17:00:10 #startmeeting Cloud WG 17:00:10 Meeting started Wed Aug 26 17:00:10 2015 UTC. The chair is roshi. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:10 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:19 #meetingname Cloud WG 17:00:19 The meeting name has been set to 'cloud_wg' 17:00:29 #topic Roll Call 17:00:32 who's around? 17:00:34 .hello roshi 17:00:35 roshi: roshi 'Mike Ruckman' 17:00:37 .hello sayan 17:00:38 sayan: sayan 'Sayan Chowdhury' 17:00:45 .hello mmicene 17:00:46 nzwulfin: mmicene 'Matt Micene' 17:00:48 .hello adimania 17:00:49 adimania: adimania 'Aditya Patawari' 17:00:53 .hello jasonbrooks 17:00:54 jbrooks: jasonbrooks 'Jason Brooks' 17:00:55 .fas rtnpro 17:01:00 rtnpro: rtnpro 'Ratnadeep Debnath' 17:01:08 goto turnout today :) 17:01:19 .hellomynameis dustymabe 17:01:20 dustymabe: dustymabe 'Dusty Mabe' 17:01:39 s/goto/good/ 17:01:43 * roshi can type 17:01:59 .hellomynameis jzb 17:02:00 jzb: jzb 'Joe Brockmeier' 17:02:35 .hello jzb 17:02:36 jzb: jzb 'Joe Brockmeier' 17:03:04 just because you .hello twice doesn't mean you get two votes, jzb 17:03:07 hrm 17:03:08 brb 17:03:30 did my "hello" come through? 17:03:34 yeah 17:03:36 yes 17:03:36 having some irc client weirdness 17:03:46 shoulda used irssi :p 17:03:58 #topic Previous meeting followup 17:04:15 roshi: I am 17:04:16 roshi also clarified that you would get only one vote irrespective of number of hellos. :D 17:04:18 there were no action items last week - but does anyone have anything they wanted to touch on from last time? 17:04:19 roshi: :-P 17:04:35 we discussed a bit on ticket 84. 17:04:55 and I managed to write a shell script to run basic build test. 17:05:11 it would be nice if someone can have a look at that. 17:05:14 adimania: why is mesos listed 2x? 17:05:30 ah, nice script 17:05:33 I'm not sure firefox + libvirt are "serious" use cases for Docker (though interesting) 17:05:39 * roshi will take a look 17:05:49 that is because mesos has two dockerfiles. 17:05:55 one for master and one for slave. 17:06:08 adimania: should we change the name of at least one of them so that's clear? 17:06:11 pretty much my only personal use case for docker containers is desktop apps - like Chrome (so I can control it's memory usage easily) 17:06:26 jzb, I would do that. 17:06:33 * jsmith is here (a bit late) 17:06:36 roshi, +1 17:06:37 thanks for working on it adimania 17:07:26 roshi, no problem. I am just happy that the dockerfiles I wrote in the repo are built fine. :D 17:08:21 :) 17:08:31 anything else from last week? 17:08:34 * roshi has nothing 17:08:41 roshi: Can we discuss the Atomic proposal 17:08:42 ? 17:08:57 roshi: I have a meeting in ... 22 minutes with some folks wondering what's up with that and if we need to make Website plans :-) 17:09:02 sure thing - that and tomorrow's test day were the two I was going to do first 17:09:13 roshi: we have a test day tomorrow? 17:09:17 #topic Current Atomic proposal 17:09:23 yep, been scheduled for a while 17:09:30 dusty and I will be getting things around today 17:09:37 so - I *hope* everybody in the channel/in this meeting has read the proposal 17:09:47 jzb: the email you sent? 17:09:50 jzb: got a link for the minutes? 17:09:51 dustymabe: yeah 17:10:28 .hello maxamillion 17:10:29 maxamillion: maxamillion 'Adam Miller' 17:10:32 #link https://lists.fedoraproject.org/pipermail/cloud/2015-August/005660.html 17:10:36 (sorry I'm late, in multiple meetings) 17:11:00 I feel like we are in agreement, but I'd like to get consensus + start moving forward. 17:11:05 maxamillion: welcome. we were just about to start talking about jzb proposal to make atomic the focus of the Cloud WG 17:11:39 I hold by what I said at Flock - I'll back whatever the WG wants to do 17:11:52 but we need to also nail down who's going to be chasing what to make sure things happen 17:11:52 jzb: if I can I'd like to restate 17:12:00 just to make sure I am clear 17:12:18 1. The new primary focus of the WG will be the atomic cloud image 17:12:20 oh crazy 17:12:38 some of the language in that thread concerns me a bit (as an outsider) - is the vision that all cloud users would eventually move over to atomic? 