17:00:38 #startmeeting fedora_atomic_wg 17:00:38 Meeting started Wed Jun 7 17:00:38 2017 UTC. The chair is dustymabe. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:38 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:38 The meeting name has been set to 'fedora_atomic_wg' 17:01:04 #topic roll call 17:01:07 .hello dustymabe 17:01:08 dustymabe: dustymabe 'Dusty Mabe' 17:01:11 .hello yzhang 17:01:12 yzhang: yzhang 'Yu Qi Zhang' 17:01:13 .hello miabbott 17:01:15 miabbott: miabbott 'Micah Abbott' 17:01:17 .hello jasonbrooks 17:01:18 jbrooks: jasonbrooks 'Jason Brooks' 17:01:27 .hello roshi 17:01:28 roshi: roshi 'Mike Ruckman' 17:01:29 .hello jlebon 17:01:32 jlebon: jlebon 'None' 17:01:32 .hello mmicene 17:01:32 .hello jberkus 17:01:34 nzwulfin: mmicene 'Matt Micene' 17:01:38 jberkus: jberkus 'Josh Berkus' 17:01:38 .hello sinnykumari 17:01:40 ksinny: sinnykumari 'Sinny Kumari' 17:02:18 hiya ksinny :) i was looking for jwboyer earlier to tell him about the meeting :) 17:02:32 dustymabe: Hi :) 17:02:49 dustymabe: It's a bit late in India, so mostly not able to join meeting 17:03:11 but I will try to be part of it 17:03:50 ksinny: yeah - we have talked about re-proposing a meeting time 17:04:18 dustymabe: that's awesome 17:04:23 but that is a hard problem to solve because everyone has their day packed already :( 17:04:36 we are glad to have you here ksinny 17:04:43 sometimes kushal and sayan join us too 17:04:56 ok - let's do previous meeting action items 17:05:03 dustymabe: indeed, no worry if we can't change the meeting time 17:05:05 #topic previous meeting action items 17:05:18 * jberkus to ask mattdm about how to get PRD approved 17:05:20 * maxamillion to update container guidelines to enforce ENV and label 17:05:22 maintainer usage https://pagure.io/atomic-wg/issue/277 17:05:24 * dustymabe to propose a VFAD for wiki and docs discussion 17:05:26 * dusty and roshi to plan test day for coming weeks 17:05:28 * everyone to discuss flock proposals next week 17:05:41 * roshi still has to put the testday on the calendar and whatnot 17:06:24 #info dusty and roshi are planning test day in https://pagure.io/atomic-wg/issue/280 17:06:38 going to re-action the VFAD thing 17:06:45 #action dustymabe to propose a VFAD for wiki and docs discussion 17:07:15 #info ticket to track flock discussions/proposals https://pagure.io/atomic-wg/issue/279 17:07:37 jberkus: want to update us with info about PRD? 17:07:59 yah, right now it's marinating for a week while Council members +1 it 17:08:05 so far, some +1s, no arguments 17:08:12 want to link to council ticket? 17:08:25 the one request was that we add some clear instructions on how to *join* the WG, which apparently we don't have 17:08:46 https://pagure.io/Fedora-Council/tickets/issue/119 17:09:07 #info our PRD is being voted on in https://pagure.io/Fedora-Council/tickets/issue/119 17:09:34 jberkus: do you want to add some info on that? we don't really have anything too formal 17:09:45 * yzhang volunteers to test out whatever method we come up with 17:09:50 maybe things will get more formal if we decide we need more rules 17:09:57 dustymabe: we can just paste in the bit you wrote up in email 17:10:15 jberkus: sounds good to me - although josh did poke holes in one of my points 17:10:18 so maybe we remove that one 17:10:31 jberkus: but i'm +1 if you want to do that 17:10:46 ok 17:11:07 where is our general wiki page? 17:11:16 https://fedoraproject.org/wiki/Atomic needs a TOC 17:11:32 jberkus: :) hence why we need an index off of the pagure home page 17:11:36 is there some way to see all pages in a namespace? 17:11:49 docs VFAD should help with that 17:12:32 ok we covered most action items 17:12:45 maxamillion seems to be MIA - i think he told me he would be out 17:12:57 #action maxamillion to update container guidelines to enforce ENV and label maintainer usage https://pagure.io/atomic-wg/issue/277 17:13:31 moving on to meeting items 17:13:32 dammit, there was a major issue to bring up today, but I've a migraine and can't think of it 17:13:41 language :) 17:14:02 #topic clarify policy on atomic host support for older Fedora "number" releases 17:14:06 #link https://pagure.io/atomic-wg/issue/228 17:14:20 jbrooks has published blog post on pa.io 17:14:44 #info jbrooks has published blog post on pa.io http://www.projectatomic.io/blog/2017/06/future-plans-for-fedora-atomic-release-life-cycle/ 17:14:50 so a copy of that needs to live somewhere referencable. wiki? 17:14:55 pagure? 