[openstack-dev] [neutron] What does being a neutron-core member mean? [WAS: Re: [neutron] I am pleased to propose two new Neutron API/DB/RPC core reviewers!]

Ihar Hrachyshka ihrachys at redhat.com
Wed Aug 19 17:37:53 UTC 2015


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 08/14/2015 05:25 PM, Russell Bryant wrote:
> While this includes me, I'm really not taking this personally.
> I'm thinking about it in the general sense.

Thanks, and sorry for the thread, but I felt that we should clarify
the matter asap if the new world does not completely stick to some
minds in the team, like mine.

> 
> On 08/14/2015 11:03 AM, Kyle Mestery wrote:
>>> I'd argue the system is built on a web of trust. If you trust
>>> me, and I trust Russell and Brandon, then you should likely
>>> trust Russell and Brandon as well. This is EXACTLY what the
>>> Lieutenant system was meant to convey, though I now feel like
>>> perhaps people missed this key ingredient of the new world we
>>> find ourselves in.
> 
> This is a huge and important point.  The N to N trust model we've
> been operating under doesn't scale.  Neutron is trying to move to a
> different trust model that has proven to scale much further than
> we've been able to do within a single OpenStack project so far.
> 

Indeed that's the case, and I believe Kyle's efforts are the main
reason we have a team that is so open to newcomers and different
perspectives. That's why I feel that even if I am not completely
bought in the new world doctrine, I should expect that if Kyle does
something, he makes it for good. That's indeed trust, but based on
past experience, not the potential future.

> If Kyle and others leading a section of Neutron trust me, I'm happy
> to jump in and do more reviews.  If they trust me, I'd hope others
> not as familiar with me or my work can trust by proxy.  The same
> applies to Brandon.  I honestly don't know Brandon very well, but I
> have a high level of trust for Kyle.  Kyle trusts him, so +1 from
> me.
> 

I tried to digest it these days, and I still have problems with it.

In my world, trust indeed can be based on proxy referrals, but not
solely. Usually whenever there are nominations sent, I already have a
clue of how candidates contribute to the repo, their goods and bads.
If not, I always have a source of truth to fix the lack of knowledge
(stackalytics, git log, ...) If I have no ways to know, I don't feel
like I'm in the position to +1 a nomination. If the vote is to be
based solely on my trust in Kyle, then I'm better not to vote at all
(that's what I did) and defer to Kyle alone to make a judgement. With
this new world, do we even need to vote?

> Kyle has a tough role here.  It means he gives up some control, but
> it's the way the project will scale.  Kyle doesn't have to develop
> a close trust relationship with everyone merging code into the main
> neutron repo, much less all the other repos.  He can delegate some
> of that.  It only works if everyone buys into this way of
> thinking.
> 

Agreed, though there is still should be some kind of relationship
between team members, if not with Kyle.

I think the discussion is not about delegation though; but about
criteria that we apply to core reviewers, both existing and new ones.
Kyle removed several members from the core team before because of
their low review stats (in the past, not in the future), and I believe
it was the right thing to do. But then we should probably apply
similar requirements to new candidates.

Now, I see Kyle's point in trying to push for more proxy trust in
neutron, and I also think that I may miss new developments in project
governance, and indeed adding more cores in the team is a good thing,
especially from those contributors that showed their effectiveness in
other openstack projects (ovn, nova for Russell; lbaas for Brandon).

I only want to make sure that by doing it we don't lower criteria to
core reviewers in terms of number of reviews etc. I see we lag on it.

And also I am concerned that we may suggest to others that non-core
reviews are somehow insignificant and that collecting some review
stats for several months or even weeks is considered a pointless
effort unless it's done with a core hammer.

I value reviews of multiple openstack folks who are not cores in
neutron, f.e. @otherwiseguy and @gsagie for anything ovsdb related; or
@zzzeek for all things sql-ish; or @salv-orlando for all things
API/policy/how the world started to exist; or @jlibosva for anything
python-esoteric; or @moshele for sr-iov... The list follows. I don't
want them to feel their reviews are not that valuable if not +2.

Ihar
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJV1L7sAAoJEC5aWaUY1u57rhYIAOCxX5K8HbHNcUYh33/rvTRc
InqvoKKu5NWFU4s+iUZly1Fu5/3JXzLR5h8/+1b/gWR5EDYJ2Q6FRy5Bmn6Fduuw
cHV6trqGfEmA+NJsvNSNeg1Ux8hQ0hjdcF0mcrMhM0li+DFDMwogHMxPUAzKF4Cm
fcMDr+aZW3zkUDi0Y/iUjxqWwQFOBwPg0gWhKqhasVTqbOfd0W62z4gT9o6ZYkdn
mJr6jU+qc1hV+St03qpgLN9h/S023Ha1PCnMBUfjldbthmVBtJEsgmyfHlqWpWmc
UGSH7vTv6EmSSVcvZmAuCSZQKeG4UaodbApNusELFTA4zSK6GeKgJL9hr9MfkCs=
=ZXGT
-----END PGP SIGNATURE-----



More information about the OpenStack-dev mailing list