17:01:22 #startmeeting atomic_wg 17:01:22 Meeting started Wed Apr 5 17:01:22 2017 UTC. The chair is jbrooks. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:22 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:01:22 The meeting name has been set to 'atomic_wg' 17:01:42 #topic roll call 17:01:43 .hello bowlofeggs 17:01:43 .hello roshi 17:01:45 bowlofeggs: bowlofeggs 'Randy Barlow' 17:01:47 .hello yzhang 17:01:48 roshi: roshi 'Mike Ruckman' 17:01:49 .hello trishnag 17:01:51 yzhang: yzhang 'Yu Qi Zhang' 17:01:53 .hello dustymabe 17:01:54 trishnag: trishnag 'Trishna Guha' 17:01:57 dustymabe: dustymabe 'Dusty Mabe' 17:02:01 .fas jasonbrooks 17:02:01 jbrooks: jasonbrooks 'Jason Brooks' 17:02:13 .hello miabbot 17:02:15 miabbott: Sorry, but you don't exist 17:02:15 .hello miabbott 17:02:18 miabbott: miabbott 'Micah Abbott' 17:02:39 dustymabe no longer using the .hellomynameis anymore 17:02:43 he too has converted to short-hand 17:02:53 #chair bowlofeggs roshi yzhang trishnag dustymabe miabbott 17:02:53 Current chairs: bowlofeggs dustymabe jbrooks miabbott roshi trishnag yzhang 17:02:56 haha :) 17:03:00 did I get everyone? 17:03:16 looks like it 17:03:17 #topic previous meeting action items 17:03:42 yzhang and I were to talk about containerized kube 17:03:44 And we did 17:04:01 yessir 17:04:03 I'm going to try out a couple of the things we discussed, and get his and giuseppe's feedback 17:04:27 Bascially around making the system container config a bit easier 17:04:54 just curious jbrooks, I forgot to ask, have you ever though about making the entirety of kube in system containers? 17:05:31 yzhang, Yeah, for sure, I was waiting on a howto, but now you've provided that, I'm going to look into converting them 17:05:51 jbrooks: yzhang: but they can still be run as non-system containers, right? 17:05:53 It'll be nicer, because we can reuse the base image that's in ostree 17:05:55 jbrooks: nice, also giuseppe has made some system containers for openshift 17:05:56 Yes 17:05:57 https://github.com/giuseppe/atomic-openshift-system-containers 17:05:58 here 17:06:06 Cool 17:06:07 and yes dustymabe, they can 17:06:10 cool 17:06:24 Yeah, I really like them -- the systemd integration is awesome 17:07:00 OK, tickets 17:07:09 .hello sayanchowdhury 17:07:10 sayan: sayanchowdhury 'Sayan Chowdhury' 17:07:23 #chair sayan 17:07:23 Current chairs: bowlofeggs dustymabe jbrooks miabbott roshi sayan trishnag yzhang 17:07:57 #topic meeting tickets 17:08:05 #link https://pagure.io/atomic-wg/issues?status=Open&tags=meeting 17:08:37 #topic 264 2WK Atomic Release Criteria 17:08:54 roshi, want to talk about it? 17:09:06 can we come back to that in a minute? 17:09:09 I just got a phone call 17:09:16 Yeah, np 17:09:38 #topic 254 Label/comment for primary RPM? 17:09:43 i just pinged adam to see if he can join us 17:10:01 #link https://pagure.io/atomic-wg/issue/254 17:10:08 Cool, do we need him for this one? 17:10:14 .hello maxamillion 17:10:15 maxamillion: maxamillion 'Adam Miller' 17:10:18 sorry I'm late 17:10:21 maxamillion: np 17:10:27 #chair maxamillion 17:10:27 Current chairs: bowlofeggs dustymabe jbrooks maxamillion miabbott roshi sayan trishnag yzhang 17:10:36 topic is 254 Label/comment for primary RPM? 17:10:40 https://pagure.io/atomic-wg/issue/254 17:11:04 sorry, for some reason it didn't signal me 17:11:07 .hello jberkus 17:11:08 now that we're in time change, this overlaps with a weekly video call I'm on ... so ... :( 17:11:08 jberkus: jberkus 'Josh Berkus' 17:11:23 #chair jberkus 17:11:23 Current chairs: bowlofeggs dustymabe jberkus jbrooks maxamillion miabbott roshi sayan trishnag yzhang 17:11:52 So it looks like we're waiting on #249 for this one 17:12:09 Maybe we remove the meeting label and make a note in #249 to re-add it once that's settled 17:12:11 yep, remove the meeting tag? 17:12:16 yeah 17:12:33 cool 17:12:39 that's going to be a while, we did a priority backlog planning yesterday and this work is at minimum 4-6 weeks away from getting started 17:12:47 unless someone else wants to pick it up 17:12:47 OK 17:13:12 #topic 253 lower case vs Initial Cap in Labels 17:13:22 #link https://pagure.io/atomic-wg/issue/253 17:14:28 so I think this is just boiling down to: do we want com.redhat.component or something else like org.fedoraproject.component 17:14:30 i put my comment in there 17:14:42 my biggest thing is "don't block" 17:14:58 +1 17:15:12 +1 to dustymabe 17:15:36 back 17:15:41 +1 17:15:45 was having another meeting 17:15:47 +1 to dustymabe 17:15:52 Ok, assuming we're in agreement, who takes the action? 