Bug 717462

Summary: RHEL 6 harness repo is missing python-setuptools package
Product: [Retired] Beaker Reporter: Prudhvi Tella <prudhvi_tella>
Component: lab controllerAssignee: Bill Peck <bpeck>
Status: CLOSED NOTABUG QA Contact:
Severity: medium Docs Contact:
Priority: unspecified    
Version: 0.6CC: bpeck, dcallagh, mcsontos, rmancy, stl
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-07-18 22:21:21 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Prudhvi Tella 2011-06-28 21:12:49 UTC
Description of problem:
Installation of beah on provisioned RHEL 6 system fails due to missing pyton-setuptools package in the harness repo.

Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
1.Add RHEL 6 repo to cobbler
2.Just add RedHatEnterpriseLinux6Server harness repo using beaker-repo-update
3.Provision a system
  
Actual results:
21:17:12,895 ERROR   : Error: Package: beah-0.6.27.1-1.el6eso.noarch (beaker-harness)
21:17:12,895 INFO    :  You could try using --skip-broken to work around the problem
21:17:12,899 ERROR   :            Requires: python-setuptools
21:17:14,262 INFO    :  You could try running: rpm -Va --nofiles --nodigest


Expected results:
beah to install

Additional info:

Comment 1 Dan Callaghan 2011-06-28 22:00:49 UTC
We recently stopped shipping python-setuptools (and other similar dependencies) in the Beaker repos because these packages are available in the distro itself and should be installed from there.

What this has revealed is that very often distros are not being imported correctly and the system ends up with wrong/missing yum configuration after install. I suspect that is the case again here.

Can you please link to the job where you saw this and we can investigate further.

Comment 2 Dan Callaghan 2011-06-28 22:17:41 UTC
Sorry, it just occurred to me that this problem is probably occurring on a non-Red Hat instance of Beaker, so I guess you won't be able to link to the job in question.

Nevertheless I think the problem is with the distro. To verify, check the ks.cfg file which is uploaded to Beaker after the system is provisioned. This is the kickstart file which was generated by Cobbler. You should find a section that looks like this:

# Add Distro Repos
cat << EOF > /etc/yum.repos.d/beaker-Server.repo
[beaker-Server]
name=beaker-Server
baseurl=http://example/RHEL/6/Server/x86_64/os/Server
enabled=1
gpgcheck=0
skip_if_unavailable=1
EOF
...

These correspond to the 'tree_repos' kickstart variable defined in cobbler. If that variable is not set, these distro repos will not be created and beah will fail to be installed.

Comment 3 Prudhvi Tella 2011-06-29 14:52:03 UTC
tree_repos variable is not defined as you suspected. Modifying the documentation to suggest users to define tree_repos variable in ks_meta will solve this issue.

Comment 4 Dan Callaghan 2011-06-29 22:30:30 UTC
The tree_repos variable in Cobbler is supposed to be set by this script:

/var/lib/cobbler/triggers/sync/post/osversion.trigger

which should be run by Cobbler after new distros have been imported. It uses some heuristics to find the yum repos in a distro, but it's not perfect. Something must have gone wrong with it in this case.

Comment 5 Marian Csontos 2011-07-15 11:38:10 UTC
Is this sorted out? Can we close the bug?

Comment 6 Prudhvi Tella 2011-07-15 19:06:47 UTC
My issue has been addressed. It can be closed.

Comment 7 Marian Csontos 2011-07-18 07:13:02 UTC
(In reply to comment #4)
> The tree_repos variable in Cobbler is supposed to be set by this script:
> 
> /var/lib/cobbler/triggers/sync/post/osversion.trigger
> 
> which should be run by Cobbler after new distros have been imported. It uses
> some heuristics to find the yum repos in a distro, but it's not perfect.
> Something must have gone wrong with it in this case.

Dan, is this problem in (LC|Cobbler) something to act on?

Comment 8 Dan Callaghan 2011-07-18 22:21:21 UTC
Let's assume for now that this was caused by a configuration error.

Prudhvi, if you have more information indicating that there is a bug in the osversion.trigger script or that the documentation should be changed, please re-open or file a new bug.