[Spacewalk-list] [Spacewalk-devel] Spacewalk got forked?! WTF is going on?!

[Neal Gompa]

On Mon, May 28, 2018 at 12:04 PM Duncan Mac-Vicar P. <dmacvicar at suse.de>
wrote:


> Neal Gompa writes:
> > Something is seriously wrong here.
> >
> > Today, I woke up to the news that SUSE forked Spacewalk[0] into Uyuni[1]
> > (which frankly, is a hard name to say...). The explicit reason they gave
> > for this was that despite requests to the community to step up to take
over
> > the Spacewalk project, the current Spacewalk leadership refused SUSE's
> > offer to lead the community.
> >
> > To me, something sounds fishy about this, because I've historically
known
> > Red Hat folks to be very much community-first, so if Red Hat as a
company
> > was no longer interested in Satellite 5/Spacewalk, transitioning it to
> > someone else who was interested in it should have been easy. However,
the
> > evidence they gave was pretty compelling[2]. At the same time, I have no
> > means of verifying whether or not what SUSE is saying is true.
> >
> > As a Spacewalk user, I'm incredibly disappointed that the leadership
> > managed to successfully drive away an interested party. At my scale
(tens
> > of servers), Spacewalk serves my needs perfectly.
> >
> > However, for those who think that Uyuni will be any better, my past
> > experience with most SUSE teams has not been pleasant. They're usually
not
> > the most responsive group with their open source projects (KIWI[3]
> > excepted). It's not a good sign that the SUSE Manager project
development
> > didn't already exist as an open branch to begin with.

> Hi Neal,

> SUSE Manager branch was not public because it was supposed to be, just
like the Satellite branches (which I understand are also not public), a
product branch used for releases. The collaboration point was always
intended to be Spacewalk.


Some aspects of it exist within the Spacewalk repo, but I suppose you're
correct.

> We never planned for our branch to start diverging. It just happened, and
at this point we think SUSE Manager code-base makes up for a better
"master" branch.


That's fair.

> > A splinter in the small development community around Spacewalk sucks,
and
> > the cursory glance at Uyuni seems to indicate that things are pretty
broken
> > for non-SUSE distributions.

> I would kindly ask you to give the people working on this some compassion
and benefit of the doubt. Things are not there yet.
> Since we realized we had to fork, we decided to use the openSUSE
Conference as a deadline to make things happen from our side. I think our
FAQ [1] covers some expectations about when things will be in place.


So you didn't actually plan to fork from the beginning, then? That makes me
feel a lot better about this.

For what it's worth, most of the SUSE projects I've seen and used are
fairly well maintained, but I've historically had a lot of difficulty being
able to communicate and contribute to projects. Not because there's any
particular barriers or anything like that. It's the simple problem of
people never being around to steward the project. That is, code reviews
linger, communication channels like IRC and mailing lists are dead, and
things stay broken. How do you plan to avoid that trap as you govern Uyuni?

Those complaints might sound familiar, because they're basically the same
ones you've had from the Red Hat folks about Spacewalk.

Then again, I don't think I've ever had someone like you actually reply to
me in a timely manner, so maybe things are going to be better this go
around. :)

> Uyuni server-side targets openSUSE Leap 42.3 for now. On the client side,
things should work on Redhat-like and SUSE-like systems, however, mind that
we haven't yet setup things to properly test that we support anything
outside what SUSE Manager officially supports.


Personally, I run mostly Fedora servers and desktops/laptops, with some
Mageia and openSUSE machines in the mix. So for me, the main advantage of
Spacewalk is being able to pull all those in and support them easily.

My main beef with Spacewalk right now is that I can't run the server side
stuff on Fedora, only CentOS. Someday, I hope to have everything on
Fedora-based infrastructure. But then again, I'm rather unusual. :)

> Once things are in place, we hope to take advantage of OBS, Salt and
collaborators to reach as many platforms as possible.


I'm certainly excited to see what improvements you guys have made to
Spacewalk, but what I really want to see is Red Hat and SUSE coming
together for a transition plan for the management of the Spacewalk project.

Please Red Hat folks, what can we do to make this happen? I'd much rather
see Uyuni be part of Spacewalk rather than a separate project. It's fairly
clear that SUSE is willing to pull out all the stops to support the
Spacewalk community.


-- 
真実はいつも一つ！/ Always, there's only one truth!