17:16:04 and what is the action? 17:16:09 heh 17:16:09 is the action just to update the docs? 17:16:11 totally 17:16:13 I think the koji support is already there right maxamillion 17:16:18 so just docs I guess 17:16:18 yzhang: correct 17:16:21 #chair kushal 17:16:21 Current chairs: bowlofeggs dustymabe jberkus jbrooks kushal maxamillion miabbott roshi sayan trishnag yzhang 17:16:23 two: (1) updating the guidelines, and (2) search-and-replace in existing images 17:16:55 well ... search-and-replace takes special permissions, that's not something we normally allow 17:17:03 We can ping ppl to change their own 17:17:12 I can change mine 17:17:13 package maintainers can't just drive by commit to other people's packages 17:17:36 if they don't change their files, what breaks? 17:17:43 dustymabe: nothing 17:17:44 yeah I think theres like me jbrooks and maxamillion's containers, as well as someone's python classroom? 17:17:48 dustymabe: there's backwards compat 17:17:49 nothing breaks 17:17:57 ok then, we send a message to existing maintainers and ask them to update 17:18:01 just not up-to-guidelines 17:18:03 +1 17:18:09 so action items 17:18:13 And who'll update the guidelines? 17:18:15 (1) update guidelines 17:18:21 I'll update the guidelines 17:18:25 so I can remember to do them 17:18:26 (2) send email to existing maintainers + cc cloud list 17:18:30 for system containers too 17:18:47 #action yzhang to update guidelines RE #253 17:19:00 I can send an email to ppl 17:19:18 +1 17:19:20 #action jbrooks to notify container maintainers of guideline change 17:19:30 thanks :) 17:19:35 hashtag progress 17:19:38 if we're going to notify all maintianers, I'd like to get through with *all* the changes 17:19:46 jberkus, good point 17:19:46 or at least all the ones we're currently discussing 17:20:02 jberkus: wants to only notify maintainers once of all changes they need to make 17:20:13 does the previously raised point of versioning count 17:20:31 yzhang: i think we can notify them that they need to use version 0 ? 17:20:43 sure 17:20:59 OK, I'll keep an eye on it and prepare a mini running list of things to notify about -- I'll start an issue for that, I guess? To make sure we get them all 17:21:10 they might rebel against that a bit? was that just guidance or required? 17:21:37 dustymabe: well, if we're going to replace it with automatic, then yeah, we'll need them to do it 17:21:45 however, there will be exceptions 17:21:54 We can say, these are the revised guidelines, here's what's backward compat and what is not 17:22:05 sounds good to me 17:22:13 what is left to resolve on the "container" front? 17:22:27 * dustymabe would really like to have that out of the way 17:22:37 #topic 248 Container Guidelines: Layered Images used as a base for other Layered Builds 17:22:43 that's another, dustymabe 17:23:04 #link https://pagure.io/atomic-wg/issue/248 17:23:25 i think this one is done right? 17:23:29 We haven't heard back from the issue submitter yet 17:23:43 We're just agreeing that the way mine is done is right? 17:24:04 Do we reference this in the guidelines? 17:24:07 Layered images 17:24:08 yeah - i think the issue submitter wants a section in the guidelines about it? 17:24:09 we should 17:24:29 yzhang: mind adding a section on it? 17:24:40 basically 1 - only use base images that were built in fedora 17:24:47 2 - here is an example of how to do it 17:24:57 I can do it -- I'll take his suggested section and make it look like mine 17:25:05 jbrooks++ 17:25:09 jbrooks++ 17:25:20 #action jbrooks to add layered image section to guidelines 17:25:34 jbrooks: can you add the ticket number to that 17:25:44 roshi: can we use undo? to undo the action? 17:25:52 #undo 17:25:53 Removing item from minutes: ACTION by jbrooks at 17:25:20 : jbrooks to add layered image section to guidelines 17:26:03 #action jbrooks to add layered image section to guidelines RE 248 17:26:05 yep, #undo is what you want 17:26:08 yay 17:26:08 * roshi is off the phone 17:26:10 sorry about that 17:26:32 #topic 264 2WK Atomic Release Criteria 17:26:38 dustymabe: regarding about the base images, should we be encouraging using fully-qualified image names? 