13:00:06 <mattdm> #startmeeting Council (2017-09-13)
13:00:06 <zodbot> Meeting started Wed Sep 13 13:00:06 2017 UTC.  The chair is mattdm. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:06 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
13:00:06 <zodbot> The meeting name has been set to 'council_(2017-09-13)'
13:00:08 <mattdm> #meetingname council
13:00:08 <zodbot> The meeting name has been set to 'council'
13:00:10 <mattdm> #chair mattdm jkurik jwb langdon robyduck bexelbie jwf
13:00:10 <zodbot> Current chairs: bexelbie jkurik jwb jwf langdon mattdm robyduck
13:00:12 <mattdm> #topic Introductions, Welcomes
13:00:14 <zodbot> mattdm: Error: Can't start another meeting, one is in progress.
13:00:16 <mattdm> #meetingname council
13:00:16 <zodbot> The meeting name has been set to 'council'
13:00:18 <mattdm> #chair mattdm jkurik jwb langdon robyduck bexelbie jwf
13:00:18 <zodbot> Current chairs: bexelbie jkurik jwb jwf langdon mattdm robyduck
13:00:20 <mattdm> #topic Introductions, Welcomes
13:00:24 <mattdm> whoo double meeting. sorry about that :)
13:00:26 <mattdm> hi everyone!
13:00:27 <bexelbie> .hello bex
13:00:28 <zodbot> bexelbie: bex 'Brian (bex) Exelbierd' <bex@pobox.com>
13:00:37 <jkurik> .hello2
13:00:39 <zodbot> jkurik: jkurik 'Jan Kurik' <jkurik@redhat.com>
13:00:39 <langdon> .hello2
13:00:41 <zodbot> langdon: langdon 'Langdon White' <langdon@redhat.com>
13:00:59 <mattdm> Josh declined the meeting so I think he's going to miss again.
13:01:10 <mattdm> Waiting on jwf and robyduck
13:01:29 * mattdm runs away from the keyboard for two minutes. brb.
13:03:47 <ignatenkobrain> .hello2
13:03:48 <zodbot> ignatenkobrain: ignatenkobrain 'Igor Gnatenko' <ignatenko@redhat.com>
13:04:07 * mattdm is back
13:04:15 <mattdm> do we have justin and robert?
13:05:16 <mattdm> Okay, well, continuing with out
13:05:24 <mattdm> #topic Today's Open Floor Agenda
13:05:49 <mattdm> You know the process ... suggest items, I'll arbitrarily rank them, then we go :)
13:06:03 <langdon> i would like to ask for a change for modular server
13:06:10 <langdon> not sure how to phrase it
13:06:25 <langdon> basically, talk about the server wg decision from yesterday re: modular server
13:06:39 <mattdm> "Something About Modular Server"?
13:06:56 <mattdm> How about "Modular Server Edition as Council Priority?"
13:07:15 <langdon> there ya go
13:07:18 <bexelbie> An evening with Modular Server? :P
13:07:23 <mattdm> heh
13:07:27 <langdon> ha
13:07:29 * bexelbie wonders if Modular Server can play a piano
13:07:39 <langdon> not til f28
13:07:41 <mattdm> bexelbie, jkurik Any topics of interest?
13:07:54 <bexelbie> I just got back to Brno last night - I am still digging out from Flock, etc.
13:07:56 <bexelbie> so no
13:08:11 <jkurik> nothing for the Council
13:08:32 <mattdm> Ok. Let's start with the Modularity thing and see where that goes :)
13:08:42 <mattdm> #topic Modular Server Edition as Council Priority?
13:09:14 <jwf|matrix> .hello jflory7
13:09:15 <zodbot> jwf|matrix: jflory7 'Justin W. Flory' <jflory7@gmail.com>
13:09:40 <mattdm> hi jwf|matrix!
13:09:49 <mattdm> langdon, you want to introduce the situation here?
13:09:55 <langdon> sure
13:10:00 <jkurik> langdon: do you mean this one: https://meetbot.fedoraproject.org/teams/serversig/serversig.2017-09-12-20.00.log.html#l-21 ?