17:12:41 I think it's a really cool idea but I also think it's going to catch a lot of people off gaurd who aren't necessarily paying attention 17:12:42 2. The cloud base image will remain and at least for now will be intentionally boring 17:12:53 dustymabe: +1 17:12:55 tflink: no 17:13:22 dustymabe: well, not "intentionally" boring, but it's just sort of shaped up that way. 17:13:37 dustymabe: if somebody rolls up with a proposal to do something interesting in that, we're not going to prevent it. 17:13:39 as a consumer of that base image, I'm fine with boring :) 17:13:40 jzb: Can you elaborate on *when* you expect to have the two-week cadence working well? 17:13:49 jsmith: "soon" :-) 17:13:50 jzb: yes. well boring because we aren't focusing on it.. boring until we come up with some new good ideas for it 17:14:21 jsmith: they're producing dailies now, we lack specific testing to say "ok, this is a good 2-week release" 17:14:45 jzb: so are my statements relatively correct as far as the proposal goes? 17:15:10 dustymabe: except 17:15:13 cccccccbncgbegdfgniilbghndcjbijrebvnreihikhe 17:15:24 ...? 17:15:27 maxamillion: yubikey? 17:15:38 dustymabe: 3. The group will also be looking to improve and de-embiggen the Fedora Docker image. 17:15:47 maxamillion: easy for you to say 17:15:49 tflink: yup, sorry .. was moving between rooms with my laptop 17:15:54 are more folks committing to testing the deliverables? the coverage for F22 was a bit sparse and oddshocks did a bunch of it 17:15:55 jzb: yes. I forgot that part 17:16:04 https://www.happyassassin.net/testcase_stats/22/Cloud.html 17:16:06 but otherwise yes 17:16:14 jzb: so I like the vision 17:16:15 jzb: I am fine with hvaing more focus on atomic , but not at the cost of base cloud image 17:16:18 jzb: ? 17:16:39 lalatenduM: scroll up. There is no cost to the base image. 17:16:40 jzb: my only concern is that traditionally "boring" == little time commitment 17:16:48 Chasers we need: Documentation, Testing, Releng, Marketing/Tutorials 17:16:49 dustymabe: how much time are you spending on it now? 17:17:02 moving to focus on atomic will be more time commitment I think 17:17:10 jzb: yeah read that, but somehow the msg is coming out like that even if it is not intentional 17:17:19 jzb: I'm not necessarily talking about me but the rest of the group as well 17:17:39 dustymabe: The reality is I don't see anybody really digging in or spending much time on the default image 17:17:42 dustymabe: +1 17:17:44 other than roshi 17:17:45 It would really be nice if we had one or two people "own" some of the responsibilities 17:17:52 jzb: yes I agree 17:17:54 I spend a lot of time with the base image :) 17:18:05 that's the "chasers" I referenced up top 17:18:05 dustymabe: but we're an all-volunteer group, it's not like we can assign work 17:18:17 yes true. 17:18:25 so my proposal basically just reflects reality as it is and trying to catch up w/r/t the site and marketing. 17:18:53 (And a hope we'll also remember the docker image needs love) 17:18:56 ok. yes. I like the vision 17:18:59 I don't see it as "assigning" it's more of "who wants to attach their names to this thing here?" 17:19:06 jzb: it would be nice to make some points of contact for "major initiatives" 17:19:11 roshi, I am ready to be an apprentice and help out with things :) 17:20:13 tflink brought up a good point though, this is moving into untested waters - do we have more people who are stepping up to test these deliverables? 