14:01:14 #startmeeting releaseteam 14:01:14 Meeting started Fri Nov 19 14:01:14 2021 UTC and is due to finish in 60 minutes. The chair is elodilles. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:14 The meeting name has been set to 'releaseteam' 14:01:16 o/ 14:01:22 o/ 14:01:23 hi o/ 14:01:30 #link https://etherpad.opendev.org/p/yoga-relmgt-tracking 14:01:46 we are at line 73 14:02:20 let's start with the 1st topic 14:02:25 #topic Review task completion 14:02:43 #info Ensure that all trailing projects have been branched for the previous series 14:02:49 that was on me 14:03:05 I've proposed the patches, I hope i did it right 14:03:14 #link https://review.opendev.org/q/topic:trailing-xena 14:03:19 these are the patches ^^^ 14:04:12 IIRC our deadline is pretty fuzzy on those, at least the patches will start the discussion 14:04:43 +1 14:04:49 yes, the trailing delivery deadline is December 16th 14:05:11 so we are one month ahead of that 14:05:58 ok, if nothing else for this then i'll continue with the next task 14:06:37 #info Propose autoreleases for cycle-with-intermediary libraries which did not release since the previous release 14:07:15 i've generated the patches a bit late for this, instead of beginning of week, i created those yesterday afternoon 14:07:31 #link https://review.opendev.org/q/topic:yoga-milestone-1 14:08:00 ok so we might give them a bit more time before force-approving those without -1 14:08:23 (in theory we would have done that now) 14:08:32 (and give those with -1 a few extra days) 14:08:38 note: there were quite many that were listed as not released yet, but either didn't have any change at all or just CI/tests/branch setup releated patches were merged 14:08:54 yeah those can be ignored 14:09:12 the automaton patch can be abandonned 14:09:18 I left a comment 14:09:18 ttx: yes, thanks, let's give some time until the force-approve 14:09:40 https://review.opendev.org/c/openstack/releases/+/818403 14:09:55 hberaud: looking 14:11:10 hberaud: thanks, abandoned! 14:11:15 ty 14:12:58 do we want to agree on a force-approval deadline, like Tuesday, or something? 14:13:32 I'd say approve all without -1s by Tuesday, then finalize those with -1s by end of week 14:13:35 Wednesday, seems more appropriate 14:13:50 or the end of week 14:13:52 I don;t want to delay too much 14:14:07 basically push by two business days 14:14:08 wfm 14:14:16 ok 14:15:07 #agreed yoga-milestone-1 patches: approve all without -1s by Tuesday, then finalize those with -1s by end of week 14:15:39 ok, next topic 14:15:48 #info To catch if there are acl issues in newly created repositories 14:15:58 ttx: this was yours 14:16:10 yep, the script revealed one gap 14:16:20 https://review.opendev.org/c/openstack/project-config/+/818356 (tap-as-a-service) 14:16:56 (basically the idea is to make sure releases go through openstack/releases, so remove direct tagging rights 14:17:02 ) 14:17:21 nothing more to say 14:17:32 #link https://review.opendev.org/c/openstack/project-config/+/818356 14:17:39 ttx: cool, thanks! 14:17:58 for the explanation, too 14:18:45 ok. the countdown email is not yet ready, as I'll send it after the meeting and when we are there :) 14:18:50 so next topic 14:18:55 on a related note, was there ever a formal conclusion as to whether the release team can handle sigs' repos? 14:18:57 #topic Assign R-18 tasks 14:20:04 fungi: i don't know. hberaud , ttx , do you remember such? 14:20:14 nope 14:20:47 We only do openstack release 14:20:56 i.e. deliverables from project teams 14:21:09 that's they key difference between project teams and sigs 14:21:32 Agreed with ttx 14:21:40 so we tend to explictly stay away from SIG repos 14:21:56 okay, i recall it came up on the ml last week, i thought it was ansible-collections-openstack but might have been somthing else 14:21:58 that's why QA is still a project team 14:22:41 (and with release management the onlt horizontal team left) 14:22:59 it's because their repos are handled by us as part of "the release" 14:23:12 oh, i see where i got confused 14:23:20 (+ requirements) 14:23:38 elodilles posted on openstack-discuss about a release job failure for ansible-collections-openstack but that was one they tagged themselves i guess 14:23:52 (I tried to clean things up over the years and I think we finally got there) 14:24:20 yes release jobs failure can be triggered by openstack/releases or direct tags :) 14:25:02 yes, and since it was not familiar to me i rather forwarded it to ML o:) 14:25:02 yeah, he mis-attributed it to