17:00:21 #startmeeting fedora_atomic_wg 17:00:21 Meeting started Wed Jul 5 17:00:21 2017 UTC. The chair is yzhang. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:21 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:21 The meeting name has been set to 'fedora_atomic_wg' 17:00:31 #topic Roll Call 17:00:35 #topic Roll Call 17:00:36 spacing 17:00:37 .hellomynameis kushal 17:00:38 kushal: kushal 'Kushal Das' 17:00:40 .hello sayanchowdhury 17:00:41 .hello yzhang 17:00:41 sayan: sayanchowdhury 'Sayan Chowdhury' 17:00:44 .hello jlebon 17:00:46 .hello jasonbrooks 17:00:46 .hello miabbott 17:00:47 yzhang: yzhang 'Yu Qi Zhang' 17:00:50 jlebon: jlebon 'None' 17:00:51 .hello roshi 17:00:53 jbrooks: jasonbrooks 'Jason Brooks' 17:00:55 miabbott: miabbott 'Micah Abbott' 17:00:57 .hello maxamillion 17:00:58 .hello sinnykumari 17:00:59 roshi: roshi 'Mike Ruckman' 17:01:02 maxamillion: maxamillion 'Adam Miller' 17:01:05 ksinny: sinnykumari 'Sinny Kumari' 17:01:31 .hello dustymabe 17:01:32 dustymabe: dustymabe 'Dusty Mabe' 17:01:52 #chair kushal sayan jlebon jbrooks miabbott roshi maxamillion ksinny dustymabe 17:01:52 Current chairs: dustymabe jbrooks jlebon ksinny kushal maxamillion miabbott roshi sayan yzhang 17:02:24 i'll give it another 30 seconds 17:02:49 #topic Action items from last meeting 17:02:51 I'm gonna be listen-only folks. Apologies. 17:02:55 dustymabe to propose a VFAD for wiki and docs discussion 17:02:55 maxamillion to look at logs from last meeting and add summary of our discussion to ticket https://pagure.io/atomic-wg/issue/284 17:02:55 jberkus to create new issues, specs for docs.projectatomic.io publication 17:02:55 maxamillion roshi to come up with guidelines for meeting quorum for the atomic working group 17:03:10 dustymabe: hows the VFAD coming along 17:03:20 roshi: we forgot to do that :/ 17:03:26 yeah, we totally *did* that 17:03:33 give us a couple days to write it down... 17:03:36 * roshi coughes 17:03:37 :p 17:03:51 roshi, :) 17:03:52 #action maxamillion roshi to come up with guidelines for meeting quorum for the atomic working group 17:04:01 re-actioning ;) 17:04:09 #action dustymabe to propose a VFAD for wiki and docs discussion 17:04:16 long sigh was let out just now 17:04:24 no jberkus today 17:04:34 vacay? 17:04:52 there's also maxamillion to look at logs from last meeting and add summary of our discussion to ticket https://pagure.io/atomic-wg/issue/284 17:04:56 .hello deuscapturus 17:04:57 deuscapturus: deuscapturus 'Theodore Cowan' 17:05:03 hiya deuscapturus 17:05:11 yzhang: yeah, +1 ... I'll get that one 17:05:26 re-actioning as well 17:05:29 #action maxamillion to look at logs from last meeting and add summary of our discussion to ticket https://pagure.io/atomic-wg/issue/284 17:06:16 well moving on I guess 17:06:19 to meeting tickets 17:06:25 #topic Make fedora-minimal base image generally available 17:06:31 #link https://pagure.io/atomic-wg/issue/290 17:07:23 ok i think most of the discussion in this ticket is actually unrelated to the ask 17:07:28 am I wrong? 17:07:31 hm, none of the main discussion members are here either 17:07:51 I believe maxamillion and I had discussed this prior to the ticket being opened 17:08:11 and he seemed to be aware of work needed to make this happen 17:08:24 any updates to that maxamillion 17:08:47 since the discussion seems to have detracted to building in the pagure thread 17:08:52 yeah, it'd be pretty easy to do, I'll get it out with the next container release 17:09:01 hmm 17:09:12 well we can't 'get it out' until f26 is up, right? 17:09:12 or wait 17:09:26 dustymabe: no, I'm pushing f26 base images now 17:09:34 dustymabe: docker pull registry.fedoraproject.org/fedora:26 17:10:10 k, interesting - releng gives me a hard time about nightly vs prod qcow images, but this ok :) 17:10:24 so are we good with releasing it during the next container release still 17:10:45 one thing to note is that there will be no fedora-minimal:25 17:10:53 we never built it for f25 17:11:31 dustymabe: huh? 17:11:46 maxamillion: what's the question? 17:12:02 dustymabe: "gives me a hard time" ... what are you trying to do with them? can you not download those images? 17:12:35 dustymabe: or are you trying to push them to Atlas? 