17:20:15 rtnpro++ 17:20:15 dustymabe: Karma for rtnpro changed to 4: https://badges.fedoraproject.org/tags/cookie/any 17:20:30 a lot of the testing for F22 was me 17:20:33 rtnpro++ 17:20:33 jsmith: Karma for rtnpro changed to 5: https://badges.fedoraproject.org/tags/cookie/any 17:20:47 tflink and nirik also testing things 17:21:23 if we're moving toward atomic as the major deliverable, i'm willing to help out with trying to figure out some testing 17:21:24 I'll do some testing as well. I have some resources for setting up atomic and docker. 17:21:36 since my main focus is "upstream" docs 17:21:45 I'll help, too -- atomic is my primary interest 17:21:46 I will be involved a bit more in F23+ Atomic testing and such (fwiw) 17:21:49 but the testing of the images has been sparse, and we still need more testcases written for the base image - let alone Atomic and the docker image to run inside it 17:21:54 also, the base cloud image for F22 was so similar to the server product that it got away with fewer test cases and less testing - who's writing all the test cases for atomic and docker? 17:22:03 for all of those that just spoke up.. we do have a test day tomorrow :) 17:22:06 maxamillion: isn't testing of that under the remit of the team that's currently producing it as well? 17:22:11 * roshi hasn't really tracked atomic at all - precisely because it's not blocking 17:22:13 dustymabe: was that.. promoted anywhere? 17:22:35 jzb: it was announced 17:22:41 roshi: define announced :-) 17:22:41 blog posts will be going up today 17:22:48 jzb: depends on what you define as the team that "produces it" because if that were true then technically rel-eng produces the distro but we certainly don't test everything 17:22:50 sent to the list, discussed in meetings 17:23:06 jzb: we could have done a better job on that 17:23:10 i think dustymabe has the most comprehensive list of tires to kick on atomic we've been using on test days, it definitely could be expanded, updated and automated (?) 17:23:17 roshi: I will gently suggest we need a larger group than that :-) 17:23:21 WG's are responsible for testing - they get their stuff signed off by QA (or, that's how I understand it's supposed to work) 17:23:22 flock/linuxcon got in the way of doing proper promotions 17:23:51 roshi: I think test days are an opportunity to engage the community as well 17:23:56 jzb: sure - but we're here now, and we can have another test day - there's not a limit on them 17:23:57 is kushal about? 17:24:02 that's why we have promoted them more heavily for the past two test days 17:24:08 jzb: he's traveling afaik 17:24:13 roshi: we need a hands-on on how to automqte the testing 17:24:23 number80: welcome 17:24:35 maxamillion: man, what's up with people traveling. amirite? 17:24:35 and to automate it, we could use some more devs working on taskotron 17:24:49 roshi: what do the schedules look like. would we have time to do another test day early september? 17:24:57 lalatenduM: do you have a remit to assist with testing as part of the container tools fun? 17:24:57 roshi, dustymabe, jzb, we also need some reference on how to start and where to start. 17:25:01 yeah, we can do that easily 17:25:06 dustymabe: ^^ 17:25:08 rtnpro: ^^? 17:25:08 jzb: heh 17:25:11 we're getting off track 17:25:21 ok. lets push our official F22 test day to early sept if thats ok? 17:25:26 +! 17:25:30 +1 17:25:33 jzb: nope :/ 17:25:33 so no testday tomorrow then? 17:25:39 and tomorrow the WG members can work on testing 17:25:58 dustymabe: F22 test day? :-) 17:26:08 I'll be available to do some testing.. and we'll fill out pages just for us 17:26:12 jzb: doh 17:26:25 dustymabe: also, this sort of means we'll be versionless really 17:26:34 roshi: how does that sound? 17:26:35 we'd need to have a separate discussion about versions. 17:26:37 jzb, I am not sure (I did not get the question properly) 17:26:42 tomorrow we could just work on test cases in prep for later 17:26:45 jzb: you are talking about 2-week atomic? 17:26:52 rtnpro: any room in your $dayjob to help with Fedora Atomic testing. 17:26:54 dustymabe: yes 17:27:10 so, back to the discussion at hand 17:27:13 jzb: maxamillion: is two week atomic based off rawhide? 