15:00:09 <jkurik> #startmeeting modularity_wg
15:00:09 <zodbot> Meeting started Tue Nov  1 15:00:09 2016 UTC.  The chair is jkurik. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:09 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:09 <zodbot> The meeting name has been set to 'modularity_wg'
15:00:19 <jkurik> #meetingtopic Bi-weekly Meeting of the Modularity Working Group
15:00:29 <jkurik> #chair dgilmore contyk langdon sct
15:00:29 <zodbot> Current chairs: contyk dgilmore jkurik langdon sct
15:00:38 <jkurik> #topic Roll Call
15:00:42 <contyk> .hello psabata
15:00:43 <zodbot> contyk: psabata 'Petr Šabata' <psabata@redhat.com>
15:00:46 <langdon> oops.. i didn't realize it actually hit 11
15:00:51 <langdon> .hello langdon
15:00:52 <zodbot> langdon: langdon 'Langdon White' <langdon@fishjump.com>
15:00:55 <langdon> thanks jkurik
15:00:57 <sct> .hello sct
15:00:58 <zodbot> sct: sct 'Stephen Tweedie' <sct@redhat.com>
15:01:06 <contyk> feel free to add stuff to our short agenda :)
15:01:30 <jkurik> Hi everybody
15:01:48 <langdon> i only have 30m today anyway.. so short is good :)
15:01:56 <jkurik> The agenda today seems to be short
15:02:02 <jkurik> #topic Agenda
15:02:08 <jkurik> #link http://piratepad.net/modularity-wg-agendas
15:02:14 <jkurik> #info Base Runtime and Generational Core status update
15:02:19 <jkurik> #info Open floor
15:02:23 <langdon> agenda is down for me :/
15:02:26 <jkurik> Anything else someone would like to put to the agenda for this meeting ?
15:03:15 <langdon> general modularity should have more to report/discuss for next meeting.. still trying to sort out infra moving to factory-2 team and a new dev-env for container-ized modules
15:03:27 <sgallagh> .hello sgallagh
15:03:28 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
15:03:36 <tflink> .hello tflink
15:03:40 <zodbot> tflink: tflink 'Tim Flink' <tflink@redhat.com>
15:04:00 <jkurik> #chair sgallagh tflink
15:04:00 <zodbot> Current chairs: contyk dgilmore jkurik langdon sct sgallagh tflink
15:04:14 <jkurik> ok, so lets start with the first topic
15:04:18 <jkurik> #topic Base Runtime and Generational Core status update
15:04:24 <jkurik> contyk: are you the right person to provide the update ?
15:04:30 <contyk> I think so :)
15:04:45 <jkurik> then the IRC is yours :)
15:04:54 <contyk> I just want to share a quick update on the Base Runtime and the associated project, Generational Core
15:05:33 <contyk> if you haven't heard about the Generational Core yet, it's an extended module stack that would provide everything Base Runtime provides with a couple of tools and services nowadays considered standard and would basically define our platform core
15:06:19 <contyk> the Base Runtime module is meant to be a key part of this stack, together with additional two we've dubbed System Runtime and Shared Components for now
15:06:44 <contyk> there might be more than these three in the future making up the Generational Core stack
15:07:01 <contyk> I made a simple diagram just a couple of minutes ago that illustrates how it all fits together
15:07:06 <langdon> oooo
15:07:09 <contyk> https://psabata.fedorapeople.org/modularity/The%20initial%20prototype%20split.png
15:07:11 <langdon> pictures!
15:07:11 <sgallagh> contyk: Perhaps describing the differences between them would be helpful.
