On Wed, Jun 1, 2011 at 3:54 PM, Scott Henson <shenson(a)redhat.com> wrote:
On Wed, 1 Jun 2011 09:25:14 +0200, Jörgen Maas
<jorgen.maas(a)gmail.com> wrote:
> On Tue, May 31, 2011 at 10:28 PM, Scott Henson <shenson(a)redhat.com> wrote:
> >
> > I'm curious how that worked before as it would have treated it like a
> > red hat distro right? I didn't think we handled suse in 2.x so you would
> > have had to masquerade as a red hat distro. What cobbler version did you
> > have this working on?
> >
>
> Yes, it would use the same code path as a redhat (or any other breed),
> if my understanding is correct:
>
> kickgen.py:
> 265 raw_data = utils.read_file_contents(kickstart_path,
> self.api.logger,
> 266 self.settings.template_remote_kickstarts)
> <snip>
> 276 data = self.templar.render(raw_data, meta, None, obj)
>
> I guess that code just serves the kickstart as a basic ascii file
> without any (non-Cheetah?) parsing/processing? And I don't see any
> problem with that approach, hence my question to Uwe what the purpose
> of this patch really is. From my point of view (at this moment) it
> really only adds complexity to the codebase and takes away flexibility
> at the autoyast XML level.
I think the point is that it starts providing better support for
autoyast. You start getting validation at the cobbler level and
such. Not sure if that is actually worth it or not.
I really dont know, do we do that for redhat or debian?
> We started using cobbler with version 2.0.9 and upgraded to 2.0.11, no
> problems in those two versions. We have an environment with about 400
> SuSE boxes (no redhat yet, working on that). So I guess it's possible
> we might do weird stuff in autoyast that developers could not have
> expected. But I really think if autoyast likes the XML file (and i
> know that autoyast validates the XML aswell), cobbler really should
> too.
I wonder what autoyast is using for the validation. I'd think it would
be python, but I've never used it so I really have no clue.
I think it's c++ code.
Also, 2.1.0 is a development version. So we can definitely discuss what
the proper thing to do for SUSE is. I don't really have a horse in the
race as I don't use SUSE. I think this would really need to be discussed
on list (note that the list and Uwe have dropped off the CC list at some
point) and some conclusion come to. Also, from the autoreply I get from
Uwe, he seems to be out of the office currently. I'm sure he will weigh
in it when he gets back.
My bad, i included Uwe and the list again. Sorry about that!
> > There should be some additional information in
> > /var/log/cobbler/cobbler.log. Generally when you get an ISE from apache,
> > you'll see an exception in cobbler.log.
> >
>
> Good point, I should have included that in my first mail.
> Here's the traceback:
>
> Fri May 27 15:21:04 2011 - INFO | REMOTE generate_kickstart; user(?)
> Fri May 27 15:21:04 2011 - INFO | generate_kickstart
> Fri May 27 15:21:04 2011 - INFO | autoyast XML file found. Checkpoint:
> profile=<cobbler.item_profile.Profile instance at 0xe291b90>
> system=<cobbler.item_system.System instance at 0xe293998>
> Fri May 27 15:21:04 2011 - INFO | Exception occured:
> xml.parsers.expat.ExpatError
> Fri May 27 15:21:04 2011 - INFO | Exception value: junk after document
> element: line 10, column 0
Is there anything around line 10 that rings a bell? I'm thinking that it
might start xml validation before the cheetah templates have been
expanded? Do you think I could get a look at your autoyast template
file? Thanks.
Here's the first few lines of the autoyast XML
<?xml version="1.0"?>
<!DOCTYPE profile>
<profile
xmlns="http://www.suse.com/1.0/yast2ns"
xmlns:config="http://www.suse.com/1.0/configns">
<add-on>
<add_on_products config:type="list">
<listentry>
<media_url><![CDATA[http://xxx.yy.zzz/repos/DVD/SLES-11.1-SDK-x86_64/]]></media_url>
<product>SDK</product>
<product_dir>/</product_dir>
</listentry>
<listentry>
<media_url><![CDATA[http://xxx.yyy.zzz/repos/xxx/SLE_11_SP1/]]></media_url>
<product>xxx</product>
<product_dir>/</product_dir>
</listentry>
</add_on_products>
</add-on>
Perhaps I've overlooked something here...
Regards,
Jorgen Maas