17:27:13 the vote on atomic 17:27:14 jzb, I am not sure, but I can do it in my free time :D 17:27:34 dustymabe: no, as per request of upstream projectatomic it will be based of stable+updates 17:27:37 off* 17:27:46 dustymabe: we wanted it to be on rawhide but there were .... issues 17:27:48 jzb: then that means we will still have a major version 17:27:51 I think we should identify the following people for being in charge of: Docs, Testing, Working with releng and Marketing/Tutorials 17:28:07 roshi: +1 17:28:18 dustymabe: yeah, but I think we will confuse people if we try to adhere to that. 17:28:37 dustymabe: it's probably better to have something like a date-based release number 17:28:40 but yeah 17:28:45 roshi: can you officially start a vote? 17:28:47 jzb: sure.. but it also means we are going to have to make "major" jumps at times 17:28:52 versioning is off topic for now, I think - that's details for later 17:29:05 ok then.. vote 17:29:17 https://fedoraproject.org/wiki/Cloud/Governance 17:29:18 who will handle Documentation? 17:29:22 this is what we have for now 17:29:39 reporting to the WG at meetings on status and what's being worked on 17:29:54 I'll take the releng part 17:30:05 * roshi wants to get people in slots before we vote 17:30:08 roshi: so we need to open a ticket? 17:30:08 and try to work with maxamillion on stuff 17:30:24 open a ticket for what? 17:30:41 roshi: https://fedoraproject.org/wiki/Cloud/Governance#Making_Decisions 17:30:46 dustymabe: I'm not point for Atomic at the moment, that's dgilmore but I'm happy to help where ever and when ever I'm able ... if I can wrap up some container stuff I'll refocus on helping with Atomic 17:30:46 roshi: on the page you just linked to ;-) 17:30:55 dustymabe: from the rel-eng side that is 17:30:56 Documentation can be part of marketing and tutorials 17:31:00 that wasn't me, jzb :p 17:31:08 roshi: oh, right 17:31:11 sigh 17:31:16 roshi: i can take on documentation 17:31:21 I can work on tutorials and such as well 17:31:37 sure, ticket sounds good 17:31:41 * rtnpro waits 17:31:45 jzb can you open the ticket? 17:31:48 * dustymabe recalls jsmith liking documentation 17:32:02 Guilty as charged... 17:32:08 and we can settle who;s going to run point for what here 17:32:09 I would like to help on testing but I am not a member of wg as such. 17:32:20 wow 3 ppl willing to do docs? crazy :) 17:32:27 roshi: sure 17:32:27 nzwulfin: Scary, isn't it? 17:32:32 adimania: I can get you set up with testing 17:32:33 dustymabe: whats up? 17:32:44 bring you into the QA cult :P 17:32:50 dgilmore: we just mentioned your name in passing.. sorry to bring you in 17:32:59 dustymabe: it is okay 17:33:00 you can disregard 17:33:04 dustymabe: http://dl.fedoraproject.org/pub/alt/stage/22-20150825/Cloud_Atomic/x86_64/iso/ fyi 17:33:07 roshi, I can help with testing as well 17:33:11 roshi, great! 17:33:35 awesome rtnpro 17:33:36 roshi, i can help you up in that too! 17:33:51 sweet, we can sync up after the meeting in the fedora-qa channel 17:34:17 Is there a particular outline for the documentation? 17:34:19 ok, so we have testing/releng and looks like docs as well 17:34:19 if all of you guys could be around some tomorrow to help with testing/organizing the future test day that would be gret 17:34:21 great 17:34:29 jsmith: we don't have much, as it stands 17:34:34 Or should we start brain-storming? 17:34:44 brainstorming I think jsmith :) 17:35:14 I'm happy to start putting together a skeleton once we have a basic outline 17:35:25 jsmith: yah, brainstorming, I've got some outstanding things from PA.io too 17:35:40 And of course I'm happy to translate docs from just about any format into Docbook 17:35:51 proposed #agreed - Cloud testing group is: rtnpro adimania sayan roshi 17:36:23 roshi: +1 from me 17:36:43 proposed #agreed - Cloud docs group is: nzwulfin jsmith jbrooks 17:36:59 proposed #agreed - Cloud releng POC is: dustymabe maxamillion 17:37:12 ack or patch for any of the three 17:37:13 thats a good group right there :) 17:37:22 ack all 3 17:37:25 * roshi just wanted to get it all written down in one line for easier reading 17:37:34 yay! 17:37:34 even managed to get WG member in all 3 ;) 17:37:36 is that going into a ticket? 