17:27:14 miabbott: oh, good point 17:27:27 miabbott: like, to include the tag? 17:27:36 miabbott, like what? how does this look: https://pagure.io/atomic-wg/issue/248#comment-432837 17:27:49 i mean the FROM line should be like 'FROM: registry.fedoraproject.org/fedora:25' 17:27:56 not just 'FROM: fedora:25' 17:28:09 Oh, the base 17:28:10 I think miabbott means in general and not just the layered images 17:28:11 jbrooks: yeah! 17:28:18 well 17:28:19 yeah - basically it has to come "from fedora" which in my mind means you have to have registry.fp.o in there 17:28:23 yzhang: yes, in general 17:28:57 i'm just advocating for being very explicit 17:29:08 The official builder will get it from the right place, but if someone built it themselves, it'd come from docker hub, probably, right? 17:29:22 So it would make sense to be explicit, I think 17:29:29 since you can get just about any base image from docker.io if the registry is not mentioned in the FROM 17:29:49 heck, you can even get RHEL base images from docker.io 17:30:07 right, basically when I said: 1 - only use base images that were built in fedora 17:30:17 i meant that they had to use ones that come from registry.fp.o 17:30:24 which requires you to use a fully qualified name 17:30:38 so.... 17:30:45 So we should put that in the example 17:30:51 jbrooks: sure 17:30:53 And in our existing dockerfiles 17:30:57 yep 17:31:12 OK, I can add that to the list of things we'll want to mention to the maintainers 17:31:18 yes to all that 17:31:38 :) 17:31:42 miabbott, You want to make that add to the container guidelines? 17:31:51 can do 17:32:23 #action miabbott to add example of explicit registry inclusion to container guidelines 17:32:32 :thumbsup: 17:32:58 ok, roshi ready to talk release criteria? 17:33:02 yup 17:33:40 so Atomic has lacked a lot of the typical documentation that something in Fedora has 17:33:47 like a PRD and Release Criteria 17:34:19 while we're not a release blocking image, because of 2 week release, it'd still be good to be explicit about base functionality thathas to work for a 2wk release 17:34:30 these are an attempt at codifying these expectations 17:34:31 https://fedoraproject.org/wiki/User:Roshi/QA/Atomic_Release_Criteria 17:34:38 take a look at those, and see if I missed something 17:34:53 +1 17:34:57 but I think that basically covers what we want to work as a bare minimum for each release 17:35:05 and it something we can reference in test cases 17:35:25 I think we should make all the gating tests into release criteria. 17:35:27 it's modeled after the regular Fedora Release Criteria that we tracking blocker bugs against 17:35:39 a decent amount about Fedora Atomic breaks tradition with how things in Fedora habe been done for basically forever 17:35:50 the goal is for us to have tests that cover all the criteria 17:36:01 right 17:36:04 which is fine, but I think we should aim to statisfy the areas that make sense 17:36:10 roshi++ 17:36:17 and has resulted in confusion when other WGs/Groups look at what's being done 17:36:24 roshi++ 17:36:44 this is like a communication wrapper so other groups can see what we're aiming for and know what our targets are 17:36:44 good start roshi +1 17:36:59 roshi: that looks great 17:37:02 since for the last few cycles, Atomic is only really known by members of the WG 17:37:04 just read through it 17:37:04 +1 17:37:05 and we need to fix that 17:37:19 roshi: yeah, we've been getting a lot more involvement though 17:37:20 there's a great list of things I see as us needing, and this is just the first step 17:37:30 roshi, That is correct. 17:37:30 i anticipate awareness to rise 17:37:33 right, but we still haven't really leveled out the confusion 17:37:47 and awareness needs to be written down, not learned all tribally, is my point 17:37:54 roshi: ideas for improving confusion level? 17:37:55 But, we already have particular use-cases which we marked as gating tests for release. 17:38:17 That is why I am suggesting to have those also in the release criteria. 17:38:21 the next question that will require greater involvement from other WGs and releng, is how to treat atomic during pre-release 17:38:34 May be this will make more tests for release than other WG(s). 