the ansible team in the e-mail, but that was it 14:25:32 i thought maybe it was one which had been handled via openstack/releases but looks like it wasn't 14:25:51 confusion cleared, thanks 14:26:05 the only way to be sure bnothing falls into the cracks is really to have a finite, defined list of things we handle 14:26:25 otherwise you're just never sure 14:26:26 yes, i agree 14:26:42 yepp, that's for sure 14:26:48 it's been tricky enough to compile that list (see release-management key in projects.yaml) 14:27:00 there were exceptions all over the map :) 14:27:20 but we are pretty good now... just need to periodically check all data is aligned 14:27:28 like that ACL check we did this week 14:27:50 makes sense, that's what reminded me 14:28:57 ok, so, next topic: assigning R-18 tasks 14:29:03 there is only 1 task 14:29:23 i guess this could be "all", right? o:) 14:29:32 yes 14:29:34 "Review any remaining milestone-1 exceptions" 14:30:20 added "all" to that then. thanks 14:30:41 next topic: 14:30:49 #topic Review countdown email contents 14:30:59 #link https://etherpad.opendev.org/p/relmgmt-weekly-emails 14:31:25 i've added and edited the mail ^^^ 14:31:51 there is one thing: we don't have release cycle goal (page) 14:32:16 It's hard to say we are past Yoga-1 14:32:17 so I'll remove that if you confirm that is OK 14:32:29 I'd wait until next week to send this one 14:32:48 Like wait until we process things on tuesday 14:32:49 ttx: hmm, yes, since the shifted deadline to Tuesday 14:33:17 ttx: that works for me 14:33:22 WFM 14:33:27 otherwise looks good! 14:33:33 ok, thanks 14:33:39 +1 14:34:18 #action to elod: Send countdown mail on Tuesday, after Yoga-1 milestone patches have been merged 14:34:39 (hope i remember correctly for the keywords :)) 14:34:47 anyway 14:34:51 next topic then 14:35:23 #topic Branches without releases https://review.opendev.org/c/openstack/releases/+/818445 14:35:51 yeah I was wondering about that one 14:36:14 I think we only branch on release points, on purpose 14:37:02 openstack-ansible-roles have similar branching, i mean, without any release, 14:37:13 just the hashes, set as location 14:37:36 the error looks more like a bug than a validation rule failure 14:37:45 latest_release = deliv.releases[-1] 14:37:48 IndexError: list index out of range 14:38:07 the openstack-ansible-role is tagless 14:38:17 https://opendev.org/openstack/releases/src/branch/master/deliverables/wallaby/openstack-ansible-roles.yaml#L3 14:38:27 hberaud: you are right 14:38:46 the tripleo-docs is a cwi deliverable 14:38:59 this is why it complain with it 14:39:34 I don't think this is possible to branch without releasing something 14:40:07 (in the tripleo-docs context) 14:40:26 at the very minimum we should catch that case because currently the script just errors out 14:40:50 indeed 14:40:57 but yes, the error matches 14:41:11 "stable branches must be created from the latest tagged release..." 14:41:25 here the script errors out trying to grab the latest release 14:41:34 but there is none 14:42:04 If it was tagless it would pass 14:42:14 yes 14:42:28 or "upstream" 14:42:32 on the other hand do we want to allow this late branch cutting (for wallaby, in this case)? 14:42:46 it can't hurt 14:42:55 ok 14:43:07 ok, who replies? 14:43:12 this is a trailing deliverable 14:43:20 this is why we missed to branch it 14:43:50 i see 14:44:06 it went under our radar 14:45:21 ok, so I can comment on the patch that this requires a release as well. <- is this OK? 14:45:36 wfm 14:46:11 wfm 14:47:50 good. and another thing is the validation, it could be fixed with some clear error message then, i guess 14:49:11 i'll try to look into it. of course, if someone has an idea for how to fix, then please propose a patch for that :) 14:49:48 ok, next topic 14:49:54 #topic Proposing skipping meeting next week (just need to make sure we process all yoga-1) 14:50:40 I won;t be around anyway, but there is also no task the week after 14:50:54 wfm 14:50:57 trick is we still need to doublecheck that all yoga-1 is handled 14:51:23 yes, I mean, now we'll have the yoga-milestone-1 patches, but that can be handled without the meeting :) 14:52:12 ++ 14:52:33 ok 14:54:08 i've added a note to meeting agenda after (skip + make sure everything is good with yoga-milestone-1) 14:54:20 s/after// 14:54:35 last topic: 14:54:39 #topic Open Discussion 14:55:13 anything to mention? anyone? 14:55:37 nope 14:56:03 nothing here 14:57:13 okay, then thanks everyone, let's close the meeting 14:57:22 #endmeeting