14:01:44 <dhellmann> #startmeeting releaseteam
14:01:45 <openstack> Meeting started Fri May 27 14:01:44 2016 UTC and is due to finish in 60 minutes.  The chair is dhellmann. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:46 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:49 <openstack> The meeting name has been set to 'releaseteam'
14:01:53 <dhellmann> courtesy ping: ttx, dims, lifeless, tonyb, stevemar, fungi
14:01:53 <dhellmann> If you would like for me to add you to the courtesy ping please say so
14:02:00 <dhellmann> our agenda is at https://etherpad.openstack.org/p/newton-relmgt-tracking
14:02:05 <fungi> heyhey
14:02:05 <dhellmann> under R-19
14:02:16 <dhellmann> hi, fungi!
14:02:25 <ttx> o/
14:02:39 <dhellmann> I'm going to go out of order to let fungi drop off if he wants after...
14:02:41 <dhellmann> #topic automation update (fungi)
14:03:14 <dhellmann> I think our goal was to have the special nodes set up for N2, right?
14:03:17 <fungi> the very brief update is that i'm setting aside a good chunk of today away from our server upgrades sprint to get the signing slave work implemented
14:03:22 <fungi> yep
14:03:27 <dhellmann> cool
14:03:47 <fungi> so i should have some progress to show for next meeting
14:03:49 <dhellmann> one small thing changed since the summit, in that now when we release a lib we also submit a patch to the requirements repo to update the constraints
14:04:07 <dhellmann> I think you said these nodes would have all the gerrit credentials in place to do that?
14:04:08 <fungi> oh, cool, i know we'd discussed that as a possibility in the past
14:04:37 <dhellmann> the current implementation introduces a bunch of merge conflicts, so I may redo it, we'll see, but at least now we don't need folks to remember to do it by hand
14:04:41 <fungi> yeah, we'll need a gerrit account capable of pushing tags, so also pushing a change for constraints isn't a big deal
14:04:50 <dhellmann> that's what I figured, thanks for confirming
14:05:07 <dhellmann> I'm just using "git review" to submit it now, is that going to work in the gate, too?
14:05:20 <dhellmann> sorry, in CI
14:05:34 <fungi> yep
14:05:41 <dhellmann> perfect
14:05:42 <fungi> that's actually what our current proposal jobs do anyway
14:05:49 <dhellmann> I hoped as much :-)
14:06:01 <dhellmann> ok, is there anything you need from us?
14:06:31 <fungi> not at present. i think i have a good chunk to hack on before i'm blocked on anything release-side
14:06:40 <dhellmann> sounds good
14:06:45 <dhellmann> moving on then
14:06:51 <dhellmann> #topic remaining tasks for N-1
14:06:58 <dhellmann> we have a short list with a bunch of notes
14:07:09 <dhellmann> #link https://etherpad.openstack.org/p/newton-relmgt-plan
14:07:10 <ttx> did add some context there
14:07:15 <dhellmann> #info (ttx/dhellmann) audit the projects like this for their release model tags
14:07:29 <ttx> right so I went through the list and see
14:07:29 <dhellmann> I was reading over the notes for this earlier today
14:07:56 <ttx> I think we should standardize for Horizon plugins and leave the rest undefined
14:08:01 <ttx> for the time being
14:08:21 <ttx> 3 options on how to handle the horizon stuff
14:08:24 <dhellmann> yeah, that makes sense -- we have enough of them to suggest consistency
14:08:30 <dhellmann> you list 3 options there
14:08:58 <dhellmann> I like option 2 or 3, I think
14:09:02 <ttx> the first one is to skip defining type:horizon-plugin and encourage everything to be shipped within same deliverable
14:09:12 <ttx> since about a third already do that
14:09:24 <ttx> But that will likely be painful version-wise for little gain
14:09:44 <dhellmann> I agree
14:09:47 <ttx> between 2 and 3 the question is... should we push for release model alignment at all ?
14:10:08 <dhellmann> what benefit does aligning give us?
14:10:24 <ttx> I'm not sure we should push, but it seems valuable to point out the inconsistency and check if it's not an oversight
14:10:28 <dhellmann> and why does aligning for a horizon plugin make more sense than for say a client lib?
14:10:35 <dhellmann> sure, that makes sense
14:10:49 <dhellmann> so 3, with a nudge to align?
