15:00:01 <nils> #startmeeting modularity_wg
15:00:01 <zodbot> Meeting started Tue Apr  4 15:00:01 2017 UTC.  The chair is nils. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:01 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:01 <zodbot> The meeting name has been set to 'modularity_wg'
15:00:01 <nils> #meetingtopic Meeting of the Modularity Working Group (once every two weeks)
15:00:01 <nils> #chair dgilmore langdon tflink
15:00:01 <zodbot> Current chairs: dgilmore langdon nils tflink
15:00:12 <nils> #topic Roll Call
15:00:12 <langdon> .hello langdon
15:00:12 <zodbot> langdon: langdon 'Langdon White' <langdon@fishjump.com>
15:00:16 <nils> .hello nphilipp
15:00:17 <zodbot> nils: nphilipp 'Nils Philippsen' <nphilipp@redhat.com>
15:00:22 <jkaluza> .hello jkaluza
15:00:23 <zodbot> jkaluza: jkaluza 'Jan Kaluža' <jkaluza@redhat.com>
15:00:59 <tflink> .hello tflink
15:01:01 <zodbot> tflink: tflink 'Tim Flink' <tflink@redhat.com>
15:01:27 <nils> #topic Agenda
15:01:27 <nils> #info We've lost our pad again! (nils)
15:01:27 <nils> #info Follow DST with meeting schedule (asamalik)
15:01:27 <nils> #info Base Runtime status update (contyk)
15:01:39 <contyk> .hello psabata
15:01:40 <zodbot> contyk: psabata 'Petr Šabata' <psabata@redhat.com>
15:01:43 <nils> anything else for the agenda, especially in the light of topic #1?
15:02:16 <nils> No? Good.
15:02:26 <nils> #topic We've lost our pad again!
15:02:34 <contyk> :(
15:02:38 <nils> Subject says it all, basically.
15:02:58 <langdon> where is the new pad?
15:02:58 <nils> The piratepad.net pad broke and only gives me error messages. So I setup a new pad.
15:03:13 <nils> #link https://board.net/p/modularity-wg-agendas New WG Meeting Agenda Pad
15:03:23 <nils> ^^ there
15:03:31 <langdon> yay agenda!
15:03:34 <nils> Let's see how long that one lives :)
15:03:43 <langdon> did you update the wiki already?
15:03:48 <nils> yes I did
15:04:06 <nils> we should update the meeting invite
15:04:20 <nils> jkurik isn't here, let me check who can modify it
15:04:46 <nils> I can, how great is that!
15:04:49 <langdon> ha
15:05:03 <nils> #action nils fix the meeting invite with the link of the new agenda pad
15:05:22 <nils> ok, let's move on...
15:05:25 <tflink> should we be figuring out a method for archiving that pad in case it disappears again?
15:05:50 <nils> Do we need old Agendas? I mean, meetbot archives what we actually discussed.
15:06:25 * tflink can't remember if there were more things than just old agendas in there - notes, to-be-discussed topics etc.
15:06:34 <nils> To me, the greatest hassle is filling the new pad with the template etc., and losing the next/current agenda items.
15:07:14 <langdon> we havent generally put anything on the pad besides agenda and use #info, #action, etc for notes
15:07:30 <tflink> ok, just checking
15:07:46 <nils> You can import/export pad contents, but I don't think there's automation for it yet.
15:08:10 <nils> A more pointed question would be why we don't have an Etherpad in Fedora Infra, but that's a bit out of scope here ;).
15:08:26 <asamalik> .hello asamalik
15:08:27 <zodbot> asamalik: asamalik 'Adam Samalik' <asamalik@redhat.com>
15:08:32 <langdon> ha
15:08:40 <tflink> IIRC, it's not fun to admin and there are some incompatibilities with how fedora infra is set up
15:08:45 <nils> As long as the world produces volunteers who host the pad for us, ... :D
15:08:50 * tflink does not recall the details, though
15:09:01 * asamalik apologizes for coming late
15:09:25 <nils> asamalik: just in time for your topic, just give us a few secs to wrap this one up
15:09:43 <nils> so, anything else besides fixing the meeting invite?
15:09:46 * asamalik just discovered he has a topic
15:09:54 <nils> asamalik: you brought it on ya
15:10:10 * asamalik remembered which one it was
15:10:19 <nils> looks like nothing else for this one
15:10:32 <nils> #topic Follow DST with meeting schedule
15:10:35 <nils> #chair asamalik
15:10:35 <zodbot> Current chairs: asamalik dgilmore langdon nils tflink
15:10:54 <nils> Adam?