17:15:12 wiki 17:15:26 copy of what? doesn't PRD have a section on that 17:15:30 that talks about updates 17:15:38 yeah 17:15:46 yah, but it's brief 17:15:49 but it's a clear departure from every other part of fedora 17:15:59 so it'd be good to have it in a couple places 17:16:02 we want a copy of the blog post on a separate page (linked) for the detailed explanation 17:16:03 I think, anyways 17:16:15 sure - sounds fine 17:16:29 i believe jbrooks was also going to post on the fedora magazine - i asked mattdm about this 17:16:41 basically we want to broadcast to the fedora community as much as possible 17:16:42 roshi, supporting this is part of the goal of containerizing things - you don't have to update everything at once, we just have a faster moving host 17:16:47 dustymabe, Yeah, I'm dumping the text into a post right now 17:16:54 walters: sure, makes sense to me 17:16:58 the other fedora editions don't currently emphasize this 17:17:07 but unless you're in the fold, you won't just *know* that 17:17:54 ok - I think we are mostly done with this topic? any open questions? 17:18:32 dustymabe: never been part of VFAD but would love to. To get some idea would like to know that how long VFAD runs and usually what's the timing? 17:18:53 ksinny: cool - let's chat about it in open floor :) 17:19:01 yep 17:19:06 moving on to next meeting item 17:19:20 #topic fedimg: don't use 'builder' instance for uploading AMIs 17:19:34 #link https://pagure.io/atomic-wg/issue/269 17:19:54 looks like sayan is getting close to closing off the loose ends 17:20:03 he did a new release of fedimg earlier today 17:20:24 #info sayan is close to getting #269 resolved still waiting on one more thing 17:20:39 moving on to next topic 17:20:50 #topic Fedora 26 Atomic/Cloud Test Day 17:20:58 #link https://pagure.io/atomic-wg/issue/280 17:21:17 so for this, I basically need to update the QA calendar, and create a wiki page with what we want tested 17:21:25 basically we want to do this soon, but roshi can't do it tue-thur next week and I can't do it friday 17:21:34 along with some instructions on what we're trying to ensure works 17:21:38 I'm on PTO next week 17:21:41 :D 17:21:46 so next possible day was monday - and then QA has meetings all day monday 17:21:50 :( 17:21:58 soo.. that lands us on tuesday the 20th 17:21:58 pretty much :) 17:22:10 which gives us 7 days before Final Freeze 17:22:14 to find/fix bugs 17:22:31 dustymabe: I'm at a conference on the 20th, although a low-key one. Just means I won't be able to test all day 17:22:33 well, if there are blockers, those can get pulled in during freeze 17:22:46 freeze exceptions as well 17:23:15 roshi: yeah, that's a bit of a can of worms since AH is not release blocking 17:23:17 so I see that as less of an issue (being close to final freeze) 17:23:29 so blocker for us is considered differently 17:23:45 FEs for AH don't have much of an impediment to being accepted 17:24:05 roshi: yes, as long as the component we are updating only is included in atomic host 17:24:15 yeah 17:24:20 but if it is say - anaconda, then good luck 17:24:25 :) 17:24:37 * roshi doesn't recall a bunch of bugs that we couldn't get fixed because of this issue though 17:24:37 so, all of this to say: i think we'll have an official test day on the 20th 17:24:47 but i'm going to send out information way before that with links and such 17:24:55 so we can be testing before the 20th 17:25:24 for sure 17:25:25 and hopefully find anything that is crazy alarming 17:25:48 also, side note - i updated a cluster running openshift from f25 to f26 and rolled back 17:25:59 yah? 17:26:00 and openshift continued to work through each transaction :) 17:26:03 remind me