[4594] in testers
Re: linux 8.4.{24,25} -> 9.0.1
daemon@ATHENA.MIT.EDU (Greg Hudson)
Tue May 1 10:17:32 2001
Message-Id: <200105011417.KAA08941@egyptian-gods.MIT.EDU>
To: Abby Fox <ajfox@MIT.EDU>
cc: testers@MIT.EDU
In-Reply-To: Your message of "Tue, 01 May 2001 09:45:14 EDT."
<200105011345.JAA05229@whirlwind.mit.edu>
Date: Tue, 01 May 2001 10:17:21 -0400
From: Greg Hudson <ghudson@MIT.EDU>
> The first one doesn't look like any of the glitches mentioned in
> previous mails but let me know if I'm missing something (like,
> should I have updated to 8.4.25 first?).
No, it's my fault; I failed to release a volume necessary to make the
update to 9.0.1 work. (So it worked fine for me in testing, but not
for actual alpha testers.) Sorry about that, and thanks for providing
the update log. I've released the volume, so this shouldn't be a
problem in the future.
(The problem is that the update didn't freshen rpmupdate before going.
So you had all the rpmlib problems I'm so familiar with.)
I believe you can recover with:
Run /sbin/mkkerneldoth
Edit /etc/lilo.conf and changel kernel-2.2.17-14 to kernel-2.4.2-2
Run lilo
Run kudzu -q (not terribly important)
rpm -e kernel-2.2.17-14
Edit /etc/athena/version and change the last 8.4.24 to 9.0.1
Unfortunately, I don't have any ideas why rpmupdate didn't freshen (I
think that support was in 8.4.23 and higher), and the evidence is all
gone now. Hopefully this problem is restricted to bleeding machines.
> Update would break dependencies:
> Mesa = 3.2 is needed by Mesa-devel-3.2-2
Yeah, that's pretty clearly something Alex did (installed a version of
Mesa greater than 3.2, or locally removed Mesa). Installing Mesa at
3.2 should fix the problem.
