[Spacewalk-list] CentOS 5.2 - a warning]

[John Hodrien]

On Wed, 22 Apr 2009, m.roth2006 at rcn.com wrote:

>>> Well,yes, they are. I know what the 686's are. The big thing was this: when
>>> they were in the Spacewalk repository, the query that Spacewalk did on its
>>> d/b got them uniquely. So when I told it to upgrade, it did it with those
>>> packages. Then, when I went to reboot, that failed coming back up, with the
>>> error "request_module: runaway loop modprobe binfmt-464c" repeated three or
>>> so times.
>>
>> That's a bug in the multiarch handling of spacewalk then.
>>
> That bug is in the 686 being installed. It was the install of the 686 that
> killed everything.

This is where I reckon you're wrong.  It's not the install of the i686
package, it's the uninstall (that I'm guessing has happened) of the x86_64
package that caused the problem.  But I really am guessing.

>> Yep, I had noted that...  I really wouldn't run against 0.4 out of choice.
>> I used 0.1/2/3/5.  0.4 was fun for testing.
>
> Not long after I walked in the door where I've been contracting, they told
> me to do Spacewalk. I went to the site, and 0.4 was the latest, greatest,
> and stable version. I saw nothing that suggested it was a development
> version, like the way the Linux kernel x.<odd> == dev, and x.<even> ==
> stable.

I'd argue there hasn't been a 'stable' release of spacewalk yet, they're all
devel releases.  If you're not willing to get mucky, you shouldn't be using
spacewalk.  If you'd been told to use spacewalk by people who didn't get that,
it doesn't change things.

I'm not providing any official word on this because I'm just somebody who uses
spacewalk.  0.4 changed lots (integrating cobbler) and broke lots (my
opinion).  0.5 was then intended to be a fix of all the bugs in 0.4.  I stuck
with 0.3 (which worked for me) rather than upgrade to 0.4 (which didn't).  0.5
looks like it'll be worth an upgrade, so I'm now in the process of doing this.

> Then you and I have a disagreement. I *know* that when I did the upgrade,
> and got the 686 packages installed, it failed to successfully boot, and when
> I removed them, it booted with no issues at all. As far as I'm concerned,
> that is proof that it should not have been installed. That makes it a CentOS
> problem, not a Spacewalk problem, since they have packages in one repository
> that should not be there.  <snip>

But you said you installed kernel.i686 packages, which aren't in x86_64.  I
point you to my stock 5.2 installed desktop machine:

$ rpm -q --qf %{NAME}-%{ARCH}\\n glibc
glibc-x86_64
glibc-i686

Having i686 packages installed does not destroy your machine and make the
world cave into a black hole.  This is just something you *have* to accept.

I can't tell you what broke your setup from a distance, all I can tell you is
that you're wide of the mark blaming the i686 packages that /are/ part of
x86_64 CentOS and *should* be part of x86_64 CentOS.  If these were i386
packages instead, they would be functionally identical.  This should make
sense.  If it doesn't, ponder the difference between an i386 and an i686 RPM.

I hope you're not about to suggest that installing i386 packages on a x86_64
box is wrong, else we're completely lost.

jh

-- 
"Waste no more time arguing what a good man should be. Be one."
                                                      -- Aurelius