17:38:39 roshi: yes, that is something we are going to work on for f27 17:38:41 kushal: I'll handle that later, these already cover most of the gating tests 17:38:50 dustymabe: I do have ideas to lower the confusion 17:39:00 we have had discussions with dennis and adam w about it 17:39:17 mostly, writing docs and more visibility with other WGs about what the goals are 17:39:20 we are going to create requirements for releng so we can have a better relationship 17:39:22 and why those goals are good for fedora 17:39:26 right 17:39:26 roshi, but this means we are breaking away from our current criteria. 17:39:35 kushal: what current criteria? 17:39:36 roshi: I should create a ticket for the requirements doc 17:39:43 roshi, Release criteria. 17:39:46 which ones? 17:39:51 there are no Atomic criteria 17:39:55 For Fedora Atomic based on Gating tests. 17:39:57 because it's never been release blocking 17:40:03 tests arent' criteria 17:40:11 hold up one second guys 17:40:23 roshi's "criteria" is different than kushal 17:40:30 kushal, tests vs criteria can be discussed at a later point 17:40:31 right 17:40:42 roshi's "criteria" is criteria that is defined like the rest of fedora 17:40:49 dustymabe I understand. 17:40:51 like what is done to block alpha beta GA 17:41:03 That is where we started our points (from Cloud days). 17:41:08 i know we have our own tests that we rely on to block release 17:41:36 but i think what roshi is saying is that because we don't have things like this explicitly written down, it makes it hard for the rest of the fedora community to understand us 17:41:44 including roshi himself 17:41:49 right, this is bringing atomic more in line with the rest of fedora, so it doesn't feel like two projects quite as much 17:41:56 exactly 17:41:57 i am all for that 17:42:03 kushal: does that make sense? 17:42:13 in the beginning, atomic just kinda came out of no where for those of us who weren't already working on it 17:42:51 +1 17:43:19 roshi: dustymabe We can have the doc written down in https://github.com/projectatomic/atomic-host-docs on *fedora* branch which is going to be hosted on docs.projectatomic.io/fedora_atomic/ 17:43:48 All right, what's our next action? 17:43:54 trishnag: the doc should probably live in the wiki with the rest of the criteria 17:44:05 trishnag: yeah, i'm with roshi on that one 17:44:12 * roshi dreams of the Fedora wiki being as good as the Arch wiki 17:44:19 roshi: OMG so much that 17:44:23 Arch Wiki ++ 17:44:25 Heh 17:44:25 maxamillion: yeah me 2 17:44:31 trishnag I think docs.projectatomc.io is for user docs, not releng docs 17:44:34 * dustymabe wishes we could get the arch folks to write docs for us 17:44:36 credit where credit is due, they do great work 17:44:38 roshi: dustymabe Ack. just like Release Criteria we have for Fedora. 17:44:39 but people neglect the wiki and if we'd put some time into it, I think we can start to get there 17:44:42 at least for atomic 17:44:44 right 17:44:49 jberkus: Understood. 17:44:54 roshi: +1 17:45:03 but that's what I had for this ticket 17:45:03 roshi: i'm the biggest offender of neglecting the wiki 17:45:06 more issues to come 17:45:08 trishnag: thanks for continuing to work on that while I was gone 17:45:10 dustymabe: that you are :p 17:45:11 please poke me if you want me to update a page 17:45:16 always trying to not use it :p 17:45:19 jberkus: Certainly! :) 17:45:26 roshi: i actually voted for wiki this time :) 17:45:38 I noticed, now I need to check that you're OK :p 17:45:53 actually ... we could probably write some kind of updater service script that updated the wiki pages from Markdown, ASCIIdoc, or reStructuredText files in the atomic-wg pagure repo using pandoc 17:46:02 that could be an interesting thing 17:46:08 maxamillion: +1 17:46:15 maxamillion: QA has a ton of ways to automate working on the wiki 17:46:18 so we can do that 17:46:20 maxamillion: yeah - that would be nice 17:46:33 but that's where we're at now, take a look, see if anything is missing 17:46:57 as far as plain english "this and this must work for an Atomic release to be good for public consumption" 17:47:08 OK, so, people need to check out https://fedoraproject.org/wiki/User:Roshi/QA/Atomic_Release_Criteria 17:47:25 you can edit the wiki, and I'll work with QA to find out where exactly those should live officially 17:47:34 and not under User:Roshi/QA/ 17:47:39 dustymabe, Okay[ 17:47:50 OK, and we'll leave the meeting tag on here, and revisit next week? 17:48:01 sure, should be quick if people take a look and sign off 17:48:10 if it looks good, leave a +1 in the pagure ticket 17:48:15 that work? 17:48:26 sounds good 17:48:43 roshi: +1 17:48:52 do we record "everyone" actions like this? 17:49:08 sure, why not 17:49:08 sure 17:49:09 jbrooks, originally the WG is supposed to vote for these items. 