16:00:13 <smcginnis> #startmeeting releaseteam
16:00:13 <openstack> Meeting started Thu Apr 11 16:00:13 2019 UTC and is due to finish in 60 minutes.  The chair is smcginnis. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:14 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:16 <openstack> The meeting name has been set to 'releaseteam'
16:00:19 <smcginnis> Ping list: smcginnis ttx dhellmann diablo_rojo hberaud evrardjp fungi armstrong
16:00:25 <evrardjp> o/
16:00:28 <smcginnis> #link https://etherpad.openstack.org/p/stein-relmgt-tracking Agenda
16:00:32 <ttx> o/
16:00:37 <fungi> aloha
16:00:40 <diablo_rojo> o/
16:00:58 <smcginnis> #topic Release Postmortem
16:01:04 <dhellmann> o/
16:01:32 <ttx> Any objection to approving https://review.openstack.org/#/c/650173 now?
16:01:34 <smcginnis> I think overall the final release went well. A few items yet to wrap up.
16:01:46 <smcginnis> evrardjp had a comment on there.
16:02:05 <smcginnis> These were supposed to have happened with the final cwi release, so I don't think we need to wait for any PTLs on that.
16:02:06 <evrardjp> it's fine
16:02:11 <ttx> ++
16:02:19 <evrardjp> yeah we don't need to wait
16:02:27 <ttx> smcginnis: you do it or should i?
16:02:31 <evrardjp> I am not sure why I am always careful in those branching cases :)
16:02:39 <smcginnis> ttx: Feel free to take care of that.
16:02:50 <ttx> done
16:03:09 <smcginnis> evrardjp: It's good to be careful earlier on. Really a pain if your repo gets branched and suddenly you realize you need to backport a bunch of things in a time crunch.
16:03:12 <smcginnis> At this point though...
16:03:29 <ttx> next up is two release page cleanups
16:03:49 <smcginnis> I can approve those. Unless dhellmann wanted a chance to look?
16:04:16 <smcginnis> #link https://review.openstack.org/#/c/651551/
16:04:21 <smcginnis> #link https://review.openstack.org/#/c/651549/
16:05:05 <smcginnis> OK, I'll approve.
16:05:32 <smcginnis> ttx: Want to talk about Post-release tasks?
16:05:58 <ttx> sure
16:06:16 <ttx> there were a couple lines above though
16:06:30 <ttx> Are we all done ?
16:06:32 <ttx> Anything missing in the process ?
16:06:51 <smcginnis> Oh, stable/stein releases. Sorry, I skipped that. We can circle back.
16:07:06 <smcginnis> I do think we are ready to init-series for train.
16:07:27 <smcginnis> I can grab that unless anyone else wants it.
16:07:28 <ttx> yeah...
16:07:34 <ttx> Once the stable branch process
16:07:46 <smcginnis> Yeah
16:07:49 <ttx> I mean https://review.openstack.org/#/c/650173/ is in
16:07:51 <evrardjp> I would prefer if you could do it smcginnis , as I am on holiday tomorrow and next week
16:07:58 <dhellmann> was the tripleo team ready for branches yet?
16:08:08 <smcginnis> evrardjp: ack
16:08:13 <ttx> dhellmann: those were cwi ones
16:08:17 <dhellmann> ok
16:08:54 <smcginnis> We haven't done any cycle-trailing ones yet. Well, "we" haven't. I think one or two have been done by those teams.
16:08:55 <ttx> maybe was oversight though. We'll soon know
16:10:00 <ttx> So yes, init-series once the stable branches are all in
16:10:14 <smcginnis> OK, back to the stable/stein releases?
16:10:14 <ttx> then we have a few stalled requests
16:10:24 <ttx> python-searchlightclient 1.5.1 https://review.openstack.org/#/c/649487/
16:10:26 <ttx> python-cloudkittyclient 2.1.1 https://review.openstack.org/#/c/650239/
16:10:35 <ttx> that we can now unblock. Could also wait for Monday
16:10:44 <ttx> but I see no reason why
16:10:57 <evrardjp> I think it would be fine to do it nowadays?
16:11:01 <smcginnis> Neither assert stable:follows-policy
16:11:10 <dhellmann> yeah, I'm in favor of going ahead, assuming the releases look right
16:11:22 <smcginnis> Since they were caught in the end of cycle freeze, migh be good to get them out right away.
16:11:22 <evrardjp> ofc
16:11:30 <ttx> ++
16:11:34 <ttx> I can push it
16:11:54 <ttx> done
16:12:09 <smcginnis> OK, on to ocata EM.
