22:00:09 #startmeeting neutron_drivers 22:00:10 Meeting started Thu Oct 13 22:00:09 2016 UTC and is due to finish in 60 minutes. The chair is armax. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:00:11 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 22:00:13 The meeting name has been set to 'neutron_drivers' 22:00:37 hi 22:00:54 o/ 22:00:55 carl_baldwin, kevinbenton, amotoki ihrachys 22:00:57 ping 22:01:19 ok we almost have a full house 22:01:24 o/ 22:01:42 Lots of lbaas riff-raff... haha 22:02:04 o/ 22:02:18 we’re 10 days from the summit 22:02:34 and about a month from Ocata-1 22:04:07 not that people wouldn’t know but I figured I’d remind everybody 22:05:14 today I wanted to take the opportunity given by this hour to talk about 22:05:18 #link https://review.openstack.org/#/q/status:open+topic:stadium-implosion 22:05:51 over the past week I started looking at how things are going with the various projects in the lot 22:06:19 I have had a first pass on every project except bgpvpn and bagpipe 22:06:37 which I am hoping I’ll be able to process by the end of the week 22:06:54 what I need from you drivers members 22:06:58 and non 22:07:05 to review first and foremost 22:07:09 #link https://review.openstack.org/#/c/383525/ 22:07:29 make sure you’re happy with the content and the format 22:07:45 it’s a relatively straightforward list of Yes and No questions 22:08:34 the set of responses gathered should help us get an idea of the level of maturity and overall quality of a project 22:08:36 since you've done a first pass, what's the gist of the review? mostly in? out? 22:08:43 dougwig: well 22:08:56 dougwig: the in vs out is something I want to reach out with my fellow drivers 22:09:15 dougwig: I have a few takeaways I am sharing in a minute 22:09:19 dougwig: so bear with me 22:09:24 k 22:10:05 the scorecard should not be seen first and foremost as a mean to spur conversation and allow people to identify gaps 22:10:17 that might have been overlooked 22:10:28 "should not"? 22:10:33 sorry 22:10:36 should be 22:10:48 I was making sure people were paying attention 22:10:53 HenryG gets a candy 22:11:12 * armax hands over a virtual candy to HenryG 22:11:16 :) 22:11:20 * HenryG puts candy in mouth and finds it very sour 22:11:24 :) 22:11:33 come on! 22:12:44 I hope people see these as something that helps improve the quality of the projects 22:13:29 I think it definitely does. 22:13:29 it actually does, for what I see. though I am puzzled as to why it works. 22:13:44 that said, in the end we, the PTL, the Neutron drivers need to make a decision on where time and energy is well spent 22:13:53 when looking at the Stadium as a whole 22:14:21 I mean, quality should be at the core of the interest of a team. if what in reality is a simple badge is a better motivation to push for quality, I dunno. 22:14:37 and if on one end the tool is useful to improve quality for the individual project 22:14:40 ihrachys: It gives the projects a checklist, which I guess they didn't have before 22:15:28 on the other end it’s a tool that is useful for the PTL, and the drivers and the rest of the Neutron community to raise awareness and involvement in projects where one may not necessarily involved on a day to day basis 22:15:45 I’d love to spend the same amount of time on project X than I spend on Neutron proper 22:15:50 but I simply know that I can’t 22:15:55 * carl_baldwin wanders in late... 22:16:51 and I see these regular checkpoints as a way to foster discussion between the various parties invovled in the stadium 22:17:35 ok, now enough with the talking :) 22:17:44 any suggestions/comments so far? 22:17:57 you can throw virtual rotten tomatoes if you want 22:18:17 having those checkpoints at the start of a cycle makes a lot of sense, assuming projects actually follow the cycle in some way (some don't). 