14:10:51 <ttx> on a case-by-case basis
14:11:12 <dhellmann> I guess that's 2 but with a soft touch
14:11:16 <ttx> yeah, just a check that the model actually reflects what the people want
14:11:27 <dhellmann> ok
14:11:32 <ttx> low-prio for N2
14:11:54 <dhellmann> yep
14:12:49 <ttx> I can take both
14:13:17 <dhellmann> ok. I added them to https://etherpad.openstack.org/p/newton-relmgt-plan
14:13:20 <ttx> That is all I think is necessary for Newton on that topic
14:13:24 <dhellmann> line 113 and 114
14:13:41 <dhellmann> #info N1 TODO review projects with cycle-based models to ensure they are all set up for relmgt to manage their releases (dims)
14:13:57 <dhellmann> dims, did you have a chance to look at this?
14:14:34 <dims> hi dhellmann : nope. time slipped away this week. will bounce it to top of my heap for next week
14:14:56 <dhellmann> ok, sounds good
14:15:02 <dhellmann> #info N1 TODO does merging stable tags into the master branch confuse reno?
14:15:06 <dhellmann> I did not get to this.
14:15:29 <dhellmann> I'm going to PyCon next week, so I'm unlikely to do it before the N1 date, but it'll be at the top of my N2 priority list
14:15:40 <dhellmann> #info N1 TODO (ttx) write a spec proposing how locking stable/newton without changing ACLs might work
14:15:53 <dhellmann> it sounds like that's in a similar situation
14:16:05 <ttx> I'll post something early next week
14:16:06 <dhellmann> we have time, so I'm not worried
14:16:14 <ttx> I may work on it today actually
14:16:16 <dhellmann> ok, good
14:16:59 <dhellmann> that covers our n1 tasks, and I think we're actually in pretty good shape -- we did most of the things we needed, and the things that are slipping aren't necessarily critical
14:17:08 <dhellmann> #topic N1 schedule
14:17:32 <dhellmann> as I mentioned, I'll be at PyCon next week. I'm flying back Thursday, and should have wifi (cross your fingers) to help with tagging
14:18:04 <dims> dhellmann : ack, i'll help process releases
14:18:10 <dhellmann> great, thanks, dims
14:18:36 <dhellmann> My goal for this milestone is to do as little as possible to advertise the deadline, aside from the weekly email and one more email next Wednesday. Then we'll see how the liaisons and PTLs do with filing their milestone tag request, for a milestone where missing doesn't matter too much.
14:19:04 <dims> :)
14:19:13 <dhellmann> does that seem reasonable?
14:19:23 <dhellmann> is there anything else we need to line up before next thursday?
14:19:38 <ttx> yes, we want to use that as a data point to send ultimatums to disconnected teams to get their act together or we'll consider removing them
14:19:47 <ttx> (at TC level)
14:20:28 <dhellmann> if we're going to go that far, I should include that in the deadline email
14:20:31 <ttx> (we'll use mitaka final as another data point, absence of meetings too etc)
14:20:44 <ttx> so it's not the only data point
14:20:51 <ttx> just 'yet another sign'
14:21:10 <dhellmann> so how about I say that missing N1 and N2 deadlines would be grounds for the release team to suggest that the tc consider removing the project?
14:21:42 <dhellmann> and that only applies to the release:cycle-with-milestone teams
14:22:12 <ttx> if that was the release team suggesting it, sure... but it's just a data point amongst others that me and thingee expect to use to send warnings to a number of projects that are already in the red zone
14:22:27 <dhellmann> or do we want to say that all teams need to be preparing releases for n1?
14:22:29 <ttx> communications-wise
14:22:47 <ttx> so I'm not sure a warning from the release team is necessary
14:23:04 <dhellmann> the warning will come from elsewhere?
14:23:10 <ttx> if missing n1 is the first issue we wouldn't send anything
14:23:29 <dhellmann> ok, so we'd warn around n2
14:23:41 <ttx> dhellmann: It would come from the TC. we'll suggest a warning to a number of teams based on past issues
14:23:43 <dhellmann> makes sense -- let's see if we need to worry about warnings at all
14:24:16 <dhellmann> ok, so just the normal email about the schedule for now
14:24:22 <ttx> that's why we delayed those warnings to post-n1
14:24:30 <ttx> to have yet another data point
14:24:33 <dhellmann> we should think about the non-milestone release models and what we expect from those teams, too
14:25:28 <dhellmann> ok, anything else for n1?
14:25:31 <ttx> nope
14:25:34 <dhellmann> #topic priority reviews
14:25:41 <dhellmann> I had a few here, but some merged
14:25:47 <dhellmann> it looks like the django-openstack-auth release may be needed to unblock horizon https://review.openstack.org/#/c/322152/
14:25:53 <dhellmann> based on something I saw on the ML earlier today
14:26:14 <dhellmann> if they're ready for a release later today, we should probably consider it even though it's friday
14:26:39 <ttx> +1
14:26:44 <dhellmann> will either of you be available for that? I'm flying today...
