[lug] kickstart and group files

[Hugh Brown]

I believe kickstart is just anaconda.  So anaconda is the next place to look.  I think you can also set this in kickstart "logging --level=debug" to see if it will give you more info (and a Ctrl-Alt-F2 to a shell after it fails).

I agree that it would be nice to know what's necessary to make this work.

General thoughts: does your new comps.xml validate against the DTD, you can turn up logging on the Apache to see which files are being grabbed, you can sniff the traffic between the client and the server to see if you get a list of files.

Hugh


On Wed, Mar 26, 2008 at 12:33:58PM -0600, Jeffrey Haemer wrote:
> Hugh,
> 
> Thanks!  There's also a bzgrep.  I'd forgotten.
> 
> But createrepo doesn't fail at all -- it succeeds just fine.  It's the
> kickstart process that's choking.  (I'm inferring from this that kickstart
> looks at files that createrepo isn't modifying.)
> 
> It'd be useful to see what files in the repository kickstart is reading.
> The repository is on an F8 box, and I don't know how to get the atimes that
> would tell me.
> 
> And yep, there are hack-arounds.  (What you suggest is roughly one of the
> ones we've used, though we're not using yum.)  You have to admit, though,
> it'd be nice to do it in what seems to be the intended and documented way.
> And to understand why that way is failing in such whacky ways.
> 
> Plus, if I could make it work, I might learn something.  First time for
> everything.  :-)
> 
> 
> 
> On Wed, Mar 26, 2008 at 12:10 PM, Hugh Brown <hugh at math.byu.edu> wrote:
> 
> > On Wed, Mar 26, 2008 at 11:43:16AM -0600, Jeffrey Haemer wrote:
> > > Charlie,
> > >
> > > I'm getting all this from an HTTP server, not a CD or some other hard
> > > media.  It means I build my metadata on a real, F8 server and use it
> > from
> > > there.  Unfortunately, because it's F8, I can't tell what's actually
> > being
> > > read from the atime, but I actually have to use this server setup.
> > >
> >
> > Our solution was to create a separate yum repository in the %post section
> > of the kickstart and then use yum to install the extra software (though we
> > use nfs to do it).  You could write the script(s) to do all of this, put it
> > on the web server and then use a wget <script> in the post section and then
> > call the script after that.
> >
> > zgrep will let you look in gzipped files.  I also would have run the
> > createrepo with the failing aztek group name through strace to see which
> > files it was opening when it spit out the failure.
> >
> > Hugh
> > _______________________________________________
> > Web Page:  http://lug.boulder.co.us
> > Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> > Join us on IRC: lug.boulder.co.us port=6667 channel=#colug
> >
> 
> 
> 
> -- 
> Jeffrey Haemer <jeffrey.haemer at gmail.com>
> 720-837-8908 [cell]
> http://goyishekop.blogspot.com

> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> Join us on IRC: lug.boulder.co.us port=6667 channel=#colug