Build server requires calling cobbler sync before deploying nodes

[subconscious]

After multiple installs of a build server: each time an additional "cobbler sync" needs to be executed in order for dnsmasq to offer a lease.

Associated errors:

Fri Apr 26 18:18:23 2013 - INFO | Exception Info:
  File "/usr/lib/python2.7/dist-packages/cobbler/remote.py", line 1944, in _dispatch
    return method_handle(*params)
   File "/usr/lib/python2.7/dist-packages/cobbler/remote.py", line 873, in xapi_object_edit
    handle = self.get_item_handle(object_type, object_name)
   File "/usr/lib/python2.7/dist-packages/cobbler/remote.py", line 655, in get_item_handle
    raise CX("internal error, unknown %s name %s" % (what,name))

[Chris Ricker (chris-ricker)]

We'll probably need more steps to reproduce

In various light testing (start from bare, add more nodes, remove nodes, add nodes back) I can't get to the state of node existing cobbler but dnsmasq not aware of it / offering lease, or produce the above exception

[subconscious (subconscious)]

I confirmed this again this morning on a virgin install. The only step I needed to correct it was to run cobbler sync on the build server then rerun ./clean-server control-server. I am thinking that instead of using

before => Service[cobbler], we use "require". I'll test this out and post results.

[subconscious (subconscious)]

https://github.com/CiscoSystems/puppet-cobbler/pull/45

This modification appears to fix it. I tested it as working now. Basically, I dumped the package requirement and required the service instead of using before.

[Abdi Ibrahim (abdi-w)]

I'm running g.2 release for Grizzly and this issue still exist. After multiple install of the build-server you'll need to run 'cobbler sync' in order the PXE boot process to start and working. I'm assuming that this problem still exist.