22:18:42 right, for now I am happy with the scorecard format I produced 22:18:56 if you guys are happy too, we can make incremental improvements over time 22:18:58 overall, I feel that the stadium finally gets some technical substance, so + on that 22:19:18 but for now I’d say if I can get eyes on https://review.openstack.org/#/c/383525/ 22:19:23 so that I can take it out of the way 22:19:36 I’ll be focussing on the actual assessments going forward 22:20:04 as for now I am filling the blanks at high level and to the best of my abilities 22:20:48 I’ll be reaching out to the various subproject liasions to make sure we capture a pitcure as objective and as trustyworthy of reality as possible 22:22:28 In the ‘final remarks’ section, I’ll will summarize at high level the status of the project and I’ll refrain from making a decision in vs out 22:22:55 which is something I want to agree in coordination with the drivers team once everyone has had the chance to review the scorecard 22:23:02 there will be exceptions though 22:23:38 especially if a project has clear indication of being stale 22:23:53 or not having chance of survival 22:24:09 one more thing and that I can answer dougwig’s question 22:24:49 the same scorecard would be used for projects that may want to be considered for inclusion, i.e. are not part of the governance list for Neutron and they want to be 22:25:09 ok 22:25:34 any question/comments so far? Otherwise I’ll start the drum rolls 22:25:55 armax: question about latest statement: does it mean, that projects can somehow "bypass" governance and will be included into Neutron Stadium Projects? 22:26:00 *last 22:26:20 clearly not 22:26:35 dasm: but I am not sure I understand your statement 22:27:08 what I mean is that if some project foo wants to be part of the governance list for Neutron 22:27:20 they would need to file an RFE and post a scorecard to the neutron-specs repo 22:27:45 armax: yeah. i re-read that. sorry for messing, i've misunderstood it, and thought about openstack governance list, not neutron governance list 22:27:49 in time for the first milestone of the given cycle 22:28:07 so say you have project X and want to be part of the Neutron stadium in Ocata 22:28:47 you have time until Ocata 1 to file and RFE, post a scorecard with all the dots being filled and we, the drivers team, we’ll talk to you and dive into the guts of the projects to see how cool it is 22:29:10 armax: ok, makes sense to me. 22:29:46 spin offs, like have happened in the past, e.g. dynamic routing, would get a free pass so to speak 22:30:14 but the assessment is at every cycle so if the project being spun off goes downhill 22:30:19 it can get to be removed 22:30:53 dougwig: what was your question again? 22:30:59 how much we suck? 22:31:05 or how much we rock? 22:31:09 What was your overall feel of the current list? 22:32:00 overall I noticed that some projects felt the pressure of the stadium and worked hard to cross most items on the checklist 22:32:13 some others didn’t give a rat’s ass 22:32:22 so it’s roughly 50/50 22:32:36 one thing that was consistent throughout the stadium projects 22:32:43 is that no-one gives a rat’s ass about neutron-lib 22:32:50 and that’s most likely is our fault 22:32:53 and we need to fix that 22:33:03 * HenryG sobs 22:33:12 if we want neutron-lib to be anything other than a repo that holds constants and api-ref 22:33:44 That's easy to fix. Just release the freeze on interface changes in neutron proper. 22:34:06 I look forward to talking about this in more details in a few weeks time 22:34:13 armax: the question is, do we steer neutron-lib evolution, or do we necessarily hardlift it ourselves? can we make use of that broader stadium community we claim we have? 22:34:14 I 22:34:16 but something that ihrachys suggested today during our secret meeting 22:34:56 is that we can consciously make a decision that neutron-lib is the way forward and we can force projects to break if they do not comply 22:35:10 ihrachys didn’t exactly suggested that 22:35:30 but the evil side of me decided to interpret his statement that way 22:35:34 Right. That's what I just said too 22:35:35 * ihrachys sounds tough in your mouth 22:35:46 dougwig: I don’t listen to you anymore 22:35:51 dougwig: you’re not a driver 22:35:55 Can we do quicker turn-around on deprecation removals? 22:35:58 lol 22:36:01 :) 22:36:33 HenryG: first and foremost I see neutron-lib as the key component to kill the coupling between projects and Neutron 22:36:42 if we don’t go over the hump 22:36:55 quick turnaround is not solving the problem 22:37:10 we need projects to get to 90+% import rate for neutron lib 22:37:40 if we want to stand a chance for it to be seen as a useful piece of our development and deployment puzzle 22:37:57 I think we should build neutron-lib, and make stadium projects pay for it 22:38:13 a toll 22:38:15 :) 22:38:22 I mean, make use of their passion to comply with the scorecard ;) 22:38:34 but if all deprecated functions/classes will be removed from neutron code, than