14:27:02 <ttx> I'll probably be off by the time that is ready
14:27:20 <ttx> and at the very least unable to handle any fallout
14:27:32 <ttx> so if it's just me safer to skip
14:27:38 <dhellmann> ok, if it's really late they can wait for monday anyway (or I'll do it Sunday since I expect to be online then)
14:28:06 <dhellmann> I'll check in over the weekend or monday morning
14:28:14 <dhellmann> that segues nicely into the next topic
14:28:15 <dhellmann> #topic Spread out release days across team members
14:28:27 <dhellmann> ttx, I think you added this one?
14:28:33 <ttx> yeah, so I added that because I slack on releasing
14:28:54 <ttx> I figured we could try to have a specific day where we handle stuff
14:29:07 <dhellmann> that's definitely worth trying
14:29:08 <ttx> that way the others are mostly off-duty
14:29:37 <ttx> We have 3 days (Tue-Wed-Thu
14:29:39 <ttx> )
14:29:56 <ttx> Given my position on the globe I should probably not take Thursday
14:30:09 <ttx> Tuesdays are generally busy for me
14:30:18 <ttx> but I can certainly be on-ècall on Wednesday
14:30:52 <ttx> it's just a "barring anything else, I should be the one processing release requests on that day"
14:30:52 <dhellmann> ok, I can do Tuesday or Thursday
14:31:05 <dhellmann> right
14:31:05 <ttx> Tuesday also happens to be the busiest day with backlog
14:31:36 <dhellmann> I'll take Tuesday, since I do try to release some of the backlog on Mondays as well
14:31:47 <dhellmann> dims, can you take the lead on Thursdays?
14:32:01 <dims> dhellmann : sure, let's try
14:32:05 <dhellmann> cool, thanks
14:32:30 <ttx> The idea is that unless specifically pinged (absences etc) you should be able to ignore releases requests on off-days
14:32:42 <dims> i can definitely help with things late in the day any day of the week if things pile up
14:32:49 <ttx> If you can't do, it's your duty to pass the torch
14:33:07 <dhellmann> dims : would you prefer to split tue/thu with me where I do mornings and you do afternoons?
14:33:16 <ttx> Right, my day being generally earlier than others, I expect some releases requests to pile up on Wednesday evening too
14:33:27 <dims> dhellmann : just thu is fine
14:33:33 <dhellmann> ok, that's simpler
14:33:39 <dhellmann> let's give this a try for a couple of weeks and see how it goes
14:33:41 <dhellmann> good suggestion, ttx
14:33:43 <ttx> dhellmann: that should improve our general reactivity I think
14:33:51 <ttx> And also force me to do my share
14:33:54 <dhellmann> :-)
14:34:11 <ttx> because currently I'm more like.. "someone will pick them up"
14:34:15 <dhellmann> I won't publish a schedule for now, so we can see how it goes and adjust if needed
14:34:20 <ttx> always something else more urgent to do
14:34:21 <dims> :)
14:34:37 <dhellmann> right
14:34:45 <dhellmann> ok, that's the end of our formal agenda
14:34:48 <dhellmann> #topic open discussion
14:34:57 <dhellmann> is there anything else we need to cover this week?
14:35:24 <ttx> nope
14:35:28 <dhellmann> heh, now that I've promised to do tuesdays, I'll be speaking next tues so may need someone to help cover
14:35:35 <ttx> haha
14:35:42 <dhellmann> we'll see -- I should be able to do some work while listening to other talks
14:35:52 <dims> lol
14:36:20 <dims> dhellmann : if you can do a +2 and ping me on the review, i can do the legwork
14:36:37 <dhellmann> ok, that sounds good
14:37:05 <ttx> dhellmann: we are still good on single-review for release approvals, right
14:37:28 <dhellmann> ttx: yeah, although I've been waiting for the stable team to review the stable releases
14:37:37 <ttx> ok, noted
14:38:02 <dhellmann> I figure when we have the automation done, we'll give them +2 and then they can help out with stable releases, but there's no sense in getting them set up with the tooling in the mean time
14:38:17 <dhellmann> they have enough other work on their plates
14:38:30 <dhellmann> tonyb agreed with that general plan, I think
14:39:18 <dhellmann> ok, if there's nothing else we can call the meeting done early
14:40:23 <dhellmann> I leave for the airport in about 2 hours, so I'll be in and out between now and then, but ping me if you need something
14:40:27 <dhellmann> thanks!
14:40:28 <dhellmann> #endmeeting