13:10:18 <jwf|matrix> Hiya! Sorry I'm late.
13:10:20 <langdon> jkurik: yeah
13:11:38 <langdon> so, the server wg decided yesterday to no block the fedora release to support a modular server. Basically, we have had a bumpy few weeks both in traditonal and modular builds for f27, the modular ones are still not complete and the server wg decided to not do the modular server as the official release because they didn't think it should block workstation and cloud
13:11:47 <mattdm> #link https://meetbot.fedoraproject.org/teams/serversig/serversig.2017-09-12-20.00.html
13:12:16 <langdon> I can't say i disagree with this decision as one WG shouldn't block others for their needs.. however,
13:12:43 <langdon> i think the council should consider a) if modules are important to fedora, should this modular server block fedora in general? or
13:13:03 <langdon> b) somehow decouple the fedora server edition for f27 to release after workstation and cloud
13:13:26 <langdon> all that said, there are a number of blockers on the beta for traditional fedora which may result in a delay anyway
13:13:27 <mattdm> I have some strongly conflicting priorties I am not sure how to reconcile here
13:13:40 <mattdm> For one thing, I'm very interested in keeping close to the schedule
13:14:00 <mattdm> and from my understanding of the state of modular server, there's a lot of work yet to do *and* a lot of unknown unknowns
13:14:28 <langdon> yeah
13:14:31 <mattdm> If we were taking a year to put this together I'd be inclined to say "this is important enuogh that let's hold the train"
13:14:43 <jkurik> langdon: can we have the modular version delivered in paralel to "traditional" version
13:14:46 <langdon> i think it is two at this point
13:14:47 <mattdm> but with two months to go? Let's talk about this "decouple" idea.
13:14:56 <robyduck> .fas robyduck
13:14:57 <zodbot> robyduck: robyduck 'Robert Mayr' <robyduck@gmail.com>
13:15:05 <mattdm> hi robyduck!
13:15:06 <langdon> jkurik: in fact, that is and was the plan.. it is just which is "official"
13:15:11 <robyduck> hi all
13:15:17 <jkurik> hi robyduck
13:15:40 <bexelbie> I don't think that shipping a Boltron like playground object for two releases in a row is bad
13:15:52 <bexelbie> if we get to F29 and have modularity problems that is a bigger deal, imho.
13:16:00 <mattdm> Right, there's two ways to do this. First, we could say "Fedora 27 Server is the same as it ever was. Also, we now have 'Boltron II: The Reckoning!' playing over in theater 2"
13:16:10 <bexelbie> s/29/28/
13:16:17 <langdon> well.. i do.. but.. we would ship better than that.. it would be a "real" release (or close) with updates and all.. it just wouldn't get the attention
13:17:11 <mattdm> OR, we could say "Fedora 27 is releasing with Fedora Workstation and Fedora Atomic editions. Fedora Cloud Base and Fedora Server are still available as options but no longer Editions. Please enjoy this placeholder graphic for Coming Soon Modular Server Edition"
13:17:13 <langdon> i am frightened of treating any option like boltron.. as it really was point in time.. but.. if we treat it like a "real" release.. but call it "the reckoning" that is ok
13:17:14 <bexelbie> langdon, does that mean you believe that what will ship in the next release, without regard to block/slip is going to be usable for the target audiences of server?
13:17:41 <langdon> bexelbie: yes, per the server wg
13:17:57 <langdon> and.. we plan to also ship a traditional server.. just not as the "official server release"
13:18:05 <langdon> *was the plan
13:18:20 <bexelbie> langdon, how will the traditional build be shipped?  It won't be an edition ... so what will it be?
13:18:45 <bexelbie> i.e. could we plan to ship both and just swap their positions if modularity isn't ready?
13:18:57 <langdon> i am not sure.. we hadn't ever gotten that far.. building it is "easy".. we just need to decide what directory it falls in to
13:19:04 <mattdm> bexelbie: Yes, that's absolutely something we could do.