17:12:42 maxamillion: sure i can - but putting out nightlies on the website for 'pre-release' testing - was given quite a hard time about that 17:12:44 dustymabe: because do note, I don't push pre-release things to Docker Hub 17:13:21 dustymabe: interesting, we should follow up off-meeting ... I'm curious what that's about, I might be doing something I shouldn't :/ 17:13:21 either way - that's a side discussion 17:13:24 maxamillion: can you update that pagure issue when you release the images 17:13:30 #action maxamillion to release the fedora-minimal base image with next container release: https://pagure.io/atomic-wg/issue/290 17:13:31 yzhang: I can 17:13:35 cool 17:13:50 any more questions/concerns regarding that issue 17:13:50 i'll update the pagure issue today with info from this discussion 17:13:51 dustymabe: just to verify, that image is built and I can query koji for it's latest build, yes? 17:14:07 maxamillion: it's part of the RC composes 17:14:14 dustymabe: rgr, thanks 17:14:20 so it has been building just fine :) 17:14:30 moving on then 17:14:31 #topic Atomic Host images omit many common locales that all other flavors include 17:14:37 #link https://pagure.io/atomic-wg/issue/282 17:15:30 Any updates jlebon? 17:16:50 Real quiet today 17:17:02 yeah - crickets 17:17:13 Too tired from 4th of July I guess 17:17:16 we punted on this last week. discussion in the ticket is kinda muddy 17:17:29 basically balance of more locales vs image size 17:17:41 I guess we can continue punting it until jberkus gets back 17:17:54 any objections 17:17:58 should we remove the meeting keyword for now? 17:18:11 right, it's a tradeoff we have to decide on 17:18:20 Is it a high priority issue at all 17:18:26 don't think so 17:18:51 well - adamw seemed to think it was a pretty big deal and we do occasionally get people who hit the problem 17:18:53 in the Grand Future, people should be able to download with whatever locale they want, as it'll be automagically generated, amirite? 17:19:06 roshi: is that planned? 17:19:12 I'm not worried about 10% bigger image 17:19:17 I detect minor sarcasm 17:19:21 esp since we have cutting kube on the horizon 17:19:56 weigh in on the ticket 17:20:02 * roshi was pontificating 17:20:02 I'm a little worried about it, similar alternatives to Atomic Host are already considerably smaller in size 17:20:11 I can see either side of the argument 17:20:14 If we still undecided we might as well leave the meeting tag there, just so we can occasionally nag people about it 17:20:33 I like the idea of a system container 17:20:50 or at least, alternatives targeting the same use case and user base ... not necessarily "similar" in the sense that they do the same stuff 17:21:08 +1 to system container 17:21:28 however, we need to figure out how to build and release those ... I have no idea how that's being done now 17:21:48 They're pretty much being built as docker containers 17:21:48 * roshi has nothing for this one 17:22:04 It should work just the same as others, like he said ^ 17:22:22 i don't really understand how a system container is going to solve the locale problem 17:22:33 locales are good for many people. 17:22:48 Does that increase the size too much? 17:22:50 Since I'm not super familiar, are locales pretty much just extra files on the host? 17:23:02 yzhang: that's how i understand it 17:23:09 Can someone link to a locales rpm? 17:23:11 system containers can install files to the host pretty easily 17:23:35 but I'm not sure if that's the best use case, if all it is doing is install files 17:23:49 yzhang: right, but there's more metadata isn't there? I'm curious how to release/distribute them 17:23:51 yzhang: where do locale files go? 17:24:08 common_bugs and list how to overlay the package? 17:24:17 what locals do the others use? like CoreOS? 17:24:32 dustymabe: I'm not sure where they would go, for any arbitrary file you can put them into the container at /exports/hostfs/... 17:24:40 and it would get installed to the host, for example 17:25:01 a system container that could install locales on demand would be a nice trick 17:25:08 maxamillion: there's always: https://fedoraproject.org/wiki/Container:System_Container 17:25:11 yeah - i'm just wondering if the files can't be written to a place because we have some filesystems that are readonly 17:25:17 ok we should move on 17:25:24 yeah let's update the issue 17:25:38 with our discussions 17:25:38 the locale data needs to be *exactly* the same as the host today or otherwise things can be partially translated 17:25:54 talk about containers here implies a lot of technical work 17:26:20 in a perfect world the system containers will always have the same base image as the (atomic) host it's on 17:26:25 but that's not a thing yet 17:27:01 anyways, we can discuss further if we have time during open floor 17:27:03 i don't think i agree; among other things, i'd say a goal is to have e.g. the system container of Kube be from a slower moving base image for example 17:27:16 (right?) 