at Open Floor that I have an aside from this discussion 17:26:06 walters: jlebon ^^ 17:26:12 you'll be proud 17:26:37 roshi: cool 17:26:37 :) 17:26:40 doing a rebase while the system is running is definitely one of the cooler things 17:26:50 ok moving to next item 17:27:01 #topic talk/workshop proposals for Flock 17:27:06 #link https://pagure.io/atomic-wg/issue/279 17:27:16 the CFP for flock is going to be extended 17:27:51 oh? flock isn't that far away 17:28:06 jberkus: yeah, but they didn't have a lot of proposals yet 17:28:26 i honestly think this move to "less talks" has got people not submitting talks 17:28:43 which makes sense, personally i wish we hadn't made that change 17:28:43 dustymabe: that's really cool. did that as well for the papr executor node and it Just Worked (tm) 17:29:04 so all of this is to say - let's propose some workshops 17:29:07 dustymabe: deadline isn't for a week 17:29:15 hmmm, no maxamillion 17:29:26 jberkus: sure, but we should probably come up with some stuff to submit before the day 17:29:37 I want to do a "submitting to FLIBS" or "reviewing containers" work session 17:29:51 jberkus: i think it would be cool to do a 'hands on atomic host' session 17:30:03 less of a working session and more a 'hey fedora, check us out' session 17:30:15 so almost like a hands on lab, but with your laptop and maybe using vagrant 17:30:42 That'd be cool 17:30:51 I haven't submitted a talk because I don't grok the format or think anything I'd have would fit 17:30:53 i'd like to get more people familiar with atomic host 17:30:55 but I don't know, really 17:30:56 Even w/o vagrant, installing a vm from the iso is pretty fast 17:30:59 dustymabe: can you prepare an outline for that? 17:31:16 or from qcow 17:31:18 jberkus: yeah 17:31:19 dustymabe: I'm just not clear what I'd teach people without having a cluster availab;le 17:31:39 jberkus: i think we really only need a single node 17:32:19 so yes there is one 'session' 17:32:31 we should probably submit a 'state of Fedora Atomic' talk 17:32:42 even though those were discouraged, i think people would show 17:33:04 ok, that's yours, oh fearless leader 17:33:22 haha - your words, not mine 17:33:35 we are nothing without everyone 17:33:45 anywho we probably should have some other sessions 17:33:47 Heh 17:33:49 so FLIBS is good 17:33:52 gholms: o/ 17:34:11 i think it would be useful for us to just have a session where we work through some issues 17:34:28 in our issue tracker - maybe almost like a bug scrub - and then breakouts based on individual bugs 17:34:35 so, I'll propose "review a container" 17:34:37 slash issues 17:34:47 question: do we want a "create a container" session? 17:35:12 also, I'm thinking one of us should do "Origin on Fedora" 17:35:25 or "Kubernetes on Fedora", either way 17:35:32 jbrooks: you wanna go to Flock? 17:35:33 jberkus: is that a talk? 17:35:39 jberkus, Yeah 17:35:43 dustymabe: yah, with demos and WIP 17:35:58 ok - that works fine for me 17:36:40 I'd like to see a session on swapping out runtimes 17:36:40 i think that would be good for introducing and detailing our system container work 17:36:46 and how atomic != docker 17:36:58 yzhang: you might be able to help jbrooks with that k8s on atomic section 17:37:04 since I still don't grok it completely or know how to swap them out 17:37:11 sure 17:37:38 roshi: hmm, well we really only have one runtime right now 17:37:45 although CRIO is getting more traction 17:37:54 I thought runc was one that worked 17:37:56 or rocket 17:37:57 dustymabe: we have two, runc 17:38:00 dustymabe: by help do you mean help set up? 17:38:07 dustymabe: and system containers use runc 17:38:09 or is it that we don't have an easy means of swapping them out? 17:38:15 its not so much as swapping 17:38:18 yzhang: we can have co-presenters 17:38:22 they work independently 17:38:32 yzhang: right, but docker also uses runc 17:38:36 docker, system containers, crio-o all use runc 17:38:39 so not really switching anything out 17:38:47 ah 17:38:48 dustymabe: if its not using rkt its probably using runc 17:38:55 k 17:39:04 so