13:19:39 <langdon> sorta .. so thta is kinda my plan.. but 2 things: 1) no one thinks having beta != ga is a good idea 2) it is a lot of work to fully qe both
13:19:43 <mattdm> Although there might be some issues with mirroring; langdon, are we putting modular builds on Fedora mirrors in devel areas at all right now?
13:19:45 <jkurik> bexelbie: afaik the problem is that tomorrow we have Go/No-Go for Beta and we do not want to change our decision (which version of server to deliver) between Beta and Final
13:20:10 <bexelbie> do we have a beta rainday?
13:20:12 <mattdm> Yeah I'm sadly also in the "beta must be like ga" camp.
13:20:22 <langdon> mattdm: i don't understand "devel areas".. but boltron is being mirrored by some mirrors in the /alt
13:20:26 <mattdm> bexelbie: Yes. And we're 99.999999% likely to activate that tomorrow.
13:20:50 <mattdm> langdon https://download.fedoraproject.org/pub/fedora/linux/development/
13:20:51 <bexelbie> I'd like to clarify something
13:20:55 <mattdm> very few mirrors carry alt
13:21:00 <robyduck> bexelbie: this is not really an option I think, we need to set up websites and can't just switch...
13:21:13 <bexelbie> is the council being asked which thing is official for hte next release, traditional server or modularity - then that makes calls on slips and blocks?
13:21:33 <DiscordBridge> <linuxmodder-kn4fto> @Rhea  thanks
13:22:02 <bexelbie> If so, I suggest we ship a traditional official server as modularity/boltron feels like a slipped change which I though we agreed we'd push to the next release from now on
13:22:09 <langdon> mattdm: ahh is see.. right, not in "devel" at the momenet
13:22:14 <DiscordBridge> <Botwinder Mk.II> @Rhea received a thank you cookie!
13:22:46 <mattdm> ugh whatever DiscordBridge is seems to be lagging from previous meeting :(
13:22:48 <robyduck> can we kick this bridge?
13:22:56 <langdon> bexelbie: most changes don't have an objective tied to them
13:23:14 <DiscordBridge> <linuxmodder-kn4fto> next release tho is 27 or 28 @bexelbie  @Mattdm
13:23:17 <bexelbie> langdon, I understand ... it feels like a change :) it is differnt
13:23:23 <mattdm> i dunno who has ops :(
13:23:46 <bexelbie> apparently linuxmodder is connected to this meeting via Discord .. this is why that birdge is talking
13:23:58 <bexelbie> bridge
13:24:11 <langdon> bridges shouldn't talk.. they are bridges ;)
13:24:36 <linuxmodder> i can leave
13:24:47 <linuxmodder> I'm on both sides and didn't realize it
13:24:52 <mattdm> linuxmodder: it's fine if it's you and not some weird bot chatter
13:25:27 <mattdm> robyduck: doing double work to prepare for websites seems like a very fair concern
13:25:27 <linuxmodder> @mattdm, I just busy recently sorry if my return seemed like bot chatter
13:25:47 <langdon> +qe
13:25:52 <dzho> d'oh
13:25:57 <linuxmodder> been doing ansible stuff for respin effort and rhcsa/rhcse study
13:26:00 <robyduck> mattdm: more than double work it is too late for us actually
13:26:03 <bexelbie> langdon, I recongize that modularity is an objective, but I think shipping on time is a bigger objective :)
13:26:10 <mattdm> no it's that a thing called "DiscordBridge" was making apparently random comments. We're in the middle of a council meeting.
13:26:27 <mattdm> bexelbie: Well, there's some wiggle room.
13:26:39 * linuxmodder backs out then
13:27:01 <bexelbie> I read most, but not all of hte meeting logs - it seemed that composes are potentially the real blocker here ... is that the heart of it?
