15:00:17 <smcginnis> #startmeeting releaseteam
15:00:18 <openstack> Meeting started Fri Aug 24 15:00:17 2018 UTC and is due to finish in 60 minutes.  The chair is smcginnis. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:21 <openstack> The meeting name has been set to 'releaseteam'
15:00:22 <annabelleB> o/
15:00:28 <smcginnis> Morning annabelleB
15:00:31 <annabelleB> howdy!
15:00:58 <fungi> mmm, meetings
15:01:08 * elbragstad meanders in
15:01:40 <smcginnis> Doug is out, but I think I still saw ttx around.
15:01:50 <smcginnis> #link https://etherpad.openstack.org/p/rocky-relmgt-tracking Agenda
15:01:52 <smcginnis> Line 718
15:01:57 <ttx> o/
15:02:06 <smcginnis> #topic Rocky Release Status
15:02:34 <smcginnis> I didn't check the stale intermediary releases, but as far as unreleased deliverables go I think we mainly have the tempest plugins yet.
15:03:00 <smcginnis> There's been some activity for some of them, no signs of response on others.
15:03:08 <ttx> no we got everything that was too old refreshed i think
15:03:17 <smcginnis> Oh, excellent!
15:03:34 <smcginnis> annabelleB: Have you taken a look at the highlights so far? How do those look?
15:03:37 <ttx> depends on your def of too old I guess
15:03:57 <smcginnis> #link https://releases.openstack.org/rocky/highlights.html Rocky highlights
15:03:58 <annabelleB> smcginnis: looking good! I think folks are starting to get the hang of it and remembering to add them
15:04:24 <ttx> heat-agents got a release 2 days ago, same for swift
15:04:54 <smcginnis> ttx: OK, good. I think those were the two of concern last week.
15:05:14 <ttx> yes those were the two with no release in last 2 months
15:05:28 <smcginnis> annabelleB: Still some lower-level details for some of those highlights, but overall it did seem like most are starting to get the idea.
15:05:43 <smcginnis> Of course having previous examples to go off of really helps, so hopefully it just keeps getting better.
15:05:48 <annabelleB> smcginnis: agree, we’ll keep editing and driving home the idea, but improvement!
15:06:15 <ttx> shall we be concerned with the tempest-plugins ?
15:06:36 <smcginnis> Yeah, not sure what to do there.
15:06:58 <smcginnis> On the one hand, it's not as release critical, so I think it's OK if they are a little late.
15:07:06 <ttx> they are not really part of the release
15:07:08 <smcginnis> But we also don't want those coming in too late. Or not at all.
15:07:22 <ttx> as in no stable/rocky branch
15:07:30 <smcginnis> Yeah
15:07:47 <ttx> I would be tempted to force on Monday if not solved
15:07:54 <smcginnis> And we can always go back and tag at an earlier commit if we have to.
15:08:07 <smcginnis> Next week seems reasonable. Things should be quiet(ish) by then.
15:08:12 <ttx> maybe discuss that on Monday with dhellmann
15:08:19 <smcginnis> That works for me.
15:08:52 <ttx> documented on etherpad.
15:08:54 <smcginnis> Tempted to just go through the list and get the pypi registration and publish-to-pypi steps done since it's just kind of mechanical.
15:09:14 <ttx> yes, we probably should asist a bit in that technical step
15:09:46 <smcginnis> Once that's out of the way, the rest just becomes normal release process.
15:10:11 <ttx> Looking at the release request pipeline now
15:10:27 <ttx> nova rc3 should be in as soon as the last commit passes gate
15:11:06 <ttx> there is a magnum 7.0.1 that should pass the deadline too
15:11:16 <smcginnis> There are some tempest plugin patches that are waiting on the registration steps to get done.
15:11:24 <smcginnis> And a few release highlights that need updates.
15:11:34 <smcginnis> The only contentious one I see is https://review.openstack.org/#/c/590702/
15:11:46 <ttx> for monasca-events-api I'm tempted to approve it
15:11:55 <ttx> speaking of the devil
15:12:06 <smcginnis> Is it needed? Kind of disappointing there hasn't been a response there.