16:12:31 <smcginnis> I broke out the openstack-ansible ones from the main patch. Since they always have to be so difficult. :P
16:12:40 <evrardjp> ha!
16:12:41 <smcginnis> https://review.openstack.org/#/c/650467/ and https://review.openstack.org/#/c/651778/
16:13:13 <smcginnis> We switched to EM about a year ago. But the docs said we would tag them but we never did that part.
16:13:13 <dhellmann> did we tag one final release on all of those before the em? or are we assuming that whatever was there is good enough?
16:13:30 <smcginnis> I think we had declared them EM, just never tagged them.
16:13:38 <smcginnis> IIRC, it was on the ML.
16:13:53 <dhellmann> what I mean is, are we tagging at a point that means we have unreleased changes in any of those branches?
16:14:42 <ttx> yes that is the big question
16:14:48 <smcginnis> I did not check, but we had already said we would not be doing any more releases at this point. So if there are, I would consider those commits as part of that extended maintenance work.
16:15:00 <ttx> that makes sense
16:15:09 <ttx> it's not lost. It's just EM
16:15:23 <dhellmann> ok
16:15:41 <dhellmann> just thinking through what we discussed the other day on the review for changes to the tool that adds the tags
16:15:43 <evrardjp> it's kinda weird for OSA though, because we are tagging really old things.
16:15:58 <evrardjp> I have to think about that
16:16:12 <smcginnis> dhellmann: That was updated to tag -eol at the latest commit and not that last tag's commit.
16:16:31 <dhellmann> yeah, we also discussed the em tagging policy in there
16:16:33 <smcginnis> evrardjp: Yeah, all of these are really old at this point.
16:16:57 <smcginnis> Since we can't release once we are in em, I think the em tag needs to be at the final release.
16:17:03 <dhellmann> the osa case is definitely unique. I wonder if we need an em policy setting, like we have for branching, to tell us the type of location allowed
16:17:09 <openstackgerrit> Merged openstack/releases master: Add missing stable/stein branches  https://review.openstack.org/650173
16:17:10 <openstackgerrit> Merged openstack/releases master: Clean up failed releases  https://review.openstack.org/651549
16:17:10 <dhellmann> or at least  whether to use HEAD or $last-tag
16:17:12 <openstackgerrit> Merged openstack/releases master: Do not generate tarball links for tripleo-ui  https://review.openstack.org/651551
16:17:14 <smcginnis> Anything after that point is unreleasable.
16:17:36 <smcginnis> If we tag em a few commits after the last release, those commits are kind of in a limbo.
16:17:44 <dhellmann> I agree, mostly, that em should tag the latest release. I'm not sure we want to tag ancient releases, though.
16:18:00 <evrardjp> osa is very weird because also we stopped tagging because the people stopped caring. But sending and EM to these old versions send a wrong message IMO
16:18:09 <dhellmann> right
16:18:14 <evrardjp> while the openstack-ansible repo one is correct
16:18:31 <evrardjp> I will request a release tomorrow latest for just this branch, if it's fine with everyone
16:18:45 <evrardjp> so that we can use those new shas for em
16:18:48 <dhellmann> at some point osa won't even have tags on some of those repos, so we need to think about how to handle that case
16:18:49 <evrardjp> sorry for being late
16:18:53 <ttx> smcginnis: you were right about them being difficult :P
16:18:59 <evrardjp> ttx: :D
16:19:00 <smcginnis> So I would abandon the osa one I have, then you would propose one just for the single repo evrardjp?
16:19:09 <evrardjp> smcginnis: correct
16:19:09 <smcginnis> ttx: ;)
16:19:29 <smcginnis> OK, that works for me.
16:19:30 <dhellmann> I think we want to tag all the osa repos. The main one should be tagged at its most recent release. The others should be tagged at their HEAD.
16:20:00 <smcginnis> I would still think their last tag, not HEAD.
16:20:01 <evrardjp> dhellmann: correct
16:20:09 <dhellmann> when stein (or train) goes EM, those OSA repos won't have any version tags at all but we still want to tag them EM
16:20:20 <dhellmann> because we're no longer tagging versions in some of those repos
16:20:35 <evrardjp> that's my plan, create a late release with HEAD on those repos, and then use those for EM
16:20:36 <smcginnis> If a repo doesn't have tags, then I don't think it needs -em.
16:20:47 <evrardjp> smcginnis: but they had tags :)
16:20:54 <dhellmann> hmm
16:21:10 <smcginnis> In the case of ocata, it doesn't seem right to tag a point -em if it hasn't been "released".