17:37:37 add dgilmore to that releng bit please 17:37:55 dustymabe: I figure I'll put stuff on the wiki instead of a ticket 17:37:56 maxamillion: well he is the person we would be working with right? 17:38:05 on the far end 17:38:22 dustymabe: he's the one running point on Atomic releng tasks 17:38:29 maxamillion: I just put in people who are involved with the cloud WG, didn't think I needed to add him to yet another thing :p 17:38:35 I doubt he will be attending the meetings though.. 17:38:41 I can though 17:38:44 roshi: ohhh ok 17:38:46 yeah, alright 17:38:50 if that makes sense 17:39:04 (I'm on another meeting but I am +1 to all things that mean two-week atomic default :-)) 17:39:19 :) - jzb do we need a ticket for that 17:39:22 if people have cloud questions, I'd rather people who see the wiki not pester dennis w/o talking to either of you first 17:39:26 where we can record votes and such? 17:39:26 dustymabe: yes 17:39:32 he has enough on his plate at any point in time 17:39:33 dustymabe: I will open as soon as I am off calls. 17:39:38 jzb: sounds good 17:39:46 thanks jzb 17:39:56 so any nacks to those proposed groups? 17:40:09 ok roshi will create wiki pages.. roshi would be nice if that information lived or was linked from https://fedoraproject.org/wiki/Cloud/Governance 17:40:25 nope 17:40:34 * roshi will put it on the wiki and send an email to the list with my rough ideas as to how they keep the WG updated 17:40:54 #action roshi to create wiki pages with new groups and link to the governance page 17:41:06 roshi: are we on to new topic now? 17:41:09 #agreed - Cloud testing group is: rtnpro adimania sayan roshi 17:41:16 #agreed - Cloud docs group is: nzwulfin jsmith jbrooks 17:41:24 #agreed - Cloud releng POC is: dustymabe maxamillion 17:41:33 we should possibly discuss cloudwg members? 17:41:45 #info actual vote on Atomic will be on a ticket (yet to be created by jzb) 17:41:54 yes, only two candidates 17:41:55 are we up for filling slots now? 17:41:56 number80: created a mail about it 17:42:01 oh, right 17:42:22 go for it dustymabe - then I want to give a brief update on the "test day" tomorrow 17:42:43 so oddshocks is gone and agrimm is inactive 17:43:03 agrimm will still manage our credentials in cloud providers 17:43:04 also haven't seen much from imcleod or scollier since I have been here 17:43:26 I know Ian and Scott are around 17:43:40 hi. 17:43:41 * jsmith apologizes, but must run to another ${DAYJOB} meeting 17:43:48 there he is 17:43:49 thanks for coming by jsmith 17:43:54 scollier: you're the new beetlejuice! 17:43:55 roshi: No worries 17:43:55 scollier: welcome :) 17:44:23 oh man.. was that really a conincidence 17:44:27 hey, sorry for my lack of attendance here lately. 17:44:34 scollier: o/ 17:44:36 not a problem 17:44:44 scollier: missed you at TXLF last weekend :) 17:44:58 maxamillion, i know. next year! 17:45:17 number80: is there an exact # of WG members or is there a minimum or maximum? 17:46:44 no min, no nax 17:46:54 as long as we keep it reasonable 17:47:05 yep 17:47:11 ok. so I say if people are here and they are active and want to be a part of the group then.. why not 17:47:23 once there's enough interest, we might have to change it - but for now that's what we're at 17:47:24 I know adimania and rtnpro sent mails to the list? 17:47:44 I know other people expressed interest but no mails 17:48:19 I guess, I'll organize a vote on trac 17:48:29 that works 17:48:30 dustymabe, yes, we did :) 17:48:35 maxamillion: might be interested as well 17:48:50 what was that one guy's name at Flock long blonde hair? 17:49:01 dameastro 17:49:06 is his irc handle 17:49:06 Blondie 17:49:12 .fas dameastro 17:49:12 dustymabe: 'dameastro' Not Found! 