19:00:03 <bcotton> #startmeeting F32 Beta Readiness Meeting
19:00:03 <zodbot> Meeting started Thu Mar 12 19:00:03 2020 UTC.
19:00:03 <zodbot> This meeting is logged and archived in a public location.
19:00:03 <zodbot> The chair is bcotton. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:03 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:00:03 <zodbot> The meeting name has been set to 'f32_beta_readiness_meeting'
19:00:05 <bcotton> #meetingname f32-beta-readiness-meeting
19:00:05 <zodbot> The meeting name has been set to 'f32-beta-readiness-meeting'
19:00:15 <bcotton> #topic Roll Call
19:00:37 <pbokoc> I'm here for docs
19:00:39 * nirik is sort of around, sort of dealing with outage, ping me when you need me.
19:00:52 <bcotton> welcome, pbokoc, nirik
19:01:17 <nirik> good turnout... 1.5 people. ;)
19:01:41 <bcotton> that's about average :-)
19:01:58 <bcotton> part of the reason i changed up the process for this release
19:02:08 * nirik nods.
19:03:01 <bcotton> i'll wait until 5 minutes after and then we'll move along
19:05:25 <bcotton> #topic Current status
19:05:54 <bcotton> #info Fedora 32 Beta is NO-GO. We're going to try again with Beta 1.2 tomorrow and may still release on 17 March
19:06:04 <bcotton> #topic Team readiness
19:06:05 <bcotton> #link https://fedoraproject.org/wiki/Release_Readiness
19:06:33 <bcotton> pbokoc: are you aware that translation wants release notes 1 week before GA?
19:07:00 <pbokoc> bcotton, yeah, I saw that, I'll keep it in mind
19:07:18 <pbokoc> we'll have to update the task list, I'm assuming they'll want to keep doing that in future releases
19:07:44 <bcotton> pbokoc: great. should i start including a call for RN writers in the weekly commblog post and office hours? or are you not ready for that yet?
19:08:25 <pbokoc> bcotton, I'll be in about 20 minutes, I just need to set up a f32 branch
19:08:35 <bcotton> okay, sounds good :-)
19:08:43 <bcotton> by "task list" do you mean https://fedorapeople.org/groups/schedule/f-32/f-32-docs-tasks.html ?
19:09:49 <pbokoc> bcotton, yeah, that one. Line 5 is set to end at release minus one day
19:09:59 <pbokoc> and also line 9
19:10:39 <bcotton> okay. jibec, are you around?
19:11:16 <pbokoc> bcotton, oh wait, it's already set to 1 week before, never mind then
19:11:25 <pbokoc> all according to plan :)
19:11:30 <bcotton> excellent!
19:12:08 <bcotton> adamw or frantisekz, are you more clear on the expectations in regard to IoT now?
19:15:45 <frantisekz> bcotton: I have no idea about IoT, never wandered around it, kparal do you know more?
19:15:55 <frantisekz> (I need to run for hour or two)
19:16:15 <bcotton> frantisekz: okay, thanks. adam added it to the wiki, but i figured i'd ping you too as the listed contact
19:18:57 <bcotton> nirik: does your current rabbit fire impact release readiness?
19:19:21 <nirik> no, it shouldn't I don't think...
19:19:44 <nirik> we should be ok to mirror and distribute fine with the existing rc (if we decide to ship it)
19:23:32 <bcotton> ok. then i'll leave you to it. please let me know if it proves to be a risk to the beta or final releases :/
19:23:46 <nirik> can do.
19:24:00 <nirik> it's currently fixed, but trying to prevent it from happening again now.
19:24:09 <bcotton> fingers crossed!
19:27:17 <adamw> sorry
19:27:19 <adamw> uh
19:27:27 <adamw> no, we are not more clear on IoT.
19:27:35 <adamw> there is this idea in the ether that IoT is going to become release blocking for F32
19:27:52 <adamw> but when i look there is, like, zero paper trail for this
19:28:17 <adamw> there's no Change. there's no FESCo tickets. there's no draft PRD or tech spec or anything like that. there's just nothing to indicate this is a Thing at all. and since we're at Beta already, it seems quite late. :P
19:29:30 <bcotton> yeah. the problem here is that the Council voted to promote IoT to an edition, but since that's not a thing we do normally, there's not much of what you would call a Process for it :/
19:31:31 <bcotton> adamw: i want to do a longer debrief with you later so we can fix this for future promotions, but what can we do in the immediate term to provide some more clarity?