16:21:15 <dhellmann> I guess I thought we wanted to mark everything at once, but if we don't care about marking tagless repos then mayb enot
16:21:22 <dhellmann> it has been released, though
16:21:24 <smcginnis> Once we get to the point where there were no releases done at all, I don't think we need an em tag.
16:21:33 <dhellmann> the fact that there's no tag doesn't mean it isn't released in this case
16:21:48 <dhellmann> it is effectively "vendored" in that main repo
16:22:01 <dhellmann> because that includes the SHA of the commit from each of the other repos to use
16:22:02 <evrardjp> yeah, so what we could do is to just EM the openstack-ansible repo
16:22:07 <dhellmann> yeah, ok
16:22:10 <evrardjp> that would be enough too
16:22:21 <evrardjp> because it doesn't set a wrong expectations in the roles
16:22:25 <openstackgerrit> Merged openstack/releases master: Release python-searchlight client 1.5.1 for stein  https://review.openstack.org/649487
16:22:26 <evrardjp> (other repos)
16:22:26 <openstackgerrit> Merged openstack/releases master: Releasing python-cloudkittyclient 2.1.1  https://review.openstack.org/650239
16:22:30 <smcginnis> openstack-ansible controls what from those other repos is used, right?
16:22:35 <evrardjp> correct
16:22:46 <smcginnis> Yeah, then I would think we would only tag openstack-ansible.
16:22:51 <evrardjp> and we are not using tags.
16:23:08 <dhellmann> what is the point of adding the em tags at all? it's to communicate that the *repo* is closed to certain types of changes and to releases, right?
16:23:21 <ttx> yes
16:23:32 <evrardjp> so this is why I think the em tags still apply in this case dhellmann
16:23:43 <dhellmann> so if we don't tag all of the repos, we're not sending the message about being closed to certain types of changes in those repos
16:23:52 <evrardjp> so two choices: either just do openstack/openstack-ansible as EM and ignore the rest, or do everything on all repos.
16:24:03 <dhellmann> or to the testing requirements, or whatever else we're changing for that repo when we go to em
16:24:25 <dhellmann> I think we should tag them all, but I don't feel strongly enough about it to fix the tool to let us do that, so :-)
16:24:29 <evrardjp> I believe it's semantically more correct to do what dhellmann said and propose earlier
16:24:50 <evrardjp> dhellmann: but can't I propose a last minute release before we do EM?
16:25:11 <smcginnis> So if release-model:untagged, the tool should grab HEAD. Otherwise, use the commit from the last release. Does that make sense?
16:25:12 <dhellmann> that only helps if you tag all of the repos. which I guess would be another approach
16:25:24 <dhellmann> smcginnis : yeah, that's effectively what I'm saying
16:25:27 <smcginnis> evrardjp: Not for ocata. It is already in em.
16:25:39 <evrardjp> smcginnis: I see, I missed the timeframe
16:25:40 <smcginnis> We just never followed through with the tagging.
16:25:45 <smcginnis> Pike on the other hand...
16:26:01 <smcginnis> That was supposed to have transitioned at the beginning of March, but we have a little time yet.
16:26:29 <smcginnis> dhellmann: OK, I'll put it on my todo to update the new-release tool to handle that for the future.
16:26:40 <dhellmann> k
16:27:06 <evrardjp> I think that's fair
16:27:14 <smcginnis> So I will abandon my osa tagging patch for now. evrardjp will follow up with the "correct" tagging for those repos.
16:27:20 <evrardjp> I am sorry to say that ttx prophecy was correct
16:27:30 <smcginnis> hehe
16:29:06 <smcginnis> OK, one more monkey wrench.
16:29:21 <smcginnis> openstack-ansible ocata was not tagless. Still allow it?
16:29:40 <dhellmann> allow?
16:30:14 <smcginnis> If we tag HEAD for tagless and last release otherwise, at least as of the state during ocata, what I have up there now is "correct".
16:30:32 <smcginnis> But knowing those eventually ended up tagless, we could retroactively apply that to ocata.
16:31:18 <dhellmann> if all the repos were always tagged for ocata, we should use the tags. but iirc, that was when we did this transition? so in that case I think we want to recon it
16:31:23 <dhellmann> er, retcon
16:31:53 <evrardjp> I am not sure when we did this, it was in the MLs. At that time we said we couldn't touch the deliverables files though
16:32:20 <evrardjp> so we couldn't remove all the repos from those older branches
16:32:46 <evrardjp> The solution was, for the future, create a tagless file for the future
16:33:02 <smcginnis> OK, let's just stick with that last plan. I've abandoned mine. evrardjp can propose a new one treating those other repos as tagless.
16:33:19 <evrardjp> but because we didn't touch the old deliverables files, we just said to "just not tag there"
16:33:57 <evrardjp> I don't like us to be _special_
16:34:00 <dhellmann> looking at that deliverable file, I see all the repos being tagged up to 15.1.21, and then only openstack/openstack-ansible being tagged after that
16:34:10 <smcginnis> And now we have validation that complains if you try to do a release and don't include all repos. But that should be fine if you tag HEAD for the ones not released. Just takes a little manual editing.
16:34:15 <evrardjp> so it's probably in that two week window dhellmann
16:34:33 <dhellmann> yeah
16:34:36 <ttx> openstack-ansible is ... special, Mrs. Gump.
16:34:45 <smcginnis> :D
16:34:48 <evrardjp> ttx: :)
16:34:55 <evrardjp> Run ttx , run
16:35:16 <smcginnis> Are any of the other deliverables in https://review.openstack.org/#/c/650467/ as "special" as osa? :)
16:35:37 <dhellmann> https://media.giphy.com/media/3oriNLnYCrAj1Rg7qE/giphy.gif
16:35:51 <dhellmann> I think osa is the only snowflake for this case
16:36:13 <smcginnis> So then we should be able to finish that tagging for the rest of those.
16:36:21 <dhellmann> +1
16:36:28 <evrardjp> agreed
16:36:42 <ttx> ++
16:36:50 <smcginnis> I'll leave it up to y'all to +2+W that then. ;)
16:36:52 <dhellmann> I've +2ed that on faith, smcginnis :-)
16:37:12 <smcginnis> hah
16:37:45 <smcginnis> #topic Preparing for train
16:38:05 <smcginnis> I did ping Tony yesterday about the script we use for setting up the tracking etherpad.
16:38:27 <smcginnis> He has that on his todo list to get set up. Checking this morning, looks like it didn't make it to the top of his todo's yet.
16:39:16 <smcginnis> When we moved to storyboard for tracking things, we stopped doing the planning etherpad. I did not do one for stein.
16:39:22 <smcginnis> I think that has worked out.
16:39:43 <dhellmann> do we have a ptg planning etherpad?
16:39:58 <dhellmann> or "agenda" or whatever?
16:40:04 <ttx> We don;t have that yet
16:40:12 <smcginnis> I know Tony has mentioned it, but I have not seen one.
16:40:14 <dhellmann> ok
16:40:15 <ttx> But we have a bunch of topics
16:40:22 <diablo_rojo> https://wiki.openstack.org/wiki/PTG/Train/Etherpads <-- according to that not yet
16:40:26 <dhellmann> I know we have some things at the bottom of the stein tracking page
16:40:28 <ttx> at the bottom of https://etherpad.openstack.org/p/stein-relmgt-tracking
16:41:13 <ttx> (personally I find it good that we don;t have that many things to discuss, means the process is now more solid
16:41:15 <smcginnis> I'll throw https://etherpad.openstack.org/p/relmgmt-train-ptg in that wiki list and we can start putting some thing there.
16:41:28 <dhellmann> ttx: +1 - less change is good here
16:41:52 <evrardjp> I would say let's wait for PTLs/release liaisons feedbacks?
16:41:58 <ttx> copy pasted
16:42:04 <smcginnis> Added to wiki
16:42:13 <evrardjp> team team team!
16:42:18 <diablo_rojo> technically cut pasted ;)
16:43:34 <smcginnis> #topic PTG org
16:43:45 <smcginnis> Notes from Tony in the etherpad.
16:43:54 <smcginnis> Friday morning as a meetup time.
16:43:57 <diablo_rojo> +2 for Friday morning
16:43:58 <smcginnis> I think that works for me.
16:44:15 <smcginnis> I've been mentally blocking actually thinking about my schedule for the week.
16:44:33 <diablo_rojo> smcginnis, lol yeah I bit the bullet and did it a couple days ago
16:44:42 <diablo_rojo> its quite painful
16:44:53 <smcginnis> ttx, dhellmann: Do you know if Friday morning will work for your schedules?
16:45:00 <dhellmann> I think so
16:45:20 <smcginnis> OK, if anyone has a conflict, I guess let Tony know ASAP.
16:45:28 <dhellmann> oh, the ptg stuff isn't on my calendar yet. does someone have the link to the schedule handy?
16:45:32 <smcginnis> The team photo is also scheduled for Friday morning.
16:45:36 <dhellmann> I expect to be spending most of my time with ironic
16:46:08 <diablo_rojo> dhellmann, https://www.openstack.org/ptg#tab_schedule
16:46:19 <dhellmann> thanks
16:46:25 <diablo_rojo> no problem :)
16:46:47 <ttx> OK reorganized topics in a way that makes sense
16:46:47 <dhellmann> yeah, I should be available friday morning
16:46:54 <smcginnis> Pretty much all teams are meeting Friday. I don't think it's possible not to have conflicts in the 3 days of this event.
16:47:28 <smcginnis> Last thing in there, Tony was asking about a team dinner.
16:47:51 <smcginnis> I would love it if we can all make it out to a dinner again.
16:48:07 <ttx> could make for a calm dinner for Friday evening
16:48:09 <smcginnis> I think the only set evening plan I have so far is Monday night.
16:48:14 <diablo_rojo> Thursday is the happy hour so we could go to dinner after that,
16:48:17 <smcginnis> Friday night should be good for me.
16:48:28 <diablo_rojo> ttx, Friday night is when the game night was gonna maybe be
16:48:33 <ttx> OOOOH
16:48:38 <ttx> then no
16:48:38 <dhellmann> yeah, small group calm friday or sat after such a long week would be good
16:48:38 <smcginnis> I leave Saturday evening, so that's the only PTG night that will not work.
16:48:45 <dhellmann> ah
16:48:49 <diablo_rojo> Saturday works for me
16:49:00 <diablo_rojo> Wednesday night?
16:49:07 <smcginnis> Thursday night might be a good way to prep for Friday morning.
16:49:49 <dhellmann> my dance card is completely empty at this point, so whenever
16:50:02 <smcginnis> How about this, each of us add a line under that item in the etherpad with the nights we are available? Then tonyb can figure something out.
16:50:20 <dhellmann> ++
16:50:59 <diablo_rojo> +2
16:51:22 <smcginnis> #topic Meeting time
16:51:46 <ttx> It changes !
16:51:57 <smcginnis> Last thing - I think it is official now. Meeting time will be moving to 1900.
16:52:15 <dhellmann> there was a review to change that, right?
16:52:21 <smcginnis> Probably about as good as we can get for all the timezones involved.
16:52:25 <ttx> yes merged now
16:52:26 <smcginnis> https://review.openstack.org/#/c/651443/
16:52:40 <ttx> some diligent meetings-core person
16:52:45 <smcginnis> :)
16:53:02 <ttx> btw...
16:53:15 <ttx> if someone else wants to be meetings-core, we recruit
16:53:29 <evrardjp> haha
16:53:33 <smcginnis> I would imagine that's pretty low volume.
16:53:49 <smcginnis> I can try to watch for things there if you want to add me to the list.
16:53:51 <ttx> It is so awesome you would not believe
16:53:55 <smcginnis> :D
16:53:57 <evrardjp> ttx: :)
16:54:12 <evrardjp> I am wondering what the duties are, with a name like this
16:54:16 <diablo_rojo> ttx, if you want another set of hands you can add me to the list too
16:54:18 <smcginnis> meetings-core must attend every meeting though. :)
16:54:41 <evrardjp> smcginnis: isn't that what you are already doing?
16:54:44 <smcginnis> You don't actually work on any projects, you just attend all of their meetings.
16:54:46 <ttx> well to be fair, I just process them all when they arrive. Could switch to some recview day set up though
16:54:47 <smcginnis> Hah!
16:55:07 <smcginnis> ttx: There's check validation that makes sure a propose meeting move doesn't conflict with another meeting ,right?
16:55:07 <dhellmann> ttx: bot approval?
16:55:10 <evrardjp> let's discuss this outside the meeting?
16:55:15 <smcginnis> Yeah
16:55:18 <ttx> dhellmann: we still check things.
16:55:19 <smcginnis> #topic Open floor
16:55:25 <smcginnis> Any final things?
16:55:29 <evrardjp> none
16:55:38 <smcginnis> This will be my last meeting. tonyb will take if from here.
16:55:41 <ttx> dhellmann: Like if the PTL is the person asking for it, if there was a thread about it, etc
16:55:47 <dhellmann> ack
16:56:00 <dhellmann> thank you, smcginnis , for taking the helm again this past cycle!
16:56:13 <dhellmann> things went quite smoothly, I think
16:56:15 <diablo_rojo> thanks smcginnis for all your hard work :)
16:56:29 <smcginnis> Thank you all for all the work this cycle and all the work prior to that that made it easy to step in and figure things out.
16:56:30 <ttx> I wonder what he will now do with all his free time. Apart from meetings-core
16:56:47 <smcginnis> haha
16:57:00 <ttx> and fearless Board member
16:57:05 <smcginnis> OK, I guess we can wrap things up then. Thank you everyone!
16:57:15 <smcginnis> #endmeeting