16:00:10 <smcginnis> #startmeeting releaseteam
16:00:11 <openstack> Meeting started Thu Feb 14 16:00:10 2019 UTC and is due to finish in 60 minutes.  The chair is smcginnis. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:12 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:14 <openstack> The meeting name has been set to 'releaseteam'
16:00:18 <smcginnis> Ping list: smcginnis ttx dhellmann diablo_rojo hberaud evrardjp fungi armstrong
16:00:31 <hberaud> o/
16:00:34 <evrardjp> o/
16:00:40 <smcginnis> #link https://etherpad.openstack.org/p/stein-relmgt-tracking Agenda
16:00:42 <dhellmann> smcginnis : I'm in a call but lurking.
16:00:44 <fungi> ohai
16:00:48 <smcginnis> dhellmann: ack
16:00:48 <ttx> o/
16:01:22 <smcginnis> #topic Do we want to do the release 'manifest'
16:01:34 <smcginnis> #link http://lists.openstack.org/pipermail/openstack-dev/2018-September/135088.html
16:01:48 <ttx> right so originally we had that idea to reaffirm liaisons toward the end of the cycle
16:02:07 <ttx> but we haven't really been actively implementing that
16:02:25 <ttx> I guess there is still time, but we need to start socializing that now-ish
16:02:32 <ttx> thoughts?
16:02:50 <ttx> (I kept it off the process change since I was not sure we'd do it for this cycle)
16:03:11 <smcginnis> With the loss of enforcing 2 milestones, we do need some way to validate what will be in the coordinated release.
16:03:36 <smcginnis> Did we ever figured out the specifics of how we would do that "manifest"?
16:03:58 <ttx> I was thinking about a list of names and email addresses, keyed by project team
16:04:12 <ttx> in deliverables/stein
16:04:23 <smcginnis> manifest.yaml or similar?
16:04:33 * diablo_rojo waves
16:04:34 <ttx> yaml would likely break automation
16:04:50 <smcginnis> Yeah, just thinking we probably have too many places that glob on *.yaml
16:04:51 <ttx> since we are expected to only have release files ni deliverables/*/*.yaml
16:05:29 <evrardjp> sorry I am not really sure to understand the point of the manifest
16:05:31 <smcginnis> So a manifest.txt where each team would need to add their acknowledgement that they plan to be part of the stein release?
16:05:32 <ttx> but could just be manifest.txt or liaisons.txt
16:05:45 <smcginnis> evrardjp: It's due to the changes with cycle-with-mielstones.
16:05:53 <evrardjp> I see the email linked there
16:06:04 <smcginnis> In the past, we had a requirement that projects needed to do at least two milestone releases.
16:06:14 <evrardjp> I don't see the direct link between having a manifest file and activity/healthiness
16:06:20 <smcginnis> That was used as an indicator to make sure they were active and were planning on being part of the coordinated release.
16:06:25 <evrardjp> couldn't reviews be enough?
16:06:28 <ttx> evrardjp: we'd ask those people to add themselves
16:06:33 <ttx> so it's a commitment
16:06:45 <ttx> in advance of actually requiring reviews
16:06:54 <smcginnis> But not that we don't require milestones, the idea to make sure we had would be for them to explicitly acknowledge it by adding to a manifest.
16:06:55 <ttx> (at which point it's too late)
16:07:09 <evrardjp> ttx: what if projects don't write their names?
16:07:24 <evrardjp> (sorry to play devil's advocate here)
16:08:11 <evrardjp> for me releases should be helping projects, not putting burden on them with <insert new procedure> . I am just worried about the attention loss on that eventual file
16:08:20 <evrardjp> but maybe I am a dreamer
16:08:57 <smcginnis> I am a little concerned about having a new "you all must do this or else you are not included" requirement.
16:08:59 <evrardjp> but the plan is the plan right now, sorry for that interruption -- please continue
16:09:21 <evrardjp> smcginnis: I am glad I am not the only one :)
16:09:28 <ttx> evrardjp: the manifest process, done early enough, gives us a chance to chase people down
16:09:31 <smcginnis> It was a plan we talked about in Denver. I think it's a question now of whether it really is something we need and want to do.
