Open Stack

On 8/22/2018 1:25 PM, Jeremy Stanley wrote:
> On 2018-08-22 11:03:43 -0700 (-0700), melanie witt wrote:
> [...]
>> I think it's about context. If two separate projects do their own priority
>> and goal setting, separately, I think they will naturally be more different
>> than they would be if they were one project. Currently, we agree on goals
>> and priorities together, in the compute context. If placement has its own
>> separate context, the priority setting and goal planning will be done in the
>> context of placement. In two separate groups, someone who is a member of
>> both the Nova and Placement teams would have to persuade Placement-only
>> members to agree to prioritize a particular item. This may sound subtle, but
>> it's a notable difference in how things work when it's one team vs two
>> separate teams. I think having shared context and alignment, at this point
>> in time, when we have outstanding closely coupled nova/placement work to do,
>> is critical in delivering for operators and users who are depending on us.
> [...]
> 
> I'm clearly missing some critical detail about the relationships in
> the Nova team. Don't the Nova+Placement contributors already have to
> convince the Placement-only contributors what to prioritize working
> on? 

Yes. But it's not a huge gun to the head kind of situation. It's more 
like, "We (nova) need X (in Placement) otherwise we can't get to Y." 
There are people that clearly work more on placement than the rest of 
nova (Chris and Tetsuro come to mind). So what normally happens is 
Chris, or Eric, or Jay, or someone will work on the Placement side stuff 
and we'll be stacking the nova-side client bits on top. That's exactly 
how [1] worked. Chris did the placement stuff that Dan need to do the 
nova stuff. For [2] Chris and Eric are both working on the placement 
stuff and Eric has done the framework stuff in nova for the virt drivers 
to interface with.

Despite what is coming up in the ML thread and the tc channel, I myself 
am not seeing a horde of feature requests breaking down the door and 
being ignored/rejected because they are placement-only things that nova 
doesn't itself need. Cyborg is probably as close to consuming/using 
placement as we have outside of nova. Apparently blazar and zun have 
thought about using placement, but I'm not aware of anything more than 
talk so far. If those projects (or other people) "feel" like their 
requests will be rejected because the mean old nova monsters don't like 
non-nova things, then I would say that feeling is unjustified until the 
specific technical feature requests are brought up.

> Or are you saying that if they disagree that's fine because the
> Nova+Placement contributors will get along just fine without the
> Placement-only contributors helping them get it done?

It's a mixed team for the most part. As I said, Jay and Eric work on 
both nova and placement. Chris and Tetsuro are mostly Placement but the 
work they are doing is to enable things that nova needs. I would not say 
"get along just fine". The technical/talent gap would be felt, which is 
true of losing any strong contributors to a piece of a project - that's 
true of any time someone leaves the community, whether on their own 
choosing (e.g. danpb/sdague) or not (e.g. alaski/johnthetubaguy).

[1] 
https://specs.openstack.org/openstack/nova-specs/specs/queens/implemented/migration-allocations.html
[2] 
https://specs.openstack.org/openstack/nova-specs/specs/rocky/approved/reshape-provider-tree.html

-- 

Thanks,

Matt