[CentOS-devel] are there any chances to see finished CentOS6 in 2011?

[Thomas Bendler]

Hi Karanbir,

2011/1/7 Karanbir Singh <mail-lists at karan.org>
> Hi Thomas, I seem to have missed your email yesterday.

no problem, happens to me quite often :).

> On 01/05/2011 01:37 PM, Thomas Bendler wrote:
> >     CentOS is not SL
> > I know but the source is equal.
> Not really. They do things to the distro that we dont - and we work
> towards different(ish ?) goals. A bulk of the work is shared and perhaps
> overlaps, which is why having them do their thing on a different track
> is good - for us and for them.

What I mean is that both flavors are build out of RHEL sources with
more or less work on the original sources.

> > I'm seeking for two different things, one thing is an alpha or beta ISO
> > (which gives you at least the ability to install a minimal CentOS 6) two
> This is the bit that I dont understand. If there is an ISO set, why
> would it be Alpha or Beta ? if the iso sets are complete - we release
> that stuff as gold.

That is exactly the point, it's released when it's ready. So looking
at the current situation, RHEL6 was released early November, now it's
early January and I still can't start with reworking my puppet
installation to match CentOS6 (I did it with RHEL6 and SL6, but I need
to check it for CentOS as well). So when CentOS6 is released (maybe
this month, maybe next month, maybe even later) I can start working on
Puppet to get it ready for this release (which may also take a month
or so until everything is in production). This means, when CentOS6 is
released I have the upgrade work as well as the configuration
management work at the same time, I would like to do the configuration
work before CentOS6 is released. Simply because when CentOS6 is
released and it's in the press that it's released a lot of people
stand in front of my desk asking me when they can use the new release
because of feature xyz ... :).

> [...]
> However, the focus isnt working with these tools - the idea of these
> tools working is upto them doing whats needed to work with CentOS. The
> CentOS focus is and must remain upstream compatible.

Yes, but this is the problem, CentOS identify itself as CentOS. When I
check the configuration management with original RHEL6 it identify
itself as RedHat. There are some ways working around that issue (i.e.
patch facter or changing redhat-release), but I would prefer having
something like download the RHEL Source RPMs, integrate CentOS
modifications (in a state as it currently is) and have it run with
whatever tool to get an ISO image as the result. This also show me if
there are still build errors which I can report to bugs.c.o or where I
can send a patch to bugs.c.o so that the build team can fix it as
well.

> [...]
> The idea of sekreet-sauce is completely unfounded. And therein lies the
> issue. Too many people think that we bless the stuff with some holy
> water from somewhere to magically churn out packages. Its simply not
> true. At best a rpmbuild -ba <foo.spec>; should give you a usable
> result, at worst a mock rebuild. As for documenting things - so many
> people are using bugs.c.o already, why do you find that unsuitable ?

It's not an idea, it's how people recognize the current situation.
Make a simple test, forget for an hour all things you know about
CentOS (will be hard in your case, I know) and go to
http://centos.org/. Check out if you find something like build logs,
some state how far the process of creating CentOS6 is, what packages
still need to be reworked, how someone can help reworking these
packages, ... and you will see that these informations are not
visible. They might exist but they are not visible for someone going
to the web page and try to get these informations. So if there is a
link on the front page pointing to a guide how to setup a build
environment for CentOS6 locally and how to report issues and patches
to bugs.c.o and how you can get build logs of the current rebuild
process, everything is fine because anyone can start working on open
items and contribute to the project to make things happen faster. To
make it clear, I think bugs.c.o is a good thing but useless for anyone
who don't have a local build environment.

> [...]
> havent we already been over this ground repeatedly a few times on the
> list already ?

No, nothing changed till now, the only statement so far was something
like this can be discussed when CentOS6 is out. So the problems still
exist and the question is, will the problems be really solved when
CentOS6 is out? If so, I will shut my mouth and we can follow up with
the discussion when CentOS6 is out. If everything stays as it is, I
have to rethink if CentOS is the best solution for my needs.

Kind regards, Thomas