17:27:38 possibly, there are always concerns of taking up extra space 17:27:38 walters: that is a valid goal I would say 17:27:51 if one wanted older kube versions 17:27:51 but for the general use case, not having to pull in a different base layer would be nice 17:27:59 anyways 17:28:01 #topic Figure out comprehensive strategy for atomic host container storage 17:28:06 #link https://pagure.io/atomic-wg/issue/281 17:28:24 most recent update was from dustymabe after our last meeting 17:28:26 we can move on - should have removed the meeting label last week 17:28:47 done 17:28:53 cool 17:28:57 #topic fedimg: don't use 'builder' instance for uploading AMIs 17:29:03 #link https://pagure.io/atomic-wg/issue/269 17:29:09 sayan: how we looking? 17:30:20 this right now stalled 17:30:56 plans to pick it back up? 17:31:31 So, the plan to resume working is after the making snapshot public issue fix 17:31:46 ok 17:31:48 thanks 17:31:54 do we have a "blocked" tag 17:32:20 and/or do we want to remove the meeting tag for now 17:32:21 looks like we do 17:32:27 yzhang: probably 17:32:30 i'll remove it 17:32:39 alrighty then 17:32:44 any other questions/concerns on this issue 17:32:47 I will add the meeting tag after F26 release 17:33:16 Alright 17:33:31 moving on then 17:33:34 #topic open floor 17:33:45 I have an item for open floor 17:33:47 That's it for the tickets 17:33:52 nothing here 17:33:54 any open floor issues other than dustymabe ? 17:34:22 * roshi has nothing aside from please test the latest RCs as they show up and leave results on the wiki https://fedoraproject.org/wiki/Test_Results:Current_Summary 17:34:49 yes - if you are interested in the cloud base image or the rest of fedora - please test things out 17:35:17 #topic kubernetes sig 17:35:23 i changed the topic for this one 17:35:33 so we have a group of people interested in kube 17:35:48 https://pagure.io/atomic-wg/issue/287 17:36:08 i'd like to identify the best way to break out this working group 17:36:21 do we create a new issue tracker/mailing list ? 17:36:30 do we piggy back on the atomic working group 17:36:39 I say piggy back 17:36:49 my goal really is to communicate to this group of people in the most efficient way possible 17:37:04 I don't suppose there is such a thing as a sub-group 17:37:07 I say new issue tracker/mailing list 17:37:12 i don't want them to miss messages or ignore messages because the signal to noise ration is too high 17:37:21 Piggy back with tags for issues. 17:37:37 anything that group has to discuss will likely be noise for me and I'm not sure how I'd properly filter it out on the mailing list 17:37:47 create a "kube-sig" tag on pagure? 17:37:49 something like that? 17:37:55 but if I'm the only one, then that's fine ... I'll figure it out 17:38:01 piggyback on the atomic wg 17:38:08 can you filter pagure notifications based on tags? 17:38:39 no clue 17:38:52 It's worth figuring out 17:38:53 dustymabe: no 17:39:01 I don't think so 17:39:02 I'm sure all of us have some things they'd like to filter 17:39:31 I guess I wouldn't mind a new tracker 17:39:35 A new meeting, please no 17:39:47 yeah yeah. no new meeting 17:39:52 :) 17:40:00 lol 17:40:06 noise just comes with the territory 17:40:25 yay :) 17:40:26 ok i'm thinking a new tracker - also sometimes if issues are important to multiple groups then I open a tracker ticket 17:40:37 i.e. sometimes I open an infra ticket and a corresponding atomic-wg ticket 17:40:37 dustymabe: maybe update the issue with the options, and have the interested member vote for it? 17:40:53 yzhang: yeah I can try to do that 17:41:09 since a lot of the people aren't in the meeting today, I'd think they'd have more to weigh in 17:41:14 And then some tracker trackers ;) 17:41:36 jbrooks: well really it would be cool if pagure supported this feature (issue linking) or something 17:41:43 anywho 17:41:45 #action dustymabe to update the kubernetes thread with tracking options 17:41:50 sounds good 17:41:56 I have one other thing 17:41:58 #undo 17:41:58 Removing item from minutes: ACTION by yzhang at 17:41:45 : dustymabe to update the kubernetes thread with tracking options 17:42:01 #action dustymabe to update the kubernetes thread with tracking options: https://pagure.io/atomic-wg/issue/287 17:42:04 put the link there to 17:42:13 sorry, please continue dustymabe 17:42:32 #topic cloud/atomic confusion 17:43:02 i'm going to open a ticket for this too, but there are a lot of people that are confused about what we've done to get here 17:43:10 we're finally advertising our confusion? :p 17:43:11 a lot of people thing cloud base image is dead and atomic is what is left 17:43:12 * roshi ducks 17:43:18 think* 17:43:38 also a decent # of people don't understand why we use the cloud irc channel and the cloud mailing list 17:43:56 yeah 17:44:02 I think fedora 26 is a great time for fedora atomic to have it's own channel and own mailing list 17:44:03 I have an Atomic Mailing list setup actualy 17:44:39 i'll open a ticket for this for us to discuss 17:44:49 #action dustymabe to open a ticket for cloud/atomic confusion 17:44:52 looks like someone also made a fedora-atomic-sig list 17:45:03 does the cloud list continue or it's closed? 