After a standard security/stability upgrade of lenny I witnessed the same problem as Xavier, namely that lilo would stop with:

  liloLoading LinuxEBDA is big; kernel setup stack overlaps LILO second
  stage

(I am not sure whether there were any points after "lilo". I think not,
 but if there were, then there were only very few - i.e. not a line or
 half a line or such)

What I did exactly was upgrading from linux-image-2.6.26-2-686 2.6.26-19lenny1 to 2.6.26-21 inside aptitude and *at the same time* deinstalling the old linux-image-2.6.18-6-686 2.6.18.dfsg.1-26etch1.

Aptitude did what I told it to do and did not complain, and ended without error. The only thing indicating a problem were the following lines (from /var/log/apt.log - since I did not catch them during the down/upgrade, and translated from czech, since that's how the locale is set):

  Preparing to replace linux-image-2.6.26-2-686 2.6.26-19lenny1 (with
  .../linux-image-2.6.26-2-686_2.6.26-21_i386.deb) ...
  The directory /lib/modules/2.6.26-2-686 still exists. Continuing as
  directed.
  Done.
  Unpacking replacement linux-image-2.6.26-2-686 ...
  ...
  Configuring package linux-image-2.6.26-2-686 (2.6.26-21) ...
  Running depmod.
  Running mkinitramfs-kpkg.
  Not updating initrd symbolic links since we are being updated/reinstalled
  (2.6.26-19lenny1 was configured last, according to dpkg)
  Not updating image symbolic links since we are being updated/reinstalled
  (2.6.26-19lenny1 was configured last, according to dpkg)
  ...
  Removing package linux-image-2.6.18-6-686
  The link /vmlinuz.old is a damaged link
  Removing symbolic link vmlinuz.old
  Unless you used the optional flag in lilo,
   you may need to re-run your boot loader[lilo]
  The link /initrd.img.ol is a damaged link
  Removing symbolic link initrd.img.ol
  Unless you used the optional flag in lilo,
   you may need to re-run your boot loader[lilo]
  Removing configuration directory of package linux-image-2.6.18-6-686 ...
  ...

And that's it, no errors anywhere.

After upgrading I rebooted and ended up with the lilo error reported by Xavier and repeated above.

Strangely enough (!!!) I could *still* boot into the removed (!) 2.6.18-6 kernel and allthough the bootprocess complained about a lot of missing kernel modules (or modules.dep ? I don't know, I can't find it anywhere in the logs), but somehow miraculously (!) it was able to install the networking stack... Then, from the shell I issued:

  # lilo
  Fatal: open /boot/vmlinuz-2.6.18-6-686

After that I reinstalled the old linux-image-2.6.18-6-686 2.6.18.dfsg.1-26etch1 package and rebooted.

Now everything was fine and I could boot into the new 2.6.26 kernel without lilo complaining.

Running lilo from the command line would show me the bootable kernels:

  # lilo
  Added 2618-6-hda5
  Added 2626-hda5
  Added windows

  # ls -l /
  [...]
  initrd.img -> boot/initrd.img-2.6.18-6-686
  initrd.img.old -> boot/initrd.img-2.6.26-2-686
  [...]
  vmlinuz -> boot/vmlinuz-2.6.18-6-686
  vmlinuz.old -> boot/vmlinuz-2.6.26-2-686

It seems that even if lilo.conf is half broken - i.e. it's missing some of the kernels, the install process will nevertheless install the new kernel but somehow break halfway through the process and not properly install the new kernel.

*t




--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to