15:11:31 <asamalik> most of us are on DST, but the meeting is scheduled on a "non-DST" calendar which means is one hour later - and that might cause collisions for us who are in a DST timezone
15:11:44 <asamalik> my proposal is to move it one hour earlier, so it's at "the same time as before"
15:12:14 <asamalik> any objections? :)
15:12:22 <tflink> not from me
15:12:27 <langdon> im fine with it... but will need to reschedule some other thing
15:12:30 <nils> the idea is that most of us are affected by "Northern Hemisphere DST" in some way, so having the IRC meeting move by not being on DST...
15:12:49 <nils> langdon: you rescheduled that other thing at the last DST change for NA?
15:12:55 <langdon> it was new
15:12:59 <nils> Ahh.
15:12:59 <langdon> its fine..
15:13:02 <langdon> i can move it
15:13:03 <tflink> does it switch with US DST, then?
15:13:17 <nils> With however we put it into the calendar.
15:13:26 <nils> I don't care too much, really.
15:13:48 <nils> It's screw with the schedule of some people during the time when Europe and NA are out of sync re: DST.
15:14:01 <tflink> i don't care so much either but we might want to be clear on which DST switch we're syncing with
15:14:06 <nils> Those two weeks are screwed up in that regard anyway.
15:14:12 <tflink> very much so
15:14:12 <asamalik> I don't really care which DST - any DST is fine. Having two weeks wrong is better than having half a year wrong... (I'm not a big fan of DST, but I'm forced to live in it, so that's why I'm saying "wrong")
15:14:12 <nils> tflink: of course
15:14:38 <nils> asamalik: "out of sync"
15:14:59 <nils> anyway, let me tally up
15:14:59 <asamalik> nils, thanks!
15:15:23 * asamalik is in europe, so he votes for european DST, of course
15:15:47 <nils> I think Europe and US are pretty much in balance as far as # of affected people goes.
15:15:47 * contyk doesn't really care
15:15:55 * langdon doesn't really mind as long as he can put it in his calendar in a way that it moves on its own
15:16:14 <asamalik> +1 langdon
15:16:20 <tflink> the fedora meetings which change with DST that I'm aware of switch with US DST
15:16:25 <tflink> if that makes a difference
15:16:36 <asamalik> so make it consistent then
15:16:41 <asamalik> I'm fine with US
15:16:49 <nils> tflink: are you affected by this, i.e. do you have meetings prior and after that are like this?
15:17:00 <nils> *prior or after
15:17:14 <tflink> nils: nothing that's critical
15:17:28 <nils> just looking for a basis for making the decision :D
15:17:32 <tflink> makes sense
15:17:52 <langdon> nils, i would +1 us-dst for consistency with fedora meetings..
15:17:58 <nils> drawing blanks so far though because everybody in here is so bloody accommodating :)
15:18:07 <nils> yeah, makes sense
15:18:32 <nils> so, agreed? make it 10am EST/EDT?
15:18:39 <tflink> WFM
15:18:52 <langdon> +1
15:18:54 <asamalik> +1
15:18:56 <nils> +1
15:18:59 <tflink> +1 rather
15:19:05 * langdon just typed +1 in some window i am not sure of :/
15:19:17 <nils> tflink: I took it that way
15:19:21 <nils> langdon: ha :)
15:19:33 <nils> okay
15:20:08 <nils> #agreed schedule future meetings to 10am EST/EDT so they move with Daylight Savings Time (or rather, feel as if they stay in the same place)
15:20:24 <nils> #action nils fix the meeting invite a little more wrt DST
15:20:36 <nils> ok next
15:20:41 <nils> #topic Base Runtime status update
15:20:44 <contyk> ok
15:20:44 <nils> #chair contyk
15:20:44 <zodbot> Current chairs: asamalik contyk dgilmore langdon nils tflink
15:20:49 <langdon> sorry
15:20:51 <langdon> one more thing
15:20:56 <nils> oops
15:20:59 <nils> #undi
15:21:01 <nils> #undo
15:21:01 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x2ba6b250>
15:21:06 <nils> langdon: go on
15:21:17 <langdon> nevermind.. you said est.. i wouldn't to make sure "which us-dst" you were using
15:21:26 <langdon> *wanted (not wouldn't)
15:21:37 <nils> langdon: I have a history with time zones, so take extra care :)
15:21:41 <langdon> ha
15:21:59 <langdon> we can move on now
15:22:00 <contyk> undo the undo!