19:31:51 <adamw> bcotton: what i would recommend is going back to the Fedora.next process
19:31:57 <adamw> that's kinda the template for how we 'do' editions
19:32:18 <adamw> what did we require of all the editions there? that should be the starting point for figuring out what we require for onboarding new editions
19:33:16 <adamw> the bits that were significant to QA were, it needs to have a PRD (which is a high-level doc defining what the edition is ultimately *for*), and a tech spec (which is a more detailed, practical breakdown of the bits it'll include to achieve the goals in the PRD)
19:33:32 <adamw> then QA ultimately derived the release criteria for the editions from the PRDs and tech specs
19:33:42 <adamw> but i expect there were other requirements i've forgotten
19:34:10 <adamw> it also seems logical to me that promotion of a new edition in a given release should be a Change for that release
19:34:28 <adamw> generating all the other bumpf could be requirements for that Change
19:34:34 <bcotton> okay, i'll look at those. it's helpful to know that QA formed the criteria from the docs, because in a quick look i didn't see them explicitly called out
19:35:29 <bcotton> adamw: i agree-ish on the change part, just from a "who-approves-it" standpoint. but let me synthesize all of this and i'll try to form a more coherent opinion
19:38:28 <bcotton> adamw: is there anything that can be provided *now* apart from assurances that QA won't be blamed if it goes horribly wrong
19:39:21 <adamw> a PRD and a tech spec would be good
19:39:40 <adamw> but at this point i gotta say i really don't think we should be *doing* this for f32
19:39:58 <adamw> we have literally no 'fedora validation testing' for IoT at this point. it doesn't exist.
19:40:18 <adamw> in the blocker review meeting we just did now? none of that reflected any testing of IoT at all.
19:40:29 <adamw> that RC we built? doesn't have IoT images in it.
19:40:36 <bcotton> adamw: here's the PRD https://docs.fedoraproject.org/en-US/iot/prd/
19:40:52 <bcotton> understood
19:40:53 <adamw> what were we planning to ship as F32 Beta IoT? does anyone know? were we planning to ship something?
19:41:15 <adamw> that's a start
19:41:25 <adamw> tech spec would be nice, but we do have the release criteria that were just written directly
19:42:02 <adamw> the 'edition promotion' doc was written to an f31 timeframe and does not cover *anything* to do with qa
19:43:15 <adamw> the release criteria reference initial-setup, but afaik that was taken out of f32 and IoT is trying to port to ignition, only it doesn't really work yet (so i hear)
19:43:21 <adamw> you know, stuff like that. :)
19:43:26 <bcotton> :-)
19:44:45 <bcotton> aiui the plan was to call one of the nightlies 32 Beta IoT. it defintely doesn't help that it lives off in its own world largely (ditto FCOS and Silverblue, which will cause similar problems at some point)
19:45:47 <bcotton> honestly, i'm inclined to tell you to treat thie edition promotion largely as a marketing exercise for F32, which is a super crappy answer
19:46:20 <pwhalen> I can speak to some of that. We have been testing, there is a compose from the beta compose which I'll be testing once I finish up with vanilla testing.
19:47:11 <adamw> pwhalen: sure, the thing we don't have is a clear *validation testing* process which makes sense
19:47:41 <adamw> there's no joined-up process by which we say, okay, this is the thing we're thinking of releasing, here is what it should do, here are the tests that check that, here are the results for those tests
19:47:50 <adamw> it all just appears to be in various people's heads
19:47:59 <adamw> and tickets and stuff, but none of it is *done*
19:48:08 <adamw> (and the tickets were mostly filed like last week :>)
19:49:12 <adamw> bcotton: on fcos and silverblue - fcos is in a way less bad because it's *more* different: they not only have a separate compose process, it's nothing to do with pungi, and they have a whole test system baked in there
19:49:22 <adamw> silverblue is actually composed as part of regular composes
19:49:30 <bcotton> would it help to get folks from IoT and QA in a (virtual) room for a little bit to try to sort some of this out? that way we'll have at least *something* before GA
19:49:51 <adamw> bcotton: i don't know that it would, i mean, i talk to peter often enough
19:49:54 <adamw> we talked at devconf
19:49:58 <adamw> and the criteria came out of that
19:50:13 <adamw> but i don't want it to be just in his head and paul's head and matt's head and whoever else's head
19:50:59 <adamw> the thing i'm being awkward about is that there should be a clearer process for this whole thing and fesco/council need to be more proactive in figuring out requirements and, you know, filing tickets and writing stuff down
19:51:00 <bcotton> oh for sure. i can try to push on some more writing of things, and definitely work on fixing this for when some mythical future thing gets promoted
19:51:10 <adamw> rather than relying on people further down the chain to talk things out on irc and bodge it up
19:51:19 <bcotton> 100% on board with that
19:51:20 <adamw> that's just a recipe for confusion and messes and future tech debt
19:51:45 <adamw> as you say, we have IoT and FCOS and Silverblue all kinda in the pipeline and they're all Different in different ways
19:52:11 <adamw> we really aren't gonna get somewhere good by just kinda rolling along trying to figure out what we're doing ad hoc
19:53:12 <adamw> ahem, anyway. we should probably move on. but: no, qa is *not* yet clear with what's going on about IoT. :)
19:53:27 <bcotton> yep. i can focus on the long term fixes, and if there's more i can do in the immediate term you know where to find me
19:55:08 <bcotton> okay, that's all of the things that were on the wiki page
19:55:17 <bcotton> any other readiness-related discussion in the lst 5 minutes?
19:57:48 <bcotton> sounds like a no! thanks everyone
19:57:51 <bcotton> #endmeeting