[openstack-dev] [all] [nova] [placement] placement below or beside compute after extraction?

Matt Riedemann mriedemos at gmail.com
Tue Aug 21 14:04:02 UTC 2018


On 8/21/2018 7:55 AM, Thierry Carrez wrote:
> Matt Riedemann wrote:
>> [...]
>> Regarding microversions I was mostly thinking of the various times 
>> I've been asked in the placement channel if something warrants a 
>> microversion or if we can just bug fix it in, like microversion 1.26. 
>> I then generally feel like I need to be defensive when I say, "yes 
>> it's a behavior change in the API so it should." That makes me 
>> question how stringent others would be about upholding 
>> interoperability concerns if I weren't around. [...]
> 
> The issue with that kind of distrust by default is that it's not 
> sustainable... In a large project you can't have every individual review 
> everything because they trust noone else.

It's not distrust by default. I said, "thinking of the *various times*". 
Which means more than once. But I also said I was asked for input, and 
failed to reflect on that until I actually wrote it down. That's my fault.

> 
> That is why in OpenStack we instituted a culture of "trust by default, 
> then escalate to PTL or TC if shit ever hits the fan". And the fact is, 
> the PTL (at team level) or the TC (between teams) rarely had to 
> arbitrate conflicts, because there aren't so many conflicts that are 
> escalated rather than solved by consensus at the lower level.

Sure, but I'm sure there are also times where people don't escalate 
simply because they want to avoid conflict. There have been many times 
where I've questioned another nova core's +2/+W on a change and rather 
than make a big deal out of it, I push that frustration way down but it 
comes out in other ways, like distrust later. Again, that's my fault, 
but I suspect I'm not the only person in OpenStack that does this. On a 
good day I'll ask the person directly in IRC, or failing that on the 
review, "hey why did you do this? Did you think about X?".

> 
> Restoring "trust by default" between placement and the rest of Nova 
> seems to be the root of the problem here. In a community, it's generally 
> done by documenting general expectations and shared understandings, so 
> that you create a common culture and trust by default people to apply it.
> 
> What would you suggest we do to improve that in this specific case?
> 

Trust falls! I don't know, Thierry. Likely just improved direct 
communication with the people involved rather than back-channeling and 
distrust/hurt feelings which lead to "sides" being setup. As I said 
above, direct communication can be hard because of the confrontation and 
awkwardness so it's easier at times to take the shitty way out and just 
complain about so-and-so to someone else that thinks the same way you do 
rather than try to gain understanding and listen to other viewpoints. We 
(I) go over this every retrospective but still fail to learn from and 
practice it.

-- 

Thanks,

Matt



More information about the OpenStack-dev mailing list