17:49:27 .fas damaestro 17:49:27 number80: jsteffan 'Jonathan Steffan' 17:49:31 dustymabe: +1 17:49:46 wtf zodbot? 17:50:08 soon, fesco meeting 17:50:12 dustymabe, you insulted zodbot? 17:50:22 I guess. man that was harsh 17:50:23 .fire dustymabe 17:50:23 adamw fires dustymabe 17:50:27 dustymabe, zodbot doesn't like you. :P 17:50:43 ok so we will organize a vote 17:50:48 action item for number80 17:51:02 dustymabe: heh 17:51:05 sounds like a plan 17:51:06 #action number80 to organize vote on Cloud WG membership renewal 17:51:30 well, it's about time for open floor 17:51:33 #topic Open Floor 17:51:35 i have some things. 17:51:40 on fedora-dockerfiles... 17:51:41 go for it 17:51:59 i created the 3 branches, f2{0,1,2} 17:52:07 need to figure out best way to maintain that. 17:52:36 need to determine if it's worth it, based on the work maxamillion and other are doing with moving dockerfiles to package maintainers. 17:52:57 if it is worth it, i'd like to see kube yaml files, nulecule files, integration with atomic, etc... 17:53:02 they have been to stagnate. 17:53:25 scollier: I think it's hard to maintain them in there current form 17:53:27 * number80 has fesco meeting soon 17:53:30 oh, and i'm not totally sure how to manage a 3 branch repo like that. 17:53:42 dustymabe, why? 17:53:54 scollier: yeah, that gets long in the tooth pretty fast, there's a lot of the build tooling under rapid development, I'm not certain it's ready for prime time, I'm aiming to have something available in staging by F23 but I don't know about moving it to PROD for general use and opening the flood gates just yet 17:53:56 having someone that is a little closer to what the application is supposed to do is better qualified at testing them 17:54:41 so eventually something like "moving dockerfiles to package maintainers" is probably a good idea 17:54:51 dustymabe, i agree with that. i will say though, that the same infrastructure / scripts used for testing packages, doesn't translate cleanly to images. 17:54:53 so that they get better tested? 17:55:00 scollier: agreed 17:55:12 just wanted to throw that out there.. 17:55:37 I am the package maintainer for fedora-dockerfiles. 17:55:46 dustymabe, i'd also like to flesh out a "best practices guide", for folks who are not familiar with doing the work. 17:55:59 of which i have started, but not a lot of content yet. 17:56:02 and at the moment, I am not sure how to manage the multi-branch setup. 17:56:14 adimania, yea, me too. where can we start? 17:56:22 adimania, we need to make sure each image builds in each branch 17:56:31 each image can launch a container, in each branch. 17:56:37 we need automation around that. 17:56:39 me thinks. 17:56:47 and I would also like to test out the dockerfiles before we actually ship them to repos. I wrote a script which is attached to trac for testing out. 17:56:59 https://fedorahosted.org/cloud/ticket/84#comment:19 17:57:04 adimania, good. i saw that. can we automate that? 17:57:20 as for automation, it would be good if we had someone working with the taskotron folks 17:57:43 they could use the extra devs and then we'd have a SME to help automate things as we need them 17:57:49 roshi, is that what we are picking over a jenkins solution? 17:57:49 and have them all in the same place 17:58:08 scollier: part of the build tooling aim is to have Dockerfiles in DistGit and built with fedpkg similar to rpms 17:58:09 I have not used taskotron, so I cannot really say. I can run it via a github hook on one of my machines though and send out an email. 17:58:28 roshi, dustymabe, one thought i had. was to have, and i don't know when, but have a fed-dockerfiles half day hackathon at the next fedora event? 17:58:33 do you think people would be interested in that? 17:58:33