[katello-devel] available release versions

[James Bowes]

On Tue, Jul 10, 2012 at 09:01:40AM -0400, Tom McKay wrote:
> 
> 
> ----- Original Message -----
> > From: "James Bowes" <jbowes at redhat.com>
> > To: "Bryan Kearney" <bkearney at redhat.com>
> > Cc: katello-devel at redhat.com
> > Sent: Monday, July 9, 2012 10:56:38 AM
> > Subject: Re: [katello-devel] available release versions
> > 
> > On Mon, Jul 09, 2012 at 06:40:08AM -0400, Bryan Kearney wrote:
> > > On 07/06/2012 11:00 AM, Tom McKay wrote:
> > > >https://bugzilla.redhat.com/show_bug.cgi?id=834012
> > > >
> > > >Both headpin and subscription-manager-gui reach out to the CDN to
> > > >get a list of available release versions (5.8, 6.2, etc.).
> > > >Katello, however, bases the available release versions on the
> > > >subset that have repos enabled.
> > > >
> > > >With that in mind, I think the BZ above is working as intended,
> > > >ie. release versions shown in subscription-manager-gui have
> > > >nothing to do with what katello has available.
> > > >
> > > >I think there have been BZs/RFEs that systems should not be able
> > > >to set a release version when registered to katello to a version
> > > >where no repos (and hence no content) are available.
> > > >
> > > >Someone with more knowledge of the feature and intricacies please
> > > >chime in.
> > > >
> > > >_______________________________________________
> > > >katello-devel mailing list
> > > >katello-devel at redhat.com
> > > >https://www.redhat.com/mailman/listinfo/katello-devel
> > > >
> > > Where does subscription manager reach out to? It would be great if
> > > it looked into the directories which are published by katello.
> > >
> > 
> > Here is roughly what subman does:
> > - for all installed products on a system, iterate over the products:
> >   - if a product has the product tag 'rhel-N' consider it to be a
> >   base OS
> >     product.
> >   - take the last found base os product
> > - if no base os products found, set available releases to an empty
> > list.
> > - with the last found base os product, find all entitlements that
> > grant
> >   access to that product
> > - create an empty list, listings
> > - for each entitlement, do:
> >   - for each _enabled_ content set on the entitlement, do:
> >     - compare the product tag to the content set's required tag, ie
> >       (rhel-6 matches rhel-6 but not rhel-5). If the tags do not
> >       match,
> >       continue to the next content set.
> >     - take the content set's url, and split it on $releasever. Take
> >     the
> >       section before the split, and append '/listing' to it. This is
> >       the
> >       location on the CDN where we expect to find a listing file for
> >       this OS. add this file location to the listings list.
> > - for each entry in listings, download the file from the cdn.
> > - create a set containing all entries from all listings files.
> > - done!
> > 
> > 
> > As far as I can tell, SAM does:
> > - for each content set, grab its listing file, if available. return a
> >   set of the results.
> > 
> > 
> > Katello/CFSE does:
> > - for each enabled content set in a system's environment, get the
> >   content set's applicable releases. return a set of the results.
> > 
> > I assume the CFSE results come from pulp, but let's also assume that
> > they'll be identical to whatever the CDN would have.
> > 
> > Therefore, the difference between CFSE and SAM is that katello only
> > examines the content sets that are enabled in a given environment.
> > This
> > makes sense, as any content not enabled is totally inaccessible for
> > all
> > consumers in that environment. SAM doesn't have environments, so the
> > difference is moot.
> > 
> > Between katello and subman, katello's results should be a superset of
> > what subman sees.
> > 
> > Subman removes the following content sets:
> > - content sets that don't map to an installed product on the system
> >   (makes sense, imo)
> > - content sets that don't map to an entitled product on the system
> >   (makes sense for subman. could probably be a toggle for katello,
> >   defaulting to only showing entitled ones, or showing all if the
> >   system
> >   doesn't have any entitlements)
> > - content sets that are not enabled on the client (i dunno if katello
> >   has access to this, but i think this could fall in with the above
> >   logic)
> > - content sets that don't map to a product tagged with rhel-N (this
> > is
> >   probably a bit broken. we should maybe be looking for a product
> >   tagged
> >   with baseOS or something? but its more of an optimization; other
> >   content sets won't have listing files. safe to ignore for katello).
> > - content sets that aren't tagged with the same rhel version as the
> >   installed product (this makes sense to do in katello if possible,
> >   and
> >   again, probably as a toggle. you likely don't want to put your
> >   system
> >   on 6Server if it has 4AS installed).
> > 
> > Phew! That was a lot of stuff.
> > 
> > So, back to the original referenced bug. If the cli can use
> > releaseVer
> > against that SAM instance, but the gui can't, that totally sounds
> > like a
> > bug! The logic in both cli and gui should be the same, so we should
> > rectify that, then if that makes the gui work, good. if not, there
> > might
> > be a bug in SAM.
> 
> So I think we are agreeing to put this BZ into the subscription-manager bucket at the moment?
> 
> Is this BZ a blocker in any way? It does need to be addressed, but in which release?
>

I'll do some triage on it quickly today, then assign it as appropriate.

As for releaseVer setting, IMO katello could use the enhancements I
outlined above, to fall more in line with subscription-manager :)

> > 
> > -James
> > 
> > _______________________________________________
> > katello-devel mailing list
> > katello-devel at redhat.com
> > https://www.redhat.com/mailman/listinfo/katello-devel
> > 

-James
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/katello-devel/attachments/20120710/e179e63e/attachment.sig>