08:04:31 #startmeeting ptl_sync 08:04:32 Meeting started Tue Aug 5 08:04:31 2014 UTC and is due to finish in 60 minutes. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:04:33 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 08:04:35 The meeting name has been set to 'ptl_sync' 08:04:46 #info Today we'll only do Nova, which we missed last week 08:04:51 #topic Nova 08:04:55 mikal: ready? 08:05:05 You're back home? 08:05:38 johnthetubaguy: o/ 08:05:58 mikal, johnthetubaguy: how did the sprint last week go? 08:06:30 Heya 08:06:35 I thought the sprint went really well 08:06:44 We got through heaps of stuff 08:06:50 * mikal finds the etherpad 08:07:02 #link https://etherpad.openstack.org/p/juno-nova-mid-cycle-meetup 08:07:43 great 08:07:48 yeah, had some good discussions 08:07:48 lots of stuff covered I see 08:07:56 OK, juno-" now 08:07:59 juno-3 08:08:05 #link https://launchpad.net/nova/+milestone/juno-3 08:08:14 You have 72 blueprints targeted at this point 08:08:19 yeah, we agreed to start punting some blueprints without code 08:08:25 So, I think there are still some spec freezae excceptions I haven't nacked 08:08:29 because we ain't merging 72 blueprints 08:08:31 But I should get to those ASAP 08:08:46 Feature proposal Freeze is August 21, so in a bit more than two weeks 08:08:58 I agree that we're going to need to be a bit brutal 08:09:03 Then juno-3 feature freeze is in 4 weeks 08:09:32 The trick is really to get stuff out of the way ASAP 08:09:42 that's the only way to avoid critical congestion 08:09:43 yeah, will reduce the priority of blueprints that are not started, and start making rude comments, and punting big blueprints that clearly have no code, at least thats my plan 08:09:56 otherwise everything is "almost ready" at the end 08:10:11 Works for me 08:10:23 So review priority is a bit of a combination between feature priority and how close it is from completion 08:10:54 the latter being sometimes a lot harder to determine from data 08:11:03 We've also talked about tweaking our meetings to help here 08:11:07 From subteam status focused 08:11:15 To what needs review to get stuff in for juno-3 08:11:22 yeah, I guess I have the first meeting using that format? 08:11:31 johnthetubaguy: yes, you're the lucky winner 08:12:20 https://blueprints.launchpad.net/nova/+spec/multiple-if-1-net has no assignee 08:12:35 Hmmm, an NFV one 08:13:00 There are also 6 with "unknown" status, which you should set to "Not started" or whatever 08:13:24 https://review.openstack.org/#/c/98488/ 08:13:25 ijw-ubuntu is the spec author of the multiple if one 08:13:32 it has code, buy yeah, I need to dig into all of those 08:14:30 But yes, I think focusing meetings on getting stuff final-approved should help 08:14:39 Yep 08:14:40 anything that merges now won't have to merge in 4 weeks 08:14:49 when it will be a lot less fun to go in gate 08:14:50 I feel like we're still feeling our way with specs and how it impacts the world 08:14:53 But getting closer 08:15:47 yeah, would be great to front load this, but the reality is that all the reviewers are busy getting their code up at this point 08:15:52 At 72 blueprints you need to complete 18 per week ;) 08:15:59 We can do it! 08:16:00 Not 08:16:14 so more than juno-1 and juno-2 put together, per week 08:16:15 hmm 08:16:17 so realistically you know you probably won't get a lot more than 40 08:16:22 right 08:16:26 the trick being to identify which of those 72 are in the 40 08:16:43 because identifying the 40 early really helps :) 08:17:00 There are a lot in "needs code review" 08:17:03 Possibly close to 40 08:17:05 honestly, the medium and high that are in "needs code review" are the ones I really care about at this point, but yeah, need to chase the others 08:17:08 (I haven't counted) 08:17:43 So this is also about calling contributors to be honest and realistic, and have them defer anything that can wait 08:17:44 29, but yeah 08:18:08 Agreed 08:18:13 I think we're on the same page here to be honest 08:18:17 ttx: +1 and not making people waste time with putting up patches that we will never have time to review 08:18:23 We just need to keep reviewers pulling in approximately the same direction 08:19:00 yes, because if the effort spreads across the whole 72, you will likely only get a handful merged 08:19:17 and 68 of them will be 90-99% there 08:19:20 IIRC, we also talked about a weekly email to -core 08:19:23 Which might help 08:19:24 which is what we want to avoid 08:19:30 We just need to actually do it 08:19:32 we are talking about having a top 10 blueprint list, so we should discuss that in the meeting 08:19:44 Do we feel like enough cores come to the meeting? 