16:09:36 <ttx> It's extra painful to do that in the release weeks
16:10:02 <evrardjp> can't we assume that if people have done an RC and branching, they are following our processes?
16:10:03 <smcginnis> I guess the question (or a question) is if it is "early enough" now or if we've missed that window.
16:10:06 <evrardjp> (that's the dreamer part)
16:10:58 <ttx> yeah, we could also skip it this time and see how it goes
16:11:18 <ttx> evrardjp: so most of the branching and RC is now automatic
16:11:35 <evrardjp> yeah I meant a review on it
16:11:40 <ttx> The big question is... is there anyone actually making sure that what we automatically ship is actually working
16:11:58 <ttx> the manifest was a way to close that loophole
16:12:03 <evrardjp> oh
16:12:08 <smcginnis> I'm wondering if that might be better this first time around. I feel like we've been introducing a lot of new things for teams to have to be aware of that are really distractions from their main focus. I'd be concerned introducing another thing they have to keep track of doing might be pushing things.
16:12:13 <ttx> if nobody abswers for a full month, let's not ship it at all
16:12:45 <smcginnis> Honestly, I'm wondering now if it was really worth switching from the milestone model.
16:13:24 <smcginnis> I would be very interested to hear feedback from teams on how they felt this has worked for them this cycle vs past ones.
16:13:40 <evrardjp> that sounds like a good PTG topic
16:13:48 <fungi> absoluetly
16:13:48 <ttx> I there was certainly less boilerplate activity as a result
16:13:50 <smcginnis> evrardjp: Good point.
16:14:14 <ttx> but maybe that removed a lot of responsibility as well
16:14:31 <evrardjp> ttx: less work for teams is good
16:14:32 <ttx> a good way to assess that is to run the stein process without the manifest
16:14:39 <evrardjp> ttx: responsibilities don't have to be in code
16:14:43 <smcginnis> Less boilerplate, but my project perspective is that milestone activity kind of helped to add a little sense of urgency to getting some things done in time where now it's kind of like "oh yeah, milestone-x was last week'/
16:14:45 <ttx> and see how screwed we end up
16:14:53 <evrardjp> ttx: :)
16:15:24 <evrardjp> smcginnis: I feel that too on my side of the pond
16:16:04 <ttx> ok, so... no manifest, revisit in Denver
16:16:08 <ttx> ?
16:16:13 <smcginnis> Proposal then: no manifest this time around and we see how it goes. We try to do some sort of retrospective at the PTG to evaluate how things went on our end AND try to get PTLs and teams involved to give feedback on how it worked for them.
16:16:30 <evrardjp> ttx: smcginnis +1
16:17:08 <ttx> It's tricky because with less resources we want to avoid useless work... but at the same time with less activity people are more likely to be lulled into sleep in absence of forced milestones
16:17:12 <diablo_rojo> Works for me
16:17:46 <ttx> I still think it was the right move though. We just need to make extra release noise
16:17:48 <evrardjp> ttx: you're right, and I think that should be into the elements of the PTG conversation
16:18:01 <ttx> to wake people up
16:18:01 <smcginnis> I think since it is mostly boilerplate, the amount of work is low versus the level of engagement it promotes.
16:18:03 <evrardjp> ttx: I agree on the noise and communication
16:18:16 <ttx> starting next week(s) with library stuff
16:18:23 <smcginnis> But happy either way if we can get feedback that we are helping teams and not causing friction.
16:18:55 <ttx> manifest is adding some bureaucratic procedure to make sure we actually have people caring. But we could also do that by making a lot of release noise
16:19:36 <ttx> (our classic dilemma of getting acknowledgement for our messaging)
16:19:54 <smcginnis> I wonder if we could do the two milestone thing by having them add to the manifest at T-1 and some sort of confirmation at T-2.
16:20:04 <ttx> aka no way to put a big dialog box in front of all teams and know for sure
16:20:14 <evrardjp> I think using noise would be good -- as we can use that to say "don't forget highlights " anyway
16:20:20 <smcginnis> Less work than doing a milestone release but still something to trigger them to be aware of the release progress.