non compliant projects will be broken, and would need to abide to brand, new world 22:38:45 *adjust 22:39:06 anyhow there’s probably more to talk about this, it’s not a difficult problem but it’s not sexy 22:39:26 we have a whole session to discuss it 22:39:51 but we need to crack this nut somehow 22:40:48 so neutron-lib being largely ignored is one piece of the finding 22:41:09 the other finding was that we partially suck at testing 22:41:41 * amuller pretends he didn't hear that 22:41:55 documentation is patchy and if a project exposes client extensions, they are not turned over to OSC yet 22:42:12 amuller don’t read that literally 22:42:44 overall there’s recognition of the fact that we need better CI 22:43:03 but most projects’ efforts are ineffective at this point 22:43:25 many projects still primarily and solely rely on unit testing 22:44:08 and for a system as complex as OpenStack that’s clearly problematic if we want to signal downstream consumers that what we produce is rock solid out of the stadium gate 22:44:54 for those projects that have no stable tempest like validation in the gate I am going to take a hard line 22:45:05 regrettably 22:45:26 they are going to escorted out of the premises 22:45:52 virtual premises so to speak 22:46:08 I have been asked on some reviews like https://review.openstack.org/#/c/383904/ 22:46:30 whether there’s a MUST TO-DO list out of the wider list of criteria 22:46:45 What is your objective? Is your objective to kick a project out, or to improve its testing? If it's the latter, it would likely be more effective to communicate this requirement to the project team and to give them sufficient time to accomplish that goal 22:47:21 I have communicated this effectively since last cycle 22:47:23 I believe 22:47:54 how long is sufficient, that depends on the project resources 22:48:06 at one point we need to draw a line 22:48:22 pushing deadlines over and over and over again is pointless 22:49:11 I think at this point teams themselves should draw a clear plan forward that could make the PTL to back up a bit. 22:50:08 I personally feel that so long as one demonstrates that all tasks are actively in progress and make steady steps forward 22:50:26 then it’s hard to draw a firm line 22:50:50 but one must have an end in sight 22:50:56 I have seen things drag on for ages 22:51:00 and ages 22:51:29 armax: what was the thinking behind the first deadline and not e.g. second? trying to understand if it's a hard deadline due to some release concerns. 22:51:35 exclusion does not preclude re-inclusion 22:51:52 you mean Ocata-1 vs Ocata-2? 22:51:55 yea 22:52:03 why -1 in the original spec 22:52:42 primarily because after Ocata-1 there’s generally less time to look into these types of issues 22:53:04 as more time gets dedicated on hands-on feature development 22:53:12 and release matters 22:53:32 ok, so there is nothing stopping us from ditching a non-complying subproject that committed to complete some gaps till e.g. Ocata-2 but has not done so. 22:54:01 * ihrachys not necessarily saying that anyone should be given a chance 22:54:37 we also want to make sure we give the TC enough time to digest/discuss 22:55:24 the related governance patch, so as of today I am giving myself Ocata-1 as the deadline by which I’ll be filing a patch to update the governance list 22:55:28 Yes, if a project is out of the stadium I would assume they would want to pivot and apply for inclusion in the Big Tent, if they have any life in them at all. 22:56:04 based on any discussion that will happen between now and then during the drivers meeting and/or the scorecard reviews themselves 22:57:39 any other comment or question, feedback? 22:58:38 if not, I would love to draw your attention on https://review.openstack.org/#/q/topic:stadium-implosion+status:open 22:59:23 It seems like a fair process overall. I know I have had to wade into fullstack testing to see how it works under the covers, to introduce it to a subproject. 22:59:24 the sooner we wrap this up, the sooner we’re gonna go back to doing what we like the most 22:59:29 My neutron-lib patches need reviewers 22:59:33 https://review.openstack.org/337731 22:59:35 https://review.openstack.org/382556 22:59:51 HenryG: ack 23:00:08 armax: attention drawn 23:00:29 njohnston: mind you, as pointed out, no project will be kicked out because of lack of fullstack testing 23:00:38 ok, thanks folks, we’re at time 23:00:40 #endmeeting