13:27:14 <bexelbie> linuxmodder, council meetings are open, though if heavily attended we'll need to use meeting protocol
13:27:20 <mattdm> There's a list of things. langdon is making a tracker bug I think
13:27:23 <langdon> bexelbie: yeah? however, without a compose we don't know what else are issues
13:27:28 <sgallagh> bexelbie: There are three things blocking modularity at the moment (to varying degrees)
13:27:39 <sgallagh> 1) We don't have a compose for various reasons
13:27:46 <robyduck> langdon: mattdm: would it be an option to ship traditional server and add the modularity during later on as additional server option?
13:27:51 <sgallagh> 2) We don't have a Bodhi that can handle modular updates
13:28:03 <bexelbie> I am just trying to decide how we would handle this if the modular server were already the official release - would it be blocking because of it or because of external causes an hwo does that affect the schedule
13:28:06 <robyduck> (without during)
13:28:09 <sgallagh> 3) We have issues with microsoft signing shim
13:28:30 <langdon> 2) we have a  pretty good handle on
13:28:47 <langdon> 3) is not an "issue" per se.. it "the time ms needs to sign shim"
13:28:48 <mattdm> robyduck: I think it could be, yeah. We have lots of options :)
13:28:49 <bexelbie> so 2 sounds like a beta level blocker to me - 3 - I don't know the ramifications of but feels like somethign we should be able to fix by GA
13:28:52 <sgallagh> Right, and 3) we may have identified a less-than-ideal workaround
13:29:05 <bexelbie> 1 is a release block for lots of things, I presume ...
13:29:15 <sgallagh> bexelbie: Signing usually takes *months*
13:29:16 <langdon> bexelbie: workaround only.. getting a proper sig is months
13:29:18 <bexelbie> based on 2 we shouldn't let it go Beta/GA if we can't reasonably believe today we can update it
13:29:24 <sgallagh> 1) Is only broken for Modular so far as I know
13:29:32 <linuxmodder> langdon,  can't we use the same fedora shim we already use for modularity
13:29:36 <langdon> bexelbie: no question.. but we think we know whats up for that
13:29:37 <sgallagh> And there are quite a few contributing problems
13:29:55 <langdon> linuxmodder: thats the workaround.. the problem is modules rebuild .. and shim we don't want to rebuild.. cause sig
13:30:04 <langdon> so we have to hack it
13:30:24 <linuxmodder> langdon,  so a shim for the shim more or less ?
13:30:28 <geppetto> sgallagh: I though #3 was a problem with "traditional server" too?
13:30:32 <langdon> linuxmodder: exactly
13:30:48 <sgallagh> geppetto: The workaround is easier in that case
13:30:54 <bexelbie> after this meeting I hope someone can point me to the short web page on "what a shim does and why I care :)"
13:31:07 <bexelbie> my Fedora is not stored on an unlevel shelving unit
13:31:08 <sgallagh> As I said, I think 3) is largely under control. I cited it for completeness
13:31:08 <langdon> bexelbie: you don't.. but trusted boot does
13:31:11 <mattdm> There are also a number of problems I'm aware of that aren't on this list yet.
13:31:19 <bexelbie> langdon, ahh, I was hoping that was the answer
13:31:26 <sgallagh> bexelbie: It's the magic piece that allows us to use secure boot on Fedora
13:31:42 <sgallagh> Because manufacturers don't ship the Fedora signing key in firmware
13:31:52 <sgallagh> So we use a shim to go from Microsoft's signed boot to our own
13:32:01 <bexelbie> sgallagh, ty
13:32:01 <sgallagh> This requires Microsoft to review, approve and sign the shim
13:32:08 * langdon notes we can leverage the new bt bug to "fix" people's firmware ;)
13:32:09 <mattdm> #link http://www.zdnet.com/article/shimming-your-way-to-linux-on-windows-8-pcs/
13:32:13 <sgallagh> mattdm: Which are those?
13:32:16 <DiscordBridge> <linuxmodder-kn4fto> @bexelbie  shim is more or less a watergate style module for secure boot
13:32:24 <sgallagh> .fire langdon
13:32:26 <zodbot> adamw fires langdon
13:32:30 <mattdm> Moar issues:
13:32:41 <langdon> mattdm: you typing or want me to
13:32:45 <mattdm> 1. DNF with modularity support is still coming from copr
13:32:54 * mattdm typign but you can add when i'm done :)
13:32:55 <bexelbie> do we believe that all modularity issues can be reasonably expected to resolve in the same window as anything else? I'd rather not force a slip over modularity if it needs more time to bake .. it needs more time ..