we've got multiple managers (engines) available, not multiple runtimes 17:39:15 k 17:39:25 There are different runtime versions 17:39:29 And implementations 17:39:45 The system containers docker stuff is about being able to swap those around 17:39:56 k, - thanks for the clarification 17:40:00 And there's the thing of docker-latest and docker-current 17:40:07 That we ship now 17:40:08 ok so we've got like 4/5 sessions to propose it seems like 17:40:45 let's boil those down 17:40:46 note, I was saying I'd like to see it - I'm *not* in a position to give such a talk 17:40:47 host, runtime, orchestration, containers 17:40:52 :) 17:41:09 #info dusty to submit session for hands on with atomic host "lab" 17:41:21 #info dusty to submit session on 'state of Fedora Atomic' 17:41:37 jberkus: jbrooks: add #info statements for yours 17:41:42 i'll add all of this to the bug 17:41:46 s/bug/issue/ 17:42:04 or just add it ot the issue, doign that now 17:42:24 well was going to add it in a single comment 17:42:27 but w/e works 17:43:05 dustymabe, you could cover enough "state of" stuff in the lab 17:43:18 if they aren't amenable to a state of talk 17:43:43 jbrooks: right, if talk doesn't get accepted we can cover some of that in the lab 17:44:29 ok jbrooks want to info your session? 17:44:38 dustymabe, which is mine? 17:44:53 jbrooks: oops, i think jberkus was proposing it be yours 17:45:06 a talk on "Kubernetes on Fedora Atomic" 17:45:13 jbrooks: Kube on Atomic 17:45:14 and/or "origin on Fedora Atomic" 17:45:14 #info jbrooks to submit session for kube on fedora 17:45:20 I man 17:45:21 mean 17:45:35 fedora atomic 17:45:49 :) 17:45:58 .hello maxamillion 17:46:00 maxamillion: maxamillion 'Adam Miller' 17:46:04 pagure seems to be down 17:46:05 ok i think we are good - please add any more suggested talks 17:46:12 sorry I'm so late, I was at the Dr for a surgical follow-up 17:46:18 maxamillion are you planning to submit a talk to flock on FLIBS? 17:46:43 maxamillion: I was thinking we should do a "Review a container" work session 17:46:49 jbrooks: is planning to do a 'session' on building/reviwing a container 17:47:01 dustymabe: jberkus 17:47:06 dustymabe: I am, I'm also going to show off the new orchestrator cluster architecture, the automatic rebuilds (which are slated to land in August) and talk about multi-arch 17:47:31 ok jberkus jbrooks one of you have to give up jb - too bad for tab completion :) 17:47:37 we can put that to a vote 17:47:44 maxamillion: would you be onboard to help with the container review work session? 17:48:05 #info maxamillion to submit a talk on FLIBS 17:48:38 jberkus: absolutely 17:48:46 well this is all just peachy :) 17:48:53 ok i think we can move to open floor 17:49:00 #topic open floor 17:49:14 who all has something 17:49:26 let's state who has something and then go through them 1 by 1 instead of free-for-all 17:49:37 I got 2 questions for maxamillion 17:49:37 * roshi had a thing 17:49:38 jberkus: had something but then old age and migraine got the best of him :) 17:49:50 * ksinny have question to ask 17:49:50 yzhang: shoot 17:50:02 ok so yzhang and roshi and ksinny have something 17:50:04 there was totally something the other day were we were "meeting tommorrow, let's discuss it there" 17:50:12 I have one other item 17:50:15 maxamillion: firstly, how does https://fedoraproject.org/wiki/Infrastructure/WhatHappenedToPkgdb affect the container registry? 17:50:16 jberkus, kube? 17:50:18 ok jbrooks has something 17:50:23 damn it, jberkus 17:50:37 ok we'll let yzhang go first 17:50:40 ^ 17:50:47 jbrooks: yes, that was it, although we're going to discuss that at the Atomic meeting 17:50:47 the question posted above 17:50:52 right 17:51:34 also for maxamillion: if you want I can take a look at helping with https://pagure.io/atomic-wg/issue/249 17:51:50 * yzhang remembers pagure is down 17:52:01 yzhang: it doesn't affect anything for containers more than it does for rpms 17:52:20 yzhang: is that the ticket about auto populating the VERSION ? 17:52:25 maxamillion: yes 17:52:44 maxamillion: yzhang is berry berry talented! 