Good morning,
we are using PXE to boot into autoyast and install the packages from local
repositories via http and are having a problem since Suse released a new kernel
package in April.
Everything works fine up to the first reboot after finishing stage 1. Then grub
has two kernels to choose from: 2.6.31.12-0.1-default, which is the kernel of
the autoyast installation, and 2.6.31.12-0.2-desktop, which is the newest kernel
freshly installed from our repositories.
Grub boots the 2.6.31.12-0.1-default-kernel, complains (with good cause) 'unable
to canonicalize path
"/lib/modules/2.6.31.12-0.1-default/systemtap/preloadtrace.ko": No such file or
directory' and continues booting, but fails to load the modules for the network
card and the graphics card (and probably others). This leads to X not starting
and -worse- autoyast failing to load some configuration files via ftp,
configured like this:
<file>
/etc/somefile
root:root
644
ftp://somehost/files/etc/somefile
</file>
Autoyast does not complain very visibly about this failure while it's running.
The files not loaded do exist on the system afterwards, alas with a size of 0
bytes. The normal system logs do not contain that first boot, so it is kinda
hard to figure out exactly why the final system does not work like it should.
If I babysit the installation and manually choose the other kernel version
(2.6.31.12-0.2-desktop) when grub comes up after stage 1, everything continues
like it should, but I don't really want to wait to hit "down down return" at the
exact right time...
So: Why is the installation kernel still configured in grub when it is not fully
functional anymore? How do I get rid of it at this early time? Or is our error
something else entirely?
Thanks in advance,
Antje
--
Antje Bendrich (Team IT-Services), Phone +49 40 808077-642
DFN-CERT Services GmbH, https://www.dfn-cert.de/, Phone +49 40 808077-555
Sitz / Register: Hamburg, AG Hamburg, HRB 88805, Ust-IdNr.: DE 232129737
Sachsenstraße 5, 20097 Hamburg/Germany, CEO: Dr. Klaus-Peter Kossakowski