17:45:08 maxamillion: interesting - maybe we can consolidate 17:45:19 dustymabe: yeah, probably should 17:45:23 walters: cloud list would continue, but for the cloud base image and related items 17:45:33 so there would be less traffic there 17:45:37 yeah 17:45:45 I made atomic@lists.fedoraproject.org and suggested a few months ago that we switch over .... don't really remember what happened there 17:45:47 walters: opinions on that? 17:45:47 that makes sense, though now there's more places to track :p 17:45:52 we just kind of never did 17:46:16 roshi: that's only because you and i are interested in both :) 17:46:36 lol 17:46:41 ok moving back to open floor 17:46:43 #topic open floor 17:47:03 Everyone ready for flock? 17:47:21 kushal: good ask - who all got talks accepte? 17:47:23 yup 17:47:28 my atomic 101 workshop got accepted 17:47:34 my state of fedora atomic host - did not 17:47:36 * roshi has one as well 17:47:40 dustymabe, We will attend 101 :) 17:47:41 jbrooks and I have a 2-in-1 17:47:45 roshi, Which one? 17:47:55 technically 2 talks I guess 17:48:20 an update on the status of the Upstream First initiative 17:48:50 tldr, getting RH tests ported to upstream and (hopefully) into dist-git, for CI/CD stuff 17:49:13 roshi, Okay. 17:49:33 cool deal 17:49:52 related to this: https://fedoraproject.org/wiki/Objectives/Continuous_Integration_and_Delivery_of_Fedora_Atomic_Host 17:49:58 maybe we can update the ticket that we had originally open to list out who got what talks accepted 17:50:13 +1 17:50:26 my talk on the multi-arch container buildsys got accepted and the workshop that jberkus and I are doing did as well 17:50:26 #info please update the ticket and say what talks were accepted for flock 17:50:29 #unfo 17:50:32 #undo 17:50:32 Removing item from minutes: INFO by dustymabe at 17:50:26 : please update the ticket and say what talks were accepted for flock 17:50:34 but there aren't really any cloud/atomic bits specific to mine 17:50:36 #info please update the ticket and say what talks were accepted for flock https://pagure.io/atomic-wg/issue/279 17:50:55 unfo also makes sense I suppose 17:51:18 zodbot not very lenient 17:51:32 anyhow, we have 9 more minutes, anything else for open floor? 17:51:35 Is there a plan or ongoing work towards making registry.fedoraproject.org available for keeping no-x86_64 container images as well? 17:52:10 maxamillion ^ 17:52:12 yzhang: I do have one more for open floor 17:52:29 yzhang: there is a plan, we'll be using manifest lists for it 17:52:48 well, more so for ksinny ^ 17:52:49 maxamillion: Nice, it will be awesome 17:52:50 stef from the fedora CI efforts has a proposal for actually delivering atomic host from a CI pipeline (i.e. better testing for Atomic Host) 17:52:58 https://lists.fedoraproject.org/archives/list/cloud@lists.fedoraproject.org/message/ZS7XY7NHXGGOPOB2YKNQUWSDUGCMYIL5/ 17:52:58 yzhang: that way we don't have to namespace a bunch of stuff 17:53:11 gotcha 17:53:13 please review that - he's going to discuss it with us next week at this meeting 17:53:27 maxamillion: Is there some place to follow-up or participate in related discussion? 17:53:40 #info stef from the fedora CI efforts has a proposal for actually delivering atomic host from a CI pipeline: https://lists.fedoraproject.org/archives/list/cloud@lists.fedoraproject.org/message/ZS7XY7NHXGGOPOB2YKNQUWSDUGCMYIL5/ please review that - he's going to discuss it with us next week at this meeting 17:53:53 dustymabe beats me to it 17:55:13 Anything else for open floor 17:55:32 nothing from me 17:55:41 Giving it 30 seconds before I start the countdown 17:55:50 25.6 17:55:54 thanks for running! 17:56:02 zodbot should play the final countdown during the last minute 17:56:09 yzhang, Thank you for running the meeting. 17:56:13 RFE music integration 17:56:19 roshi, kushal: my pleasure :D 17:56:43 alrighty then, 17:56:43 3 17:56:44 2 17:56:45 1 17:56:46 #endmeeting