[OpenIndiana-discuss] [oi-dev] Proposal: OpenIndiana Stable Branch

[Guido Berhoerster]

* Deano <deano at rattie.demon.co.uk> [2011-01-15 18:39]:
> Guido wrote
> > I agree with that, however I consider this strictly a makeshift
> > solution for those who are now desperately in need for a
> > replacement of OpenSolaris to bridge the time until the
> > Illumos-based OI has stabilized.
> 
> How does illumos-based OI change the situation we currently have, that we
> don't have the man power to manage the multitude of packages the OpenSolaris
> had/has?
> It's a classic boot strap issue, until enough package maintainers are
> available to do critical/security updates on OI, regardless of the backend
> ONNV or illumos, applications will atrophy. But if OI isn't production
> ready, then nobody will come forward as package maintainers. IMHO this can
> be seen in the state of OpenSolaris and OI package mess (dependency loops,
> etc.)
> Similar to how the wine/cross-over dev have to concentrate on the most
> important windows apps to get running, OI stable should pick apps most
> used/wanted that have active maintainers and interested parties. Nothing
> wrong with having all the other apps in repositories for use in less stable
> environments, but at least we are being honest about the status.
> 
> > > If external contributors were able and willing to commit patches/fixes
> beyond the supported list, we'd accept them with open arms, and this could
> be a great way to extend the contributor list and get more people involved.
> >
> > Note that this is not just be about accepting contributions,
> > it would also involve review and testing by regular developers,
> > in particular if we make committments about stability. I think
> > there are better ways for potential contributors to get involed
> > into the project so let's stick to the above list.
> Not sure what you mean by regular developers? At the moment we have a few
> maintainers each taking huge incorporations, that's a massive (and we are
> all very grateful) commitment. Most I suspect can't take on such a large
> chunk. However I suspect many of us could maintain a package or two
> specially if we already using them in a production OI setting, as we already
> have to do security watching and updates. However that means the smaller
> sized chunk list will be largely driven by what actual users use rather than
> a list of chosen packages.

You seem to have a misconception of how the OI project operates.
The current situation is that the largest part of maintenance
work for all consolidations except ON is still done by Oracle
although consolidation maintainers add fixes on top of that where
they deem it necessary. And we might take advantage of that as long
as it is feasible.
Adding additional packages is a different story and we are
currently working on the procedures and infrastructure to provide
an addon consolidation (see
http://wiki.openindiana.org/oi/OpenIndiana+Addon+Consolidations)

> I'm personally a good case in point, I'm happy to help OI out with a few
> how-tos and packaging up the few apps I care about and use day to day, as a
> thank you back to the massive and impressive work done getting OI where it
> is now. But currently there doesn't seem to be a spot for me to really help
> out... the apps I'd probably package and support are nginx and node.js. Not
> on the top 10 list but they are what I use and so know and could maintain
> (when I have to learnt IPS packaging).

That sounds great, if you want to contribute packages today, the
best solution would be to add tem to SFE
(http://pkgbuild.sourceforge.net/spec-files-extra/) since it
allows others to easily build your contributed package and since
SFE is likely to become one of our "upstream" projects where we
might import packages from at a later time.
I can offer to assist you with that, please mail me privately or
stop by in #oi-dev.

> > Since this is just a bridge for those who already use OpenSolaris
> > or OI in production I'm categorically against demands for feature
> > additions sucha as providing postfix and a pony that already seem
> > to pop up.
> > This is a temporary solution, so please let's treat it that way.
> If this is temporary solution, what is the permanent solution? Illumos
> solves the ONNV issue, but going forward how does OI as a distribution be
> production ready as a server platform (its obviously first use) etc.? and
> encourage user to pitch in?

The issue here is that Oracle ONNV is unmaintained and a dead
end. Every genuine fix or backport from Illumos binds resources,
resources which we need for the initial integrating Illumos into
OI. A "permanent" solution is having stable releases based on a
ON consolidation which is actively maintained by the Illumos
project.

-- 
Guido Berhoerster