15:22:07 <nils> I don't think that works
15:22:11 <nils> #topic Base Runtime status update
15:22:35 <contyk> so I just have a few quick points about what the Base Runtime team's been doing
15:22:46 * langdon kept hitting ctrl-y but it didn't seem to take
15:23:09 <contyk> first, we now have hourly composes in staging, thanks to jkaluza
15:23:26 <contyk> they're built from the latest base-runtime module content in prod, so it's all fresh, every hour
15:23:57 <contyk> we also build nightly docker base images using those composes, run a few PoC tests and if they all pass, we push the image to Docker Hub
15:24:27 <contyk> we're looking forward to doing this in Fedora infra but we're not quite there yet; hopefully soon
15:24:49 <contyk> we still need to figure out how it all works and especially how to test the image before it gets pushed to Fedora registry
15:25:03 <langdon> can i use the brt base image in docker hub as a base for a f25 container?
15:25:15 <nils> contyk: do you want to #info that or shall I summarize?
15:25:35 <contyk> langdon: that depends on what your container... contains
15:25:44 <langdon> lol
15:25:52 <contyk> langdon: Base Runtime is built from f26 components, with some f27 on top
15:26:14 <contyk> not everything in that package set is binary compatible with f25
15:26:20 <langdon> was really a side bar ... just curious
15:26:26 * contyk nods
15:26:29 <contyk> nils: you summarize :)
15:26:33 <nils> heh
15:26:35 <nils> ok
15:26:53 <nils> #info hourly composes in staging, built from the latest base-runtime module content in prod
15:27:45 <contyk> the base image is available under baseruntime/baseruntime
15:28:02 <nils> contyk: #link?
15:28:02 <contyk> ok, another thing we've been doing -- polishing the module API somewhat
15:28:20 <contyk> #link https://hub.docker.com/r/baseruntime/baseruntime/ Base Runtime-based Docker base image on Docker.io
15:28:34 <langdon> Brt module api, right? Not the spec?
15:28:49 <contyk> yes, the Base Runtime module API
15:29:11 <contyk> we now include development subpackages for nearly all the libraries we ship so building against the module should just work
15:29:33 <contyk> there are a few exceptions, though; we plan to have those fixed within the next two weeks
15:29:53 <contyk> the current state of the API is documented on github... I'll post the link
15:30:03 <nils> #info development subpackages for nearly all shipped libraries included, missing ones to be fixed
15:30:05 <contyk> #link https://github.com/fedora-modularity/base-runtime#api the current state of Base Runtime API
15:30:30 <contyk> next thing is... oh yes
15:30:34 <nils> ok, more #infos
15:31:05 <contyk> we still don't have all the arch-specific & bootloader content in the module; we're currently focusing on how to package shim, for example
15:31:32 <contyk> merlinm implemented a shell-based alternative for extlinux-bootloader; we'll see if it's actually useful
15:31:55 <nils> contyk: more to come, or shall I continue summarizing?
15:32:11 <contyk> note the module doesn't contain Anaconda and there's no higher-level module that does so we cannot actually create qcows or isos yet
15:32:33 <contyk> we'll still looking into how to solve this
15:32:35 <contyk> nils: done now :)
15:32:45 <nils> okay...
15:32:53 <contyk> (two more points after this, though)
15:33:33 <nils> #info docker base images using these composes are built nightly, pushed to Docker Hub after a few tests
15:33:46 <nils> #info doing this in Fedora infra not quite there yet
15:34:14 <nils> #info BRT image currently built from F26 and F27 components
15:34:50 <nils> #info not all arch-specific & bootloader content in the module yet
15:35:27 <nils> #info BRT doesn't contain Anaconda and there's no higher-level module that does so we cannot actually create qcows or isos yet
15:35:34 <contyk> +1
15:35:34 <nils> okay, that's it for summary
15:35:38 <contyk> ok
15:36:00 <contyk> another point is that we'll be "branching" the base-runtime module in two weeks, i.e. building it as part of the f26 stream
15:36:02 <nils> contyk: new topic for your two points?
15:36:09 <nils> aah
15:36:18 <contyk> the master stream will be changed to point to rawhide for all the components
15:36:49 <contyk> once that's done, Boltron module authors should change their [build]dependencies to point to "base-runtime: f26"
15:37:02 <nils> #info the BRT module will be branched for F26 in two weeks
15:37:06 <contyk> we'll use the master branch/stream for monitoring rawhide changes and testing
15:37:32 <contyk> and finally! :)
15:37:33 <nils> #info Boltron authors should depend on "base-runtime: f26" for building then
15:37:52 <langdon> contyk, does the f26 tag exist already?
15:38:17 <contyk> I think our focus in the near future will be, besides fixing the last bits of the Base Runtime API, getting dnf and anaconda packages, somewhere, so we can create iso and disk images in the infra
15:38:28 <contyk> langdon: the branches exist but there's no build
15:38:41 <contyk> only master builds atm
15:38:57 <contyk> and that's all from me...