08:19:49 Would an email catch cores who don't? 08:20:05 mikal: not really, as some are not in that timezone, will have to mail everyone too 08:20:13 Yeah 08:20:16 I kinda want all reviews with the same focus, core or not 08:20:17 I think I agree 08:20:19 you could also have two cores sign up for every day between now and j3 08:20:28 and have them work as a team to speed-approve stuff 08:20:30 johnthetubaguy: that email could go to -dev, I'd be fine with that 08:20:46 mikal: yeah, thats what I was meaning 08:21:00 Out of interest... 08:21:01 ttx: thats not a bad plan, we used to have that 08:21:09 I don't think any of the specs granted a freeze exception landed 08:21:13 doesn't work "all the time", but could work for the next 4 weeks 08:21:16 Did we learn spec freeze exceptions are a waste of time? 08:21:25 mikal: we landed two I think 08:21:43 mikal: not 100% sure though, maybe it was one, I can't remember now 08:21:58 Yeah, its all a jet laggy blur for me at the moment 08:22:06 mikal: the issue is that there are tactical contributors which will ask for an exception no matter what, even if they know in their hearts there is no way that feature will be there 08:22:24 Heh 08:22:28 I can think of one of those 08:22:28 if everyone was just realistic and had the project's interest in mind 08:22:36 that would be more usefu 08:22:37 l 08:22:46 Well, there are also people who only care about their thing 08:22:58 I find when I say to people that they should review other stuff to help their stuff land 08:23:03 That they often looka t me funny 08:23:13 Cause they really don't care about OpenStack 08:23:15 Just landing their thing 08:23:17 well, we do also need to be better about "resolving conflict" on key specs earlier 08:23:32 we did a review day for stable to unblock a month-worth of review pileup on stable/icehouse 08:23:38 some stuff we need, but can't agree the direction 08:23:40 granted those are easier to review 08:23:50 ttx: I feel like stable is a distro thing 08:23:55 Which I think that thread agreed on 08:24:03 but in one day we ended up doing more than in the month before 08:24:11 just by being two of us focusedd on the same reviews 08:24:21 Yep 08:24:24 mikal: sure, I was just advocating for the "two core per day" thing 08:24:27 I agree review days are a good idea 08:24:27 we had that with specs on the spec review day, it did work 08:24:43 I think too many "days" tires people though 08:24:49 So we do need to be careful 08:24:55 yes, it's not a magic bullet 08:25:04 its just I don't know I can sign up to any full days of reviews right now, just pulled in too many directions at this second in time 08:25:35 it's just that having two cores work on the same reviews at the same time is sometimes more efficient than having them work independetly 08:25:42 but we can try it, for sure 08:25:54 I agree we should try it 08:26:06 (re)try it I supose 08:26:33 how much do you use the channel to ask for that last +2/APRV review ? 08:26:43 Not much in my timezone 08:26:44 ;) 08:26:47 heh 08:27:28 because sometimes you push a +2 and you're pretty sure it just needs another formal +2, the thing haveing been reviewed by others 08:27:35 and concerns addressed 08:27:47 I have a script which surfaces those fro me 08:27:53 it does happen, but also, not so much till the US wakes up 08:27:59 So when I am not travelling I feel like they get pretty good attention 08:28:06 so signlling those (with an core email, in a meeting, with a ping on channel...) can work wonderz 08:28:33 Its interesting to me how every problem seems to be solved with better communication 08:28:36 So I agreee 08:28:39 yeah, we need better tooling here, well people have it, just not in a way to share with everyone yet 08:28:44 A status email every couple of days is a good plan 08:28:49 johnthetubaguy: yeah 08:28:52 anyway, enough brainstorming 08:28:52 we'll see how much progress is made next week 08:28:58 +1 08:28:59 johnthetubaguy: I fear being gamed if my algorithm is published 08:29:26 mikal: right, thats another fun one, but if gaming you involves doing "good" things, then thats cool 08:29:29 #info today at 72 blueprints, 1 implemented, 29 under final review 08:29:56 We'll discuss nova-network vs. neutron, and migration/transition at the meeting today 08:30:01 well, tomorrow for mikal 08:30:05 Yep 08:30:11 I will try go through the list and see how that goes 08:30:11 We did discuss that at the meetup 08:30:14 anythign else you'd like to push to the agenda? 08:30:21 Not that I can think of 08:30:31 ok then, I'll free you guys up 08:30:34 We have some ideas for specs in Kilo, but its premature and they're half baked at the moment 08:31:26 anything you wanted to ask/mention? 08:31:41 Not that I can think of 08:31:58 alrighty then 08:32:00 #endmeeting