17:52:50 uh 17:53:04 but we can discuss that later 17:53:08 since we 17:53:11 just saying i have confidence in you be able to contribute 17:53:12 're running low on time 17:53:28 so feel free to continue with other open floor topics 17:53:44 yzhang: grab him in #fedora-cloud after this meeting 17:53:51 ok roshi, you're up 17:54:00 yzhang: yeah, I'd love some help 17:54:03 just looking for a reminder, since I can't recall 17:54:18 yzhang: I have a decent idea of what needs to be done and to what code bases ... it's just a matter of having the time to hack on it 17:54:29 the issues we have with having to rebase to new fedora releases, freeze stopping fixes/features from getting in and whatnot 17:54:33 maxamillion: gotcha, we can talk afterwards 17:54:47 why don't we just build the atomic trees from rawhide and track that? 17:54:50 yzhang: +1 17:55:10 no branches, no freeze, release when our testing shows things are good, and we drive on from there? 17:55:12 roshi: mainly because rawhide has a lot more chaos going on typically 17:55:23 and bugs we find help the rest of the ecosystem 17:55:24 also, there is no gating in rawhide 17:55:37 when we get the CI stuff working, then maybe? 17:56:01 yeah, maybe once we have the CI stuff and it working reliably 17:56:05 Atomic is all about shipping stable pkgs 17:56:07 for now, this is where we are 17:56:13 As defined by bodhi 17:56:14 kk 17:56:18 was just trying to remember 17:56:21 ksinny: you're up 17:56:23 thanks 17:56:29 To get some idea and be part of next VFAD, would like to know that how long VFAD runs and usually what's the timing? 17:56:42 ksinny: so we've had two VFADs so far 17:56:51 that's good to know 17:56:53 i think one of them went fairly long 17:57:00 so maybe like 4 hours total 17:57:08 and the other one was more like 2 hours 17:57:16 the 4 hour one we broke up with a lunch break 17:57:28 ok and does it starts with US timezone? 17:57:30 would be dinner break for you :) 17:57:38 typically starts morning US 17:57:53 dustymabe: if we had enough Indian participants, it would be worth doing a 2X VFAD 17:58:00 dustymabe: ok, thanks for the detail. I will look forward to it 17:58:01 jberkus: sure - i'm not opposed 17:58:01 for sure 17:58:13 timing is based on where the majority of people will be, I think 17:58:20 that is, planning on having two sessions to the VFAD, one early morning US time, the other one late morning 17:58:24 also i'm not opposed to doing a late night FAD 17:58:36 or something like that 17:58:37 if we have enough people in US+India and not a lot from Europe 17:58:49 yah, night is easier for me w/India 17:58:59 any Europeans around? 17:59:20 so maybe we can do that sometime 17:59:22 :) 17:59:36 jberkus: did you think of what you wanted to bring up? 17:59:37 great! 18:00:02 so, two things: 18:00:38 1) The Fedora Atomic Stakeholder Review is complete and reviewed by various parties. I'll put a copy up where all contributors can access it as soon as I scrub the comments. 18:01:38 jberkus: +1 18:01:43 k 18:02:08 2) I'm working to schedule a general Atomic meeting, in order to discuss things which cut across projects 18:02:27 jberkus: +1 18:02:28 one topic for the upcoming meeting is what our plans/policy are for Kube vs. Origin 18:03:01 all atomic_wg contributors will be invited to attend once it's scheduled 18:03:16 #info jberkus trying to schedule a general project atomic meeting in order to discuss things which cut across projects, see https://lists.projectatomic.io/projectatomic-archives/atomic-devel/2017-June/msg00027.html 18:03:21 that meeting will be twice a month 18:03:38 if you're interested, please fill out the time/day availablity sheet here: https://beta.doodle.com/poll/c94n9miwxb7xdq66 18:03:42 jberkus +1 , much needed 18:03:45 +1 18:04:03 anyone else with open floor items? 18:04:05 running over 18:04:24 * roshi has nothing else 18:04:29 3.... 18:04:33 2..... 18:04:36 anyone figured out how to get 25+ hours out of a day yet? 18:04:39 1....... 18:04:40 * roshi could really use that 18:04:49 roshi: fly around a lot maybe 18:04:50 #endmeeting