16:20:36 <ttx> smcginnis: could be a way to do it
16:20:55 <evrardjp> who said spam on the channels and ML? :p
16:21:12 <smcginnis> Well, too late anyway at this point for something like that, so I think I like the plan to skip a manifest for now and reevaluate in Denver.
16:21:36 <ttx> so with stein we've pushed the pendulum far in the "less of useless work" direction. For train we might need to get a bit more insurance
16:21:57 <ttx> ok next topic
16:22:22 <smcginnis> ++
16:22:29 <smcginnis> #topic Cycle-with-rc transitions in absence of PTL answer
16:22:40 <ttx> speaking of...
16:22:42 <smcginnis> Looks like we have Watcher and Tacker still out there.
16:22:56 <ttx> Tacker had some response
16:23:24 <ttx> from not-the-PTL
16:23:30 <ttx> maybe he's the manifest person
16:24:09 <ttx> he is a tacker core
16:24:18 <ttx> I'm fine with that as confirmation
16:24:20 <evrardjp> he was manifesting his presence as liaison?
16:24:37 <ttx> too soon?
16:24:42 <evrardjp> too soon
16:24:47 <evrardjp> :)
16:25:05 <evrardjp> sorry, please continue.
16:25:28 <ttx> So I'm fine with the confirmation we got on the TAcker one. More concerned with Watcher
16:25:50 <smcginnis> I think we've published this as the plan. We've had the review out there for a week(ish) for feedback. I'm OK approving it without an official ack.
16:25:52 <ttx> are those the only two left ? Not so bad I guess
16:26:15 <ttx> oh we have a Watcher ack
16:26:28 <ttx> 6 hours ago
16:26:30 <smcginnis> Problem solved.
16:26:32 <ttx> all is good
16:26:42 <ttx> push the W+1
16:26:45 <smcginnis> Nothing to see here, move along.
16:26:52 <ttx> oh I guess I will
16:27:04 <smcginnis> Thanks :)
16:27:16 <smcginnis> #topic puppet-aodh failure
16:27:27 <smcginnis> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-February/002694.html
16:27:33 <smcginnis> Yeah, weird.
16:27:46 <smcginnis> It's gotta be some kind of test of a new release job or something.
16:27:50 <smcginnis> fungi: Any insight on this one?
16:27:53 <ttx> fungi: was there some infra re-enqueueing?
16:28:08 <ttx> dhellmann: have you been sleepwalking again?
16:28:25 <fungi> we're continuing to work through problems with the publication to puppetforge
16:28:37 <fungi> so that same release has been retriggered several times over the past month
16:28:46 <ttx> ok, so safe to ignore?
16:29:14 <fungi> yeah, currently it seems to be failing on credentials, and i'm hoping to get time today to manually decrypt and compare them to make sure it's not an encoding issue
16:29:20 <smcginnis> Problem solved again. :)
16:29:28 <ttx> fungi: feel free to quick-respond to the thread on openstack-discuss to close it for reference
16:29:28 <fungi> if only
16:29:35 <fungi> will do
16:29:40 <ttx> thanks!
16:29:46 <fungi> once i climb out of today's meeting hole
16:29:47 <ttx> we are killing it today
16:30:05 <smcginnis> fungi: Well, our problem solved. Not yours. :D
16:30:10 <fungi> heh ;)
16:30:19 <ttx> smcginnis: maybe we should assign next week tasks ?
16:30:29 <smcginnis> #topic Assigning tasks
16:30:32 <ttx> (coincidentally i won't be around)
16:30:56 <smcginnis> ttx: Are you snowboarding again?!
16:31:06 <ttx> no, visiting NYC with family
16:31:16 <ttx> so, yes.
16:31:16 <smcginnis> Oh, fun.
16:31:40 <smcginnis> Freeze reminder will be in my countdown email.
16:31:49 <fungi> have fun. i enjoy manhattan immensely
16:32:09 <smcginnis> I was surprised by how much I liked NYC. Actually didn't think I would.
16:32:13 <diablo_rojo> fungi, actually? Or is that sarcasm? lol
16:32:21 <evrardjp> diablo_rojo: I was wondering the same :D
16:32:42 <smcginnis> stable-maint groups is another one for the countdown email.