15:07:23 <contyk> sgallagh: getting there :)
15:07:37 <sgallagh> contyk: apologies
15:07:54 <contyk> we also have a brief base runtime / generational core architecture overview doc which can be found here -- https://fedoraproject.org/wiki/BaseRuntime/Architecture
15:08:39 <contyk> it explains the role of Base Runtime in more detail, or at least it tries to
15:09:36 <contyk> so, about the modules that make up this stack
15:10:22 <jkurik> Let me summarize for the minutes:
15:10:24 <contyk> Base Runtime would be our hardware enablement layer, providing the kernel, various drivers, bootloaders and a minimal, stable runtime environment -- glibc and libgcc
15:10:30 <contyk> jkurik: ok
15:10:33 <jkurik> #info Generational Core is an extended module stack that would provide everything Base Runtime provides
15:10:38 <jkurik> #info with a couple of tools and services nowadays considered standard and would basically define our platform core
15:10:42 <jkurik> #info It consist of the Base Runtime module with additional two we've dubbed System Runtime and Shared Components for now
15:10:46 <jkurik> #link https://psabata.fedorapeople.org/modularity/The%20initial%20prototype%20split.png
15:10:51 <jkurik> #info There is also brief base runtime / generational core architecture overview
15:10:55 <jkurik> #link https://fedoraproject.org/wiki/BaseRuntime/Architecture
15:11:14 <jkurik> contyk: thanks :), you can move on ....
15:11:37 <contyk> System Runtime would provide all the platform tools and services people expect and would provide the means to experience Base Runtime; this is a joint effort of the Base Runtime and the Modularity teams
15:12:04 <contyk> with the Base Runtime defining the first release the Modularity team will later extend
15:12:19 <contyk> we expect that module [stack] to be fairly large and feature rich
15:12:55 <contyk> the last one would be the Shared Components module, which ships all the runtime dependencies for components in Base Runtime and System Runtime
15:13:39 <contyk> we're currently in the process of deciding which components from the original self-hosting prototype go into which of these buckets
15:13:40 <langdon> other things would also be able to use the shared components as well
15:13:50 * contyk nods
15:14:19 <contyk> to give you all an idea, the self-hosting prototype is currently 5950 binary RPMs, or 2899 source RPMs
15:15:01 <contyk> the original package set that we used as input for dependency resolution was, all the runtime dependencies included, 232 binary RPMs, or 179 source RPMs
15:15:35 <contyk> those 179 source RPMs is what we intend to split into those three Generational Core modules, at least for the first "release"
15:16:09 <contyk> everything else will be, temporarily, in a module we call Generational Core Build, until we find better homes for those components
15:16:19 <contyk> that's for the architecture!
15:16:22 <contyk> but that's not all
15:16:48 <contyk> given that we're working on the first modules in Fedora, we've been also investigating our QE/CI options
15:17:30 <jkurik> #info - Base Runtime would be hardware enablement layer, providing the kernel,
15:17:32 <jkurik> #info various drivers, bootloaders and a minimal, stable runtime environment.
15:17:32 <contyk> we've decided to go with Avocado for now and we're in the process of thinking about what tests we actually want to run in the near future and implementing them
15:17:35 <jkurik> #info - System Runtime would provide all the platform tools and services people expect
15:17:39 <jkurik> #info - Shared Components module ships all the runtime dependencies for components in Base Runtime and System Runtime
15:18:10 <contyk> are there any questions?
15:18:45 <tflink> have there been any conversations around where those tests would be executed or what would be done with the results?
15:19:08 <jkurik> #info Work on QE/CI options is in progress. The team has chosen to use Avocado for now.
15:20:05 <contyk> tflink: we'll be using the modularity infracloud development box for now
15:20:21 <contyk> tflink: how it's going to be hooked into the infrastructure... that we don't know yet
15:20:24 <langdon> tflink, i think the formalized version will be coming from factory-2.. this is more for dev
15:20:37 <tflink> I assume that the target for initial delivery will be F26?
15:21:03 <contyk> correct
15:21:40 <contyk> the plan is to deliver a special edition of F26 Server based on Generational Core + several application-level modules that implment various server roles
15:21:41 <langdon> we need to talk to the server wg about "server edition 26" being this stuff or "modularity server edition 26" ..
