Open Stack

Sorry to have missed the lobby discussion on this this morning.

An "Ops Project" feels weird to me.

Most things I can think of going into this space seem to be better
served as part of existing projects.  For one example logging filters
& parsers should probably be distributed with Oslo which defines how
logs are output.  If I were just coming to OpenStack that's where I'd
expect to find them.

Obviously there's question of what to do with bits that may not fit
existing projects if there isn't a new Ops Project.  Also the set of
people qualified to be core reviewers of this type of contribution are
likely to be one group and not the people who are core reviewers of
python contributions.

So you could say I'm of two minds.  I think we should clearly define
what it is we want and then think hard for a bit about the best way to
implement that rather than rush right at a new incubated project.
Perhaps what we need is a new structure.

As I see it we want:

1) A central place to collect contributions so that we don't multiply effort
2) An integrated way of distributing those contributions with
OpenStack so new operators get the Ops pieces along with the related
code.
3) Peer review and automated testing of contributions.
4) "ATC" recognition of contributors

Granted other than the cross cutting nature, that does look a lot like
a Project.

-Jon

On Thu, Nov 6, 2014 at 7:59 PM, Xav Paice <xavpaice at gmail.com> wrote:
> On 07/11/14 01:42, Michael Dorman wrote:
>
> This same thing came up at the mid-cycle Ops meet up at RAX in August.
> There wasn’t much action that came from it, but we did set up a new org in
> GitHub for collecting this type of common stuff:  https://github.com/osops
>
> Excellent! Wish I'd noticed this earlier...
>
> If we've got as far as this, there's already a location for us to share
> tools and commonly needed 'things' as ops.  Michael has a great point that
> this is stuff which we all benefit from, and the level of collaboration and
> peer review available on OpenStack projects is well beyond any Stackforge
> project.
>
> It would also assist the other projects, including documentation, if we had
> a coordinated and agreed best practice for deploying and running OpenStack.
> In particular, the packaging efforts put in by the distros seems to me to be
> something that varies far too much, with many ops selecting to disregard
> distro packages in favour of our own.
>
> _______________________________________________
> OpenStack-operators mailing list
> OpenStack-operators at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
>