Bug#413311: linux-image-2.6.18-4-686: Install routine fails due to readlink error
severity 413311 serious
reassign 413311 linux-2.6
clone 413311 -1
reassign -1 release-notes
retitle -1 undeclared linux dependency on etch coreutils, affects upgrade path?
thanks
On Sat, Mar 03, 2007 at 09:10:19PM -0800, Todd A. Jacobs wrote:
> Attempting to do an upgrade produces the following error:
> Setting up linux-image-2.6.18-4-686 (2.6.18.dfsg.1-11) ...
> Running depmod.
> Finding valid ramdisk creators.
> Using mkinitrd.yaird to build the ramdisk.
> readlink: invalid option -- m
> Try `readlink --help' for more information.
> readlink: invalid option -- m
> Try `readlink --help' for more information.
> Failed to symbolic-link initrd.img-2.6.18-4-686 to initrd.img.
> dpkg: error processing linux-image-2.6.18-4-686 (--configure):
> subprocess post-installation script returned error exit status 17
> Errors were encountered while processing:
> linux-image-2.6.18-4-686
> E: Sub-process /usr/bin/dpkg returned an error code (1)
> A package failed to install. Trying to recover:
> A quick check of the man page for readlink does indeed show that there
> is no such option. So, the kernel install is left in a partial state,
> and attempting to remove linux-image-2.6.18-4-686 with aptitude attempts
> to remove other packages as well, such as linux-image-686.
This is due to a change introduced in kernel-package 10.063 back in October.
The missing dependency was fixed in kernel-package 10.065 in November, but
unfortunately the linux-2.6 kernel image packages don't use the dependency
information provided by kernel-package, and the need for a dependency update
was not communicated to the kernel team or the release team (the latter
because bug #396077 was filed at the wrong severity :/).
Having new dependencies added to the kernel package has the potential to
affect the upgrade path from sarge as a whole, so I'm cloning this bug over
to the release-notes for examination. I'm pretty confident that this
particular dependency won't hurt us, though -- coreutils shouldn't pull in
any other packages besides a few core libs, and the new glibc is already an
indirect dependency of the kernel images. Even so, we probably need for
this to be double-checked before release.
The timing of this bug report is also unfortunate, in that the hope had been
for the current linux-2.6 package to be the final one for etch r0. OTOH,
this bug has apparently been latent in testing since at least December 9th
without being reported before, so perhaps it's ignorable for r0 anyway?
FWIW, a simple workaround for this bug should be to upgrade coreutils and
then retry the install of the kernel. No removal of packages should be
necessary here.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
vorlon@debian.org http://www.debian.org/
Reply to: