[Spacewalk-list] Problem kickstarting using a Software Package Profile

[Jon Miller]

I'm inching closer to my problem / solution. What I have realized is that
the conflicting packages are actually being located from the CentOS yum
repositories on the internet. My channels are currently locked down and
sync to CentOS 6.2 whereas the default yum repos files that are distributed
via the centos-release package are pointing to the latest set of packages.

An equivalent question to solve my problem may be: How do you disable local
yum repos from conflicting with what channels you have configured in
Spacewalk?
   I currently have a post-chroot script setup in my kickstart profile that
renames any /etc/yum.repos.d/*.repo file with a ".off" extension but
apparently this isn't quick enough. Is there a more effective way in
disabling local repos quicker?

Thanks,
Jon Miller


On Thu, Apr 18, 2013 at 5:44 PM, Jon Miller <jonebird at gmail.com> wrote:

> I have been developing a CentOS 6 profile lately and toying between
> managing the versioning of the builds with either cloned channels or
> captured Package Profiles.
>
> Here is my test scenario for the later approach:
> - Created a clean Software Profile via
>   a. Kickstart a fresh machine.
>   b. Issue a single "yum -y update"
>   c. Create a Package Profile entitled "CentOS 6.2 - Updated"
>   d. Cloned kickstart profile "centos6-base" -> "centos6-base-updated"
>   e. Updated the "centos6-base-updated" profile's software tab to use the
> package profile
> - Now testing a fresh kickstart
>   + Broke out of the default PXE menu entry
>   + Updated the ks= line to reference my cloned kickstart profile
> - Result
>   + Viewing the Events History of the new machine shows a failed "Package
>     Synchronization scheduled by (none)"
>   + Looking at the error message in detail I see:
>     #+begin_example
> Client execution returned "Failed: packages requested raised dependency
> problems: [u'Protected multilib versions: nspr-4.8.9-3.el6_2.x86_64 !=
> nspr-4.9.2-1.el6.x86_64', u'Protected multilib versions:
> nss-3.14.0.0-12.el6.x86_64 != nss-3.13.1-7.el6_2.i686', u'Protected
> multilib versions: nss-3.14.0.0-12.el6.x86_64 !=
> nss-3.13.1-7.el6_2.x86_64', u'Protected multilib versions:
> nss-util-3.13.1-3.el6_2.i686 != nss-util-3.14.0.0-2.el6.x86_64',
> u'Protected multilib versions: libffi-3.0.5-3.2.el6.i686 !=
> libffi-3.0.10-2.el6.x86_64']" (code 18)
>     #+end_example
>
> The list of failed packages is exactly the number of packages I would get
> installed if I did a "yum update" on the machine right now.
>
> Any ideas as to what is wrong here? I'd also take advice on whether or not
> you think this approach is a sane one or not.
>
> Thanks,
> Jon Miller
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/spacewalk-list/attachments/20130419/a45040f7/attachment.htm>