15:12:13 <ttx> since we have no idea how necessary it is
15:12:21 <ttx> better safe than sorry
15:12:29 <smcginnis> I suppose.
15:12:36 <ttx> maybe try to chase down witold / moasca people today
15:12:41 <smcginnis> I've removed my -1.
15:12:59 <ttx> I prefer to force it today than solve it next week as an exception
15:13:30 <smcginnis> ttx: OK, I guess I'm fine with just doing that one. Maybe with a comment that it was still an unaddressed concern.
15:13:46 <ttx> let me ask around in #openstack-monasca
15:14:13 <smcginnis> OK
15:14:24 <smcginnis> And other issues or concerns anyone is aware of for the rocky release?
15:15:20 <smcginnis> #topic PTG
15:15:29 <smcginnis> #link https://etherpad.openstack.org/p/stein-relmgt-tracking Stein etherpad started
15:15:41 <smcginnis> #link https://etherpad.openstack.org/p/relmgt-stein-ptg PTG Topic etherpad
15:15:59 <ttx> ok I did send a ping and ask for answer on #openstack-release or review
15:16:00 <smcginnis> Not sure if we will have many topics, but we always have things we can work through.
15:16:06 <smcginnis> ttx: Thanks!
15:16:17 <annabelleB> smcginnis: props for being so organized with these etherpads already!
15:16:24 <smcginnis> I think we can just find a time for us to grab a quiet corner during the week.
15:16:29 <smcginnis> annabelleB: Hah, thanks.
15:17:17 <smcginnis> If there is anything you want to make sure we talk about face to face, please add it to the etherpad.
15:17:38 <fungi> give me a heads up when you do and i'll be happy to join if i don't have pressing conflicts
15:18:03 <smcginnis> Oh, I did sign us up for a team photo. All are welcome and I hope you can make it there. First thing Tuesday morning.
15:18:42 <smcginnis> I can't find the spreadsheet to verify the time, but I'll track that down. Unless anyone else has that handy.
15:18:58 <elbragstad> #link https://docs.google.com/spreadsheets/d/1qSiBYGZBdAna00jYAi59MAFiMxVjaGC5l49j4vLdFQQ/edit#gid=0
15:18:59 <smcginnis> https://docs.google.com/spreadsheets/d/1qSiBYGZBdAna00jYAi59MAFiMxVjaGC5l49j4vLdFQQ/edit
15:19:05 <smcginnis> elbragstad: Thanks!
15:19:10 <elbragstad> np
15:19:18 <smcginnis> So 9:30 Tuesday morning.
15:19:48 <armstrong> Denver will be the official release of Pike right?
15:19:53 <fungi> 15:30 utc?
15:20:17 <smcginnis> fungi: I'll trust you on that one. ;)
15:20:24 <annabelleB> fungi: i believe so!
15:20:28 <smcginnis> armstrong: Rocky you mean? That's on the 30th.
15:20:45 <armstrong> Oh yes, sorry about that
15:20:47 * fungi is starting to add ptg reminders to his schedule
15:21:44 <smcginnis> armstrong: We will cut the final release on the 29th to give a day buffer before the release announcement on the 30th.
15:21:53 <ttx> .me checks Tuesday morning for conflicts
15:22:50 <ttx> smcginnis: hmm, I may have to be somewhere else... could we move to Wednesday ?
15:23:53 <smcginnis> Looks like Wednesday morning is also open. Does that work for everyone?
15:24:03 * smcginnis was hoping to be back to back with requirements
15:24:37 <armstrong> @smcginnis: ok I see
15:24:45 <ttx> hmm maybe prometheanfire can move along
15:24:52 <smcginnis> Yeah, I'll check with him.
15:25:14 <smcginnis> OK, I will move it to Wednesday.
15:27:00 <smcginnis> #topic Open discussion
15:27:05 <smcginnis> Anything else we should discuss?