15:39:05 <langdon> contyk, ok.. so you will let us know when to switch? or do you just want to tag "latest" as "f26" now
15:39:15 <contyk> langdon: yes, on April 20
15:39:27 <nils> #info BRT team near future focus: getting dnf and anaconda as modules so infra can create ISOs and disk images
15:39:28 <langdon> and.. will the f26 tag be in osbs as well? when building there?
15:39:31 <nils> contyk: ^^ ??
15:39:46 <contyk> nils: +1
15:40:14 <contyk> langdon: are you asking whether the Base Runtime-based docker base image will be available in Fedora registry?
15:40:26 <contyk> I actually don't know... hopefully
15:40:29 <contyk> jkaluza: are you around?
15:40:46 <langdon> well.. i assumed that would be true.. really, i meant, "will it have an f26 tag there too"
15:40:54 <contyk> I actually wanted to play with our kickstarts last week but imagefactory is kinda broken right now
15:41:03 <langdon> i think it better be if we want to build module based containers
15:41:21 <langdon> contyk, you might be able to ask courtney about what she did to make it work
15:41:32 <langdon> imagefactory for base images i meant
15:41:42 <langdon> she also wrote a blog post on it IIRC
15:41:44 <contyk> I might also file a bug because the tool shouldn't throw a backtrace when you call it :)
15:42:45 <contyk> okay, I'll also consider this a priority :)
15:42:55 <contyk> although it's a topic with many unanswered questions, still
15:43:10 <contyk> so I don't feel like promising any particular date
15:44:26 <nils> contyk, langdon: anything that should be #info'ed?
15:44:46 <langdon> not from my side i don't think
15:44:53 <contyk> nope
15:44:53 <langdon> just a revisit at the next meeting?
15:44:59 <contyk> +1
15:45:26 <nils> put it on the agenda then, and knock on wood the pad survives the next two weeks :P
15:45:33 <langdon> ha
15:46:23 <contyk> btw, I can build base-runtime in the f26 stream now, if you like
15:46:31 <contyk> and then just keep updating it in both
15:46:37 <contyk> but it won't be stable
15:46:49 <contyk> I consider that April 20 being also the API "freeze" date
15:46:59 <langdon> i think thats ok.. i just don't want to have to keep changing the dockerfiles..
15:47:00 <contyk> right now... it can still change, slightly
15:47:11 <langdon> its not stable on "latest" tag either is it?
15:47:12 <langdon> :)
15:47:14 <contyk> ah, this doesn't affect the docker files
15:47:21 <langdon> no?
15:47:24 <contyk> this is about modulemd files
15:47:28 <langdon> FROM brt:f26?
15:47:45 <contyk> I don't know how that's going to work; I have very little info on that
15:47:54 <contyk> it will probably look similar to that
15:47:58 * contyk shrugs
15:48:10 <langdon> ha.. contyk exudes confidence
15:48:34 <langdon> so which "tag" were you talking about? koji? i thought this was all in context of docker base images
15:48:47 <contyk> I wasn't, you kept saying "tag" :)
15:48:51 <contyk> I was saying branch and stream
15:49:06 <contyk> right now we have base-runtime in the master branch only
15:49:11 <contyk> and we build it from that branch
15:49:29 <contyk> which means the only base-runtime build has "master" as its "stream"
15:49:50 <langdon> ok.. gotcha
15:49:51 <contyk> we'll be populating the f26 module branch and building from there
15:50:03 <contyk> the "f26" stream... and change the "master" stream to follow rawhide
15:50:27 <contyk> right now, when you write modules, you depend on "base-runtime: master"
15:50:48 <contyk> after April 20 you should use "base-runtime: f26"
15:51:03 <langdon> yeah.. .gotcha.. i just read it as a "FROM" line cause i have containers on the brain?
15:51:07 <contyk> so yeah, I can build the f26 build now, if you prefer
15:51:14 <contyk> I guess :)
15:51:37 <langdon> i would think the sooner the better.. so we can make the changes now and forget about it :)
15:51:45 <contyk> alright
15:51:47 * contyk makes a note
15:52:54 <contyk> anything else?
15:53:31 <langdon> making a note: 〰️〰️〰️〰️
15:55:05 <contyk> nils: open floor?
15:55:14 <nils> suits me
15:55:18 <nils> #topic Open Floor
15:55:33 <nils> so, does anybody have anything for open floor?
15:56:43 <nils> tick tock
15:57:37 <nils> okay, nothing is good, too
15:57:46 <contyk> :)
15:57:50 <nils> So, thanks everybody for coming and contributing!
15:57:54 <nils> #endmeeting