16:33:21 <smcginnis> fungi: Do you want to take the task to notify infra to generate an artifact signing key? Or should someone else do that?
16:34:25 <smcginnis> Not sure if that last one is an issue yet with our changes.
16:34:26 <fungi> diablo_rojo: i genuinely enjoy manhattan
16:34:29 <smcginnis> I suppose there are a few.
16:34:32 <fungi> it's a fun town
16:34:43 <fungi> smcginnis: i'll take it, yes
16:34:52 <smcginnis> fungi: Thanks
16:34:56 <fungi> (and probably just do it unless i get an eager volunteer)
16:35:12 <smcginnis> ttx or dhellmann: Do either of you want to take checking on c-w-i projects that have not been released?
16:35:20 <smcginnis> #delegation
16:36:03 <smcginnis> Everyone's gone planning vacations now. :)
16:36:52 <ttx> let me see
16:37:21 <smcginnis> Doesn't have to be right now, but Monday I guess.
16:37:34 <ttx> smcginnis: I feel like that would be a good exercise for our new recruits :)
16:37:58 <ttx> an area they have not been touching on that much
16:38:06 <ttx> #delegation
16:38:08 <evrardjp> fair
16:38:09 <diablo_rojo> What needs to be done?
16:38:21 <ttx> diablo_rojo: it's described in PROCESS I think
16:38:24 <diablo_rojo> Just go poke them and get them to do one if they haven't already?
16:38:30 <diablo_rojo> Oh, convenient lol
16:38:45 <evrardjp> ttx: is it clearly defined in process page?
16:38:52 <ttx> diablo_rojo: I think the idea is just to establish the list, for smcginnis to include in his weekly email
16:39:03 <evrardjp> I guess diablo_rojo and I we can have a look at come back to you?
16:39:06 <ttx> evrardjp: touché
16:39:34 <smcginnis> Might be a good exercise to go through to find the script to run to get the list.
16:39:48 <diablo_rojo> Making a list is definitely doable :)
16:39:51 <diablo_rojo> I love lists.
16:39:58 <smcginnis> evrardjp, diablo_rojo: Should I put your names on the task in the etherpad?
16:40:06 <diablo_rojo> smcginnis, sure!
16:40:09 <openstackgerrit> Merged openstack/releases master: Switch Watcher service to cycle-with-rc  https://review.openstack.org/635662
16:40:17 * diablo_rojo signs up evrardjp to do work with her :)
16:40:40 <smcginnis> evrardjp: OK with you?
16:40:44 <evrardjp> ofc
16:40:54 <evrardjp> after today though, it's a special day....
16:40:55 <smcginnis> :)
16:41:09 <evrardjp> (birthday!)
16:41:15 <smcginnis> It's technically a task for next week, so plenty of time?
16:41:15 <evrardjp> (not mine)
16:41:19 <smcginnis> Mine
16:41:24 <evrardjp> yeah everything good
16:41:29 <diablo_rojo> evrardjp, Cecile?
16:42:07 <smcginnis> #topic Open discussion
16:42:08 <evrardjp> wow that's not for meetings!
16:42:26 <evrardjp> let's discuss tomorrow diablo_rojo about your love of lists
16:42:36 <evrardjp> and make that thing
16:42:47 <smcginnis> Just wanted to point out to folks that I put up a proposed schedule for train
16:42:49 <diablo_rojo> evrardjp, I'll bea headed to MN tomorrow, but in the morning before I head to the airport we can :)
16:42:50 <smcginnis> https://review.openstack.org/#/c/636742/
16:43:04 <evrardjp> diablo_rojo: let's sync in a different chan :)
16:43:08 <smcginnis> Update coming, but take a look and see if milestones land on any important dates or anything liek that.
16:43:34 <diablo_rojo> evrardjp, sounds good :)
16:45:15 <smcginnis> OK, anybody have anything else?
16:45:42 <diablo_rojo> Don't think so.
16:45:53 <smcginnis> Thanks everyone.
16:45:58 <smcginnis> #endmeeting