17:49:11 I'd do an #info noting that people should look 17:49:12 just for posterity 17:49:22 kushal: what do you mean? 17:49:29 i voted in the ticket 17:49:43 #info check out https://fedoraproject.org/wiki/User:Roshi/QA/Atomic_Release_Criteria and comment in issue #264 17:50:01 kushal, Yeah, I don't think I'm a voting member 17:50:26 jbrooks: we need to get that updated 17:50:40 * roshi doesn't even remember the list of members, tbh 17:50:44 can I also apply to become a member somehow 17:50:47 :D 17:51:07 If nothing else, it needs an update for our atomic wg refresh 17:51:19 yzhang: indeed - send an email to the list and we'll review members and add new ones in a future meeting 17:51:27 miabbott: jlebon might also be interested 17:51:30 Does that mean member of atomic_wg? Then I am :). 17:51:47 Ah, yes, there's a fas thing 17:51:48 * miabbott is cool with being a member 17:51:57 jbrooks: Right. 17:52:17 miabbott, there are various blood oaths and initiation rites 17:52:40 haha 17:52:43 OK, we have two other items 17:52:45 roshi, just for official records we vote on the tickets. 17:52:46 i haven't done one of those in a while. will have to dig out my ceremonial daggers 17:52:54 jbrooks: we might have to put that off. I'm completely out of mashmellow fluff, and so is costco 17:52:55 jbrooks: I also have an item for open ticket. 17:52:56 dustymabe, the clarify older versions ticket 17:53:06 kushal: I just wasn't sure what you meant, because that's exactly what we're doing :) 17:53:13 and the FOSP, which... I don't know if there's anything new 17:53:22 I'll swtich to 17:53:26 jbrooks: hardware is not installed 17:53:28 #topic open floor 17:53:31 older revisions? 17:53:40 https://pagure.io/atomic-wg/issue/228 17:53:42 N-1 17:53:48 ahh ok 17:53:51 I have a quick question, not sure if its the correct place to bring up 17:53:56 yzhang: shoot 17:54:01 ! 17:54:03 trishnag: open floor time 17:54:05 is it possible to get runc 1.0.0 rc-3 into f26 17:54:16 # rpm -qa|grep runc 17:54:16 runc-1.0.0-5.rc2.gitc91b5be.fc26.1.x86_64 17:54:19 this is the current 17:54:20 We need to decide which blog posts we would like to have in our upcoming documentation from http://projectatomic.io/blog . 17:54:23 Here is the ticket to track so https://github.com/projectatomic/atomic-host-docs/issues/21 . 17:54:32 yzhang: talk to container team? 17:54:41 We should open tickets for the blog posts that need to be considered as Doc content item. I will include those tickets in the above ticket. 17:54:45 .whoowns runc 17:54:45 maxamillion: jchaloup 17:54:55 miabbott: jberkus or anyone can help with the same ^ ? 17:54:56 I really want to thank you for the answer about https://pagure.io/atomic-wg/issue/262 17:55:03 trishnag: sure 17:55:03 yzhang: ping the maintainer and find out, I don't see a reason why not 17:55:05 sure 17:55:07 yzhang: we should be able to if we can get it in and make sure it doesn't hose everything first 17:55:10 yzhang, probably ping lsm5 17:55:12 jberkus++ Thanks! 17:55:13 trishnag: Karma for jberkus changed to 6 (for the f25 release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:55:19 And I'm willing to know If there is a way to do what mattdm says, and how to do that 17:55:22 ? 17:55:24 I will ask jan and lsm5 17:55:36 trishnag: I'll go through the backlog, shouldn't take that long 17:55:38 trishnag: i'll have a look as well 17:55:45 x3mboy: let's chat in #fedora-cloud when this meeting is over 17:55:52 My PR to apache libcloud to modify snapshots attribute got merged. Now we can directly integrate the feature to fedimg to fix the copy ami issue that was reported 17:55:56 miabbott++ Awesome. Thank you :) 17:55:56 trishnag: Karma for miabbott changed to 4 (for the f25 release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:56:05 dustymabe, ok cool 17:56:08 ;) 17:56:18 jberkus++ 17:56:18