15:21:44 <tflink> have conversations with the affected groups been started so that they know what's coming?
15:22:10 <langdon> threebean proposed a separate edition to avoid having to worry about updates for the f26 lifecycle
15:22:25 <jkurik> #info The expected target of QE/CI initial delivery is F26.
15:22:33 <langdon> tflink, what kind of group?
15:22:43 <tflink> infra, qa, releng
15:22:46 <langdon> packagers? or infra? the latter yes, the former, not as much
15:23:03 <langdon> sorry grouping qa, releng, et al under infra
15:23:43 <tflink> as a member of one of those groups, I'm a little nervous about this huge "thing" that's supposed to be landing soon and there don't seem to be many details yet
15:23:45 <langdon> well.. i have been talking to threebean.. i think contyk has too.. and i am hoping he is coordinating as factory 2 is starting to take over that resp :)
15:24:17 <tflink> it's difficult to plan for F26 when there's an elephant lurking in the room :)
15:24:42 <contyk> tflink: it's still all very fuzzy but the good news is, as langdon says, this is going to be an experimental release, available in parallel to the classic one
15:24:50 <langdon> its not hard at all.. it is just difficult to expect the plan to work ;)
15:25:09 <tflink> langdon: i suppose that's one way to put it
15:25:14 <langdon> :)
15:25:14 <contyk> tflink: the qe/ci infra doesn't need to be perfect; another issue is the update system which we don't expect to be able to handle modules by F26 GA
15:25:24 <sgallagh> tflink: Our intention is that the modular server will be non-blocking and experimental
15:25:43 <langdon> sgallagh, is server wg on board with a "separate edition" for f26?
15:25:46 <tflink> just to make sure I understand, the goal is to have something based on F26 available around F26 release but not tied into the fedora release cycle?
15:26:04 <langdon> tflink, mostly.. tied yes.. blocking no
15:26:05 <sgallagh> langdon: I don't really know what you mean by that
15:26:17 <langdon> sgallagh, i can save it for later
15:26:30 <langdon> 16h eastern ;)
15:26:35 <sgallagh> langdon: a separate Edition actually implies a lot of administrative garbage I wouldn't want to get into
15:27:05 <langdon> sgallagh, in infra? or bureaucratic? ralph seemed to think infra was ok..
15:27:21 <langdon> we should add it to the meeting for later
15:27:30 <sgallagh> langdon: Bureaucratic and technical from a package perspective
15:27:30 <langdon> and see if we can get threebean there, eh?
15:27:49 <jkurik> just a note to Fedora updates - For example a Change for Bodhi to work with non-rpm artefacts is scheduled to F27: https://fedoraproject.org/wiki/Changes/BodhiNonRPMArtifacts
15:28:01 * contyk nods
15:28:09 <sgallagh> langdon: Edition handling has to deal with https://fedoraproject.org/wiki/Packaging:Per-Product_Configuration
15:28:31 <sgallagh> We *may* end up needing to do that, but I'd prefer if what we built simply identified itself as Fedora Server
15:28:39 <langdon> Ahh, right
15:29:53 <jkurik> contyk: thanks for the extensive update
15:29:54 * langdon needs to run in a minute but may be here partially
15:30:00 <jkurik> can we move on ?
15:30:01 <contyk> it's been a while
15:30:17 <sgallagh> /me leaves for another meeting
15:30:20 <jkurik> #topic Open floor
15:30:22 <jkurik> anything else for today ?
15:30:32 <langdon> sgallagh should we add it to server wg agenda?
15:30:47 <sgallagh> langdon: Bring it up in Open Floor
15:30:52 <langdon> K
15:30:53 <sgallagh> Or on the mailing list
15:31:31 <jkurik> if there is nothing more, I will end the meeting in 30secs
15:32:23 <jkurik> #endmeeting