13:32:57 <sgallagh> ah right
13:33:09 <mattdm> 2. There are no release criteria, without which we can't reasonably do release validation
13:33:27 <mattdm> 3. Release notes, docs, aforementioned websites
13:33:43 <mattdm> 4. Something about pungi and lorax templates
13:33:44 <sgallagh> Our plan for 2. was to apply the existing release criteria where possible
13:34:00 <sgallagh> And kind of "wing it" on the modular-specific criteria.
13:34:01 <mattdm> 5. Apparently there is no Anaconda
13:34:04 <sgallagh> Which is not a great plan
13:34:17 <langdon> 4) is the same a 1) from sgallagh's list
13:34:38 <langdon> sgallagh: i had planned to do better than that.. but haven't yet
13:34:39 <mattdm> 5b. The current plan apparently calls for a minimal anaconda that can produce images and other artifacts that we can ship, but can't actually be used by end-users to install stuff
13:35:10 <langdon> 2) i had hoped someone in server wg might have taken that on
13:35:16 <sgallagh> mattdm: I *think* the plan was to support kickstarts (and the TUI)
13:35:22 <sgallagh> But as for the full GUI... that's complicated
13:35:25 <mattdm> langdon, is that comprehensive?
13:35:38 * langdon looks at his scary list
13:36:01 <mattdm> sgallagh: That's a _big_ regression for an Edition. I don't mind it as a tech preview but I don't feel great about it for a top level artifact
13:36:04 <bexelbie> This list continues to sound like things that we would shouldn't let create a block/slip.  We wouldn't add a new edition that didn't have most of these things ...
13:36:14 <mattdm> the tui is very limited (no partitioning, for example)
13:36:14 <sgallagh> mattdm: I agree with you.
13:36:32 * mkolman notes the Anaconda team has not been contacted about the modular server effort in any way (just for the record)
13:36:45 <sgallagh> That said, limited research suggests that the overwhelming majority of Server Edition systems are deployed via kickstart
13:36:55 <mattdm> I'd be okay with it for beta as a gigantic handwave
13:36:56 <linuxmodder> 5b failed to parse
13:36:58 <sgallagh> So I was willing to tolerate it (wouldn't pass the "last blocker" test)
13:37:01 <langdon> still some issues with a) system profiles and b) koji tag inheritance.. both of those would be nice for f27 but are probably not blockers.. but it is hard to know on b) until we have a compose
13:37:36 <mattdm> sgallagh: In my experience people often try it with a by-hand install and then create kickstarts from that result
13:37:39 <langdon> mkolman: i don't believe that to be the case
13:37:41 <jkurik> mkolman: no one is blaming Anakonda team, we are just summarizing list of issues
13:39:02 <mkolman> sure, just noting this as there is a non-zero amount of time needed if anaconda changes are required
13:39:21 <sgallagh> mkolman: The problem doesn't *appear* to be requiring changes to anaconda itself at this point.
13:39:35 <mkolman> ok, good to know
13:39:36 <langdon> right... it is just "modularizing it"
13:39:39 <mattdm> mkolman: yeah, and it sounds like there might need to be, in order to make sure that packages from the anaconda module don't leak into the content being installed.
13:39:48 <sgallagh> It's a matter of building modules that support the GUI and don't conflict with other modules
13:40:03 <sgallagh> Which is harder than we expected due to an issue with the way Koji generates buildroots
13:40:11 <sgallagh> This situation is still developing
13:40:35 <sgallagh> mkolman: Do you want to be the point of contact for Anaconda on this?
13:40:40 <mkolman> isn't there a confusion between: 1) building the installation media from modules and 2) Anaconda installing modules ?
13:40:43 <sgallagh> I can try to loop you in