15:27:19 <ttx> release hour maybe
15:27:37 <ttx> annabelleB: do you know if we are doing the classic 10am CF PR
15:27:39 <ttx> CT
15:27:50 <annabelleB> AFAIK yes
15:28:01 <ttx> I haven't heard otherwise either
15:28:07 <annabelleB> might drop a few minutes before as we have the community meeting starting at 10am CT
15:28:41 <ttx> so the PR goes out 10am CT, ideally we have the releases.o.o page all set before then
15:28:49 <ttx> since it points to it
15:28:51 <annabelleB> correct
15:29:19 <ttx> Also we usually time the "it's released" tweet and ML announce post to a few minutes before that
15:29:38 <ttx> since "it's released" is a bit of a fuzzy quantity
15:30:31 <smcginnis> I can get an equivalent of https://review.openstack.org/#/c/548563/ queued up and ready to go.
15:30:32 <ttx> so let's plan to spend the few hours leading up to that online
15:30:47 <smcginnis> That's the only thing for releases.o.o, right?
15:31:14 <ttx> smcginnis: we need it to go through all the RC-> final updates too
15:31:24 <ttx> (before applying that patch)
15:31:50 <smcginnis> That will be queued up Monday. You mean the patch we want to get all the PTL +1's right?
15:32:00 <ttx> yes
15:32:22 <ttx> as many as we can would be more correct
15:32:37 <smcginnis> OK, so let's plan to be in the -release room from 1400 to 1500 on the 30th. We can approve the final patch and then get through the releases.o.o update marking it released.
15:32:43 <smcginnis> ttx: Yep
15:32:57 <smcginnis> Or wait, 29th or 30th?
15:33:10 <smcginnis> Normally we've done the final release the day before the announce I thought.
15:33:31 <ttx> wedneday, so 29
15:33:54 <smcginnis> So 1400-1500 on the 29th?
15:33:56 * ttx looks up process
15:34:41 <ttx> I think all the steps in "Final release" are for release day
15:35:23 <ttx> we can start approving things early in the day though
15:35:53 <smcginnis> OK, I think I stated the 29th as the release deadline as far as dev goes, but we can process that the morning of the 30th.
15:35:54 <ttx> I would approve the patch when you get up in the morning
15:35:59 <fungi> is that 1400-1500 utc or mdt?
15:36:04 <smcginnis> UTC
15:36:07 <fungi> thanks
15:36:20 <smcginnis> Trying to be global-friendly. :)
15:36:20 <ttx> the 30th is the Thursday, so no
15:36:32 <smcginnis> OK, I'm just getting confused then.
15:36:36 <fungi> oh, right, not at the ptg
15:36:45 * fungi has wires crossed in head
15:37:11 <smcginnis> We do our final release and our announcement to openstack-announce on the 29th, official release announcement for marketing level is on the 30th?
15:37:16 <ttx> no
15:37:27 <ttx> everything the 29th
15:37:53 <annabelleB> marketing is on the 30th?
15:37:55 <smcginnis> Hmm, we had previously announced the 30th was the official release date. Does that work for you annabelleB if we actually do it all on the 29th?
15:37:57 <fungi> that was at the request of distro package maintainers who wanted the extra day before the weekend to do final packaging work, yeah?
15:38:14 <fungi> (wednesday release instead of thursday)
15:38:20 <ttx> uh
15:38:36 <smcginnis> I thought in the past we did the release Wednesday but the announce on Thursday.
15:38:46 <ttx> that is weird
15:38:59 <ttx> not sure where that Thursday comes from
15:39:07 <annabelleB> i _think_ we can make that work—thursday is preferrable—I’d be going off the date on the site for the press team, but if we had to do a Wednesday marketing announcement, I think we could swing it
15:39:28 <ttx> well if we announced Thursday on the schedule...
15:39:39 <annabelleB> Did we want the day inbetween releasing and marketing announcing in case of…anything?
15:39:47 <ttx> we should stick to that if that works for everyone
15:39:57 <ttx> In my head it was the 29th
15:40:08 <ttx> but I read it now and it says 30th
15:40:10 <smcginnis> I just checked openstack-announce for pike and queens, and we did send that out on Wednesday.
15:40:36 <ttx> smcginnis: sure, but we also announced that it would happen on the Wednesday :_
15:40:37 <smcginnis> Maybe I messed up the release date.
15:41:02 <smcginnis> Queens schedule says the 28th (Wednesday) and we announced on the 28th.
15:41:08 <ttx> right
15:41:22 <smcginnis> Any drawback if we do the 29th even though it had said the 30th?
15:41:28 <ttx> I'm more attached to releasing on the announced day, than to releasing on the Wednesday
15:41:47 <ttx> not a big deal though from our end
15:41:53 <annabelleB> I’d prefer the extra day for the press release work
15:42:05 <ttx> we'll be ready on Wednesday, and releasing "early" is not a bad thing.
15:42:08 <smcginnis> OK, we can see if there is any pushback from distros not having that extra day and we can just apologize this time around and fix it next time.
15:42:24 <ttx> But if it makes marketing more difficult, I'd keep teh announced date
15:42:54 <smcginnis> I also put the stein release day as Thursday, so I can put up a patch to fix that.
15:43:00 <fungi> well, the "extra day" was mostly that they wanted us to start releasing on wednesdays to give them a full 2 days before the weekend to do final packaging (is what i was trying to say earlier)
15:43:11 <smcginnis> And add a note to the process doc to make sure we don't have confusion next time.
15:43:30 <fungi> before a few cycles ago we were releasing on thursdays and that left them with fairly tight timing
15:43:35 <smcginnis> fungi: Yeah, maybe things have slowed down enough now that it woudl be OK?
15:44:07 <fungi> sure, maybe. i don't maintain (openstack anyway) distro packages so hard for me to gues
15:44:09 <fungi> s
15:44:26 <smcginnis> Yeah. Not a lot of direct feedback there, unfortunately.
15:44:46 <ttx> smcginnis: my suggestion: i would stick to announced dates
15:44:54 <smcginnis> OK, let's do THURSDAY morning 1400-1500 UTC this time around. I will fix things for stein so we do Wednesday next time.
15:44:57 <evrardjp> smcginnis: I am asking on side channels
15:44:58 <ttx> It's the only way to not piss someone -- they had plenty of time to organize/complain
15:45:12 <smcginnis> Yep, it has been the published date for a long time.
15:45:19 <fungi> wfm
15:45:21 <evrardjp> that's fair
15:45:24 <evrardjp> wfm too
15:45:29 <annabelleB> thursday it is!
15:45:39 <smcginnis> My countdown email said the 29th, but I think it's fine giving the community the shorter deadline to make sure things are ready to go on the actual date.
15:45:39 <ttx> now I still think Wed is better, so we should fix Stein while we can :)
15:45:53 <smcginnis> Yep, I'll do that today.
15:46:09 <evrardjp> smcginnis: that's project management speaking right there
15:46:09 <ttx> crisis averted!
15:46:14 <smcginnis> OK, anything else I can confuse today?
15:46:15 <smcginnis> evrardjp: :)
15:46:41 <ttx> nothing comes to mind, but that was a good one to catch :)
15:46:51 <smcginnis> Yes, glad that was brought up!
15:46:56 <annabelleB> moi aussi!
15:47:02 * ttx books the Thursday in his calendar instead of Wednesday
15:47:33 <evrardjp> annabelleB: :)
15:47:41 <smcginnis> OK, thanks everyone. It's been a pretty good release cycle this time around. I appreciate everyones involvement.
15:47:51 <annabelleB> smcginnis: thanks for being our fearless leader!
15:47:58 <evrardjp> thanks smcginnis !
15:48:14 <ttx> smcginnis: great work on your side !
15:48:19 <fungi> thanks to everyone for doing the thankless things!
15:48:23 <smcginnis> #endmeeting