can ask me it.
lilo should have detected it needed to be re-run and should have prompted you about this, unless it was configured otherwise.
What does | # grep -A5 'lilo/runme' /var/cache/debconf/config.dat say?
On Thu, Jun 09, 2005 at 08:08:08PM +0200, Florian Ernst wrote: > Hello *,
> > all worked fine untill I rebooted the machine and...ooops! The boot > > loader LILO failed with the "LI" code error. > > No way to boot the machine. I solved the problem with a knoppix live > > cd, I mounted the root partition, chrooted into it, mounted /boot > > partition and run "lilo". > > After reboot the server was up&running.
> > But what did it happen? I think there is a bug somewhere. > > [...] > > I made a typescript's upgrade too but I don't think it's notable, of > > course you'll can ask me it.
> lilo should have detected it needed to be re-run and should have > prompted you about this, unless it was configured otherwise.
> What does > | # grep -A5 'lilo/runme' /var/cache/debconf/config.dat > say?
Why do people insist on using LILO? Just install grub, set it up once with symlinks and *never* mess with it again. In case anyone cares, here is my configuration:
default 0 timeout 5 color cyan/blue white/blue title Debian GNU/Linux root (hd0,1) kernel /vmlinuz root=/dev/hda3 ro vga=normal ide0=ata66 ide1=ata66 boot
title Debian GNU/Linux (recovery mode) root (hd0,1) kernel /vmlinuz root=/dev/hda3 ro single vga=normal ide0=ata66 ide1=ata66 boot title Debian GNU/Linux, OLD root (hd0,1) kernel /vmlinuz.old root=/dev/hda3 ro vga=normal ide0=ata66 ide1=ata66 boot
title Debian GNU/Linux, OLD (recovery mode) root (hd0,1) kernel /vmlinuz.old root=/dev/hda3 ro single vga=normal ide0=ata66 ide1=ata66 boot
I installed it *one* time to the MBR. I have only since had to mess with it to replace the failed harddrive in the machine. If you have /boot on a separate partition, as I do, you will want to add 'do_symlinks = No' to /etc/kernel-img.conf. Trust me on this. It is really nice to have a boot loader that can actually read your filesystem.
Hm, strange, to me it looks like this question was never asked. I guess it's best to ask the lilo maintainers <l...@packages.debian.org> about this, or you could send a bugreport against upgrade-reports...
I personally don't use lilo myself, so I wouldn't dare venture any deeper into this problem.
Florian Ernst wrote: > MicheleM wrote: > > all worked fine untill I rebooted the machine and...ooops! The boot > > loader LILO failed with the "LI" code error. > > [...] > > But what did it happen? I think there is a bug somewhere.
> lilo should have detected it needed to be re-run and should have > prompted you about this, unless it was configured otherwise.
> What does > | # grep -A5 'lilo/runme' /var/cache/debconf/config.dat > say?
I upgraded a machine running lilo and debconf mailed me the following information:
Subject: Debconf: Configuring lilo -- Deprecated parameters in LILO configuration
Deprecated files have been found on your system. You must update the 'install=' parameter in your LILO configuration file (/etc/lilo.conf) in order to properly upgrade the package.
The new 'install=' options are:
new: install=bmp old: install=/boot/boot-bmp.b
new: install=text old: install=/boot/boot-text.b
new: install=menu old: install=/boot/boot-menu.b or boot.b
-- Debconf, running at discord.proulx.com [ Debconf was not configured to display this note, so it mailed it to you. ]
Reading that message I took care of the problem (by installing grub) before rebooting the system. But I can't help but wonder if something similar occurred with the original poster? It was mailed to me because I had set my debconf to 'critical' only and so it was prevented from showing me a dialog box. Although for something like this it would justify a critical rating.
This topic may be a few days old, but this is apparently exactly what happened to a server I admin remotely. When I upgraded to the new stable packages, I saw nothing that affected the kernel, therefore rerunning LILO didn't occur to me as something that needed to be done.
Hopefully someone reported a bug report on this against LILO since Monday. Since no changes to the kernel should have affected LILO's boot configuration, it should have had no trouble booting up.
Why do I have LILO running in the first place? When I installed from the Woody CD a year ago, that's all that was offered to me. I wasn't aware, at the time, of the problems known to occur with LILO and so didn't install grub. Now that I live a couple hundred miles from this machine, but continue to admin it, I wouldn't dare attempt to install a new boot loader unless I was physically present.
Clinton
-- To UNSUBSCRIBE, email to debian-user-REQU...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
This topic may be a few days old, but this is apparently exactly what happened to a server I admin remotely. When I upgraded to the new stable packages, I saw nothing that affected the kernel, therefore rerunning LILO didn't occur to me as something that needed to be done.
Hopefully someone reported a bug report on this against LILO since Monday. Since no changes to the kernel should have affected LILO's boot configuration, it should have had no trouble booting up.
Why do I have LILO running in the first place? When I installed from the Woody CD a year ago, that's all that was offered to me. I wasn't aware, at the time, of the problems known to occur with LILO and so didn't install grub. Now that I live a couple hundred miles from this machine, but continue to admin it, I wouldn't dare attempt to install a new boot loader unless I was physically present.
Clinton
-- To UNSUBSCRIBE, email to debian-user-REQU...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
This topic may be a few days old, but this is apparently exactly what happened to a server I admin remotely. When I upgraded to the new stable packages, I saw nothing that affected the kernel, therefore rerunning LILO didn't occur to me as something that needed to be done.
Hopefully someone reported a bug report on this against LILO since Monday. Since no changes to the kernel should have affected LILO's boot configuration, it should have had no trouble booting up.
Why do I have LILO running in the first place? When I installed from the Woody CD a year ago, that's all that was offered to me. I wasn't aware, at the time, of the problems known to occur with LILO and so didn't install grub. Now that I live a couple hundred miles from this machine, but continue to admin it, I wouldn't dare attempt to install a new boot loader unless I was physically present.
Clinton
-- To UNSUBSCRIBE, email to debian-user-REQU...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
