Open Stack

Zane Bitter wrote:
> On 20/11/14 04:41, Thierry Carrez wrote:
>> Joe Gordon wrote:
>>> [...]
>>> * Appendix 4 Section 3. To play devils advocate here, if the TC goes
>>> down the 'big tent small core' model where the TC is only really
>>> actively providing direct support for a small core while making space
>>> for many other projects in the OpenStack tent, does it make sense to
>>> have everyone in the big tent vote in TC elections even if the TC has no
>>> impact on their projects? For example if the TC is mostly focusing on
>>> nova/cinder/neutron/etc and is just providing a space for project foobar
>>> but does not interact with foobar developers or the foobar developers in
>>> any meaningful way should foobar developers vote in TC elections, even
>>> though the TC doesn't represent them?
> 
> I agree this is a potential problem, but not with the bylaws. This is
> why I think projects joining OpenStack need to agree to subject
> themselves to the oversight of the TC (in exchange for voting rights),
> and why the TC needs to get a lot smarter about how it exercises that
> oversight so that it can do so in a meaningful way.
> 
> If we're going to make projects formally a part of OpenStack but refuse
> to interact with their developers or grant them representation, then in
> what sense would they be part of OpenStack? What would be the point?
> 
>> You're right: it is one of the identified issues with the
>> bigtent/smallcore structure proposal -- if we are elected by
>> contributors of the "bigtent" but focus on a "smallcore", there may be a
>> drift. Some people argued that the "smallcore" still needs to fill the
>> needs of the "bigtent", so that still makes sense.
>>
>> That said, the TC election is governed by the TC Charter, so we have
>> some wiggle room there, as long as we use the "ATCs" in the election
>> somehow. We used to have PTL seats and directly-elected seats, we could
>> have weighted votes or seats purely elected by the smallcore
>> contributors in the future if that proves necessary.
> 
> Seriously? We have a moderately-large tent (16 services) right now, yet
> if my count is not mistaken only one out of 13 members of the TC has
> ever been core on a project that was not at one time considered in-scope
> for Nova, even though those projects comprise the majority of the 16
> services. The small core hardly seems under-represented.
> 
> And y'all seem to have had no trouble starting a serious discussion
> about abandoning the OpenStack mission of building a cloud platform in
> favour of diverting all resources toward building a compute
> virtualisation service and "just providing a space for project foobar
> but ... not interact[ing] with the foobar developers in any meaningful
> way". (Not that there's any consensus on this, but the fact that the
> discussion got this far, seemingly largely without the wider community
> becoming aware, is in itself remarkable.)
> 
> So if I were to make a list of issues with TC governance to worry about,
> "too *much* representation from non-compute service projects" would be
> at the very bottom.

Yeah, that doesn't sound like a problem in need of fixing. Especially
since there is no consensus at all within the current TC that the
smallcore is such a good thing anyway.

-- 
Thierry Carrez (ttx)