[Spacewalk-list] Re: [rhn-satellite-users] channel-to-update vs cloning and importing errata

Justin Sherrill jsherril at redhat.com
Tue Oct 13 14:30:47 UTC 2009


Erling Ringen Elvsrud wrote:
> Hello list,
> 
> Previously I have used the Channel-to-update script from
> https://fedorahosted.org/spacewalk/wiki/channel-to-update to create a
> new channel on Satellite Server for a specific update of RHEL.
> A data-file for RHEL 5.4 is not available yet (the datafile contains
> the names and version numbers of the spesific pacakges contained in a
> given update of RHEL). Using the script causes the channel to contain
> a lot less pacakges than cloning the channel from the Base RHEL 5
> channel synchronized with RHN and adding all errata.
> 
> Do you know what advantages either approach has?
> Is it any drawback to clone and import all errata and having more
> pacakges in the channel? will it have any speed or reliability
> implications?
> 
> Thanks,
> 
> Erling
> 
> --
> rhn-satellite-users mailing list
> rhn-satellite-users at redhat.com
> https://www.redhat.com/mailman/listinfo/rhn-satellite-users


Hi Erling,

When cloning a channel and including all of the errata, you will get all
packages up to the latest.  So you would get foo-1.1, foo-1.2, foo-1.3
etc....

When using that script, you only get the exact package set that was
included in the update.  (i.e.  foo-1.2).

There really isn't any drawback to having all the packages in the
channel.  It may make some operations slighly faster, but if you don't
have an obscene amount of systems (i.e. > 10,000), you may not notice
any difference.

The purpose of the script is really to make it easier to control what
update level a channel is at.  This is not very easy with using the
'clone' errata feature unless you're going to the very latest update.

I'll try to update the script today with 5.4 data files.

-Justin

-- 
Justin Sherrill, RHCA          1801 Varsity Drive.
Software Engineer                Raleigh, NC 27603
Red Hat, Inc.




More information about the Spacewalk-list mailing list