15:01:56 #startmeeting releaseteam 15:01:57 Meeting started Fri Oct 27 15:01:56 2017 UTC and is due to finish in 60 minutes. The chair is smcginnis. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:58 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:58 o/ 15:02:01 The meeting name has been set to 'releaseteam' 15:02:04 o/ 15:02:13 o/ 15:02:28 Dang it, someone cleared my ping list. :) 15:02:36 * mlavalle apologizes. won't do it again next time :-( 15:02:53 mlavalle: You're fine. We never really need a full hour anyway. 15:03:01 Agenda: https://etherpad.openstack.org/p/queens-relmgt-tracking 15:03:13 We are in week R-18 15:03:23 smcginnis: i have a timed ping anyway, so no worries 15:03:26 So pretty quiet week. :D 15:03:57 #topic Queens-1 remaining items 15:04:01 * dhellmann wants to live in smcginnis' alternate universe 15:04:09 ;) 15:04:24 tripleo and sahara should be good to go 15:04:41 the other 3 are still blocked afaict 15:04:53 I think it's been kind of implied, but we're going to still do some of those today even though it's Friday, right? 15:04:54 those telemetry projects (aodh, etc.) have been publishing to pypi from the beginning, irrc 15:05:18 it seems a lot less risky to release on fridays with the constraints system in place 15:05:29 ++ 15:05:29 and services aren't that big of an issue anyway, right? 15:05:33 right 15:05:37 Right, shouldn't be. 15:05:45 it would be good to finish q1 before we all leave for sydney 15:05:48 I'm all for +2a those two 15:06:07 I sure hope Q-1 is done before I land in Sydney. 15:06:27 I think for the most part, our remaining issues are all just mismatches between service types and pypi jobs. 15:06:36 arh dhlemmbeat me to the recheck 15:06:38 I rechecked sahara 15:06:38 Nothing scary that I've noticed so far. 15:06:49 tab fail 15:07:29 So we have a good list on the etherpad. Let's work our way through that today. 15:07:30 so is the suspicion that these service types were mismatched to begin with, or that we somehow regressed them with the conversion to zuul v3 native job definitions? 15:07:35 monasca-ui should jus tneed a release-type set 15:07:42 (let's meaning I will unless someone else beats me to it) 15:07:53 Guest53850: There were some of that. 15:08:02 Oops, sorry Guest53850 15:08:08 Tab fail for me as well. 15:08:11 fungi: There was some of that. 15:08:17 smcginnis: feeling lucky? you can approve the tripleo one 15:08:19 that's it, new tab keys for everbody 15:08:29 er, everybody 15:08:40 And some were publishing to pypi on purpose. 15:08:46 Hah, new keyboard for you too fungi. 15:08:56 ttx: Yeah, probably time to bite the bullet on that one. 15:08:57 yup, i'm due 15:09:12 * lbragstad puts in an order for limit one cherry blue tab key 15:09:14 EmilienM has been waiting patiently. 15:09:15 yeah, it was a little bit of everything but I think more of the mismatches were intentional than we realized and were only exposed with the changes in the validation in the releases repo 15:09:24 lbragstad: :) 15:09:25 o/ 15:09:51 #topic Standing known issues 15:10:01 #link https://etherpad.openstack.org/p/release-job-failures 15:10:13 I've been adding a few things there. 15:10:51 to explain the situation with the job matching, the way things were built up in v2 was through composing a set of discrete jobs so teh translation to v3 where those jobs could be successfully distilled into fewer job definitions meant attempting to find common patterns across projects as to which ones applied which combinations of jobs. since there wasn't great consistency to start with, i think 15:10:52 some (incorrect) guesses were made in places 15:10:57 I do have a project-config change for vitrage, but maybe I should just fix the deliverable to publish to pypi. Thoughts on that? 15:11:32 fungi: Yeah, totally reasonable and understandable. 15:11:37 oh, hrm, yeah, I just updated monasca-ui's release request to set the release-type to python-pypi but maybe that's the wrong change 15:11:49 i think the end result though is that we're getting better consistency (finally) across the board 15:11:55 I think what we're seeing now is just the effects of trying to apply consistency to what was not consistent before. 15:12:01 right, that 15:12:59 dhellmann: Not sure. At this point maybe changing the deliverables is the path of least resistance. 15:13:09 * fungi checks to see if release job failures are getting e-mailed again 15:13:22 fungi: Oh yes, they are. They very much are. :) 15:13:28 ooh! 15:13:45 #link http://lists.openstack.org/pipermail/release-job-failures/2017-October/date.html#start 15:13:48 indeed 15:13:51 * smcginnis had to shut off his sound for awhile after the ansible jobs. 15:14:09 monasca-ui has been published to pypi in the past, so maybe that's right 15:14:19 dhellmann: Works for me. 15:14:46 do you want to go through them all together here, or one by one after the meeting 15:15:06 We can probably do that post-meeting. 15:15:41 k 15:16:13 #topic Discuss RelMgt team Sydney Forum sessions coverage 15:16:28 ttx: Want to cover this? 15:16:38 sure 15:16:54 I tried to flag sessions of potential interest for us 15:17:08 I have te first and last on my agenda 15:17:26 not sure about the middle two 15:17:36 Those two are also if interest to me. 15:17:37 also if I miss anything let me know 15:17:54 And one of the biggest issues raised by the ops midcycle folks. 15:18:34 ttx: That TC session email helped as well. Thanks for compiling those lists. 15:18:55 I still need to build my schedule for Sydney. 15:19:02 I have conflicts for the middle 2 15:19:11 I'll make sure pkovar sees the release notes & docs item 15:20:04 I'm assuming the zuul v3 one should have a lot of infra folks interested in it. 15:20:21 there's always room for zuul 15:20:27 zuulv3 I think we can skip, we sould only be grumbling how how painful the transition was for us 15:20:29 my only real feedback there is that the future looks bright but the present is a bit less so 15:20:35 would* 15:20:56 we're past the dystopian hellscape phase, I think 15:20:58 A bit of a transition period, but yeah, should be good in Rocky. 15:21:06 ok, anything on your schedule that I should definitely be attending? 15:21:25 lunch 15:21:31 :) 15:21:41 dinner too 15:21:43 there's a privsep session on monday 15:21:49 yes, I have it 15:21:52 * smcginnis makes a note 15:22:21 note that I lost expertise on rootwrap once it grew the daemon mode and became fancy 15:22:21 the stestr session may be interesting from a tc/project goals perspective 15:22:25 I'll hopefully have a plan for the week figured out by the end of today. 15:22:39 the conflict on tuesday at 11:40 is the product working group 15:22:41 Would be good to have consistency across projects there. 15:22:47 I haven't decided whether to spend time on that again or not 15:22:51 heh 15:23:27 maybe I'll take my chances and hope that I miss something useful this time 15:24:12 surprisingly I only have 3-4 time slots double booked this time 15:24:17 ttx: Do we have attendance #'s for Sydney? 15:24:29 Or some way to know how many are there for the Forum? 15:24:46 Just in casual conversations, there are quite a few "regulars" that are not going this time. 15:25:07 I don't have data yet no 15:25:07 i'm hoping we make up for that with irregulars showing up 15:25:13 haven't asked tbh 15:25:19 yes, the different mix will be interesting 15:25:54 well, we do know at least they project lower attendance just based on the space allocated and shortened duration 15:26:15 but as for specifics, no idea 15:26:27 We will have some of the regular irregulars too. :) 15:26:42 OK... 15:26:44 #topic Skipping next two weeks meetings 15:26:54 Probably a good idea. 15:27:03 If anything comes up, we can just handle in channel. 15:27:03 ok 15:27:13 when we get back to it, we'll all be off DST 15:27:18 With everything else going on, probably won't have much need for regular meetings. 15:27:28 so I was wondering if this hour would still work for yall 15:27:31 i'll be on an airplane during the next meeting anyway, home for the second one but likely not very useful 15:27:36 It works for me obviously 15:27:52 Yes, time will still work well for me after the switch. 15:28:01 ++ 15:28:03 dhellmann: ? 15:28:14 * dhellmann catches up 15:28:15 i have no conflicts sticking with 1500z on fridays 15:28:19 I know you have a meeting tunnel on Fridays 15:28:47 so, it will be an hour earlier in real life? or later? 15:28:54 it will be earlier 15:29:11 ok, I can do that 15:29:12 Fall back, so it will be 10 your time? 15:29:12 my regular meeting in that hour has been rescheduled 15:29:18 after DST, it will be later by an hour 15:29:23 ok let's keep it then 15:29:30 lbragstad: Oh right. :) 15:29:51 ical says 10:00 on the 10th 15:30:00 and I have no conflicts then 15:30:05 time to approve tripleo smcginnis 15:30:27 Yep, will do as soon as we're done. Unless you have it up and want to. 15:30:40 #topic Open discussion 15:30:49 That's it on the agenda. Anything else? 15:31:23 Just wanted to say thanks to both of you for fixing so many things over the past two weeks. It was hard to keep up with all you did while I was sleeping, so i could not help as much as I wanted 15:32:10 np; I wish it hadn't taken so long to sort out some of the issues, but I feel a lot better about where things stand today than last week 15:32:41 brb 15:32:42 All thanks to dhellmann, fungi and the infra team. 15:32:47 ++ 15:32:51 ++ thanks fungi 15:32:57 et al 15:33:05 ack 15:33:17 Alright, I think we can wrap up and get back at it if that's all. 15:33:22 Thanks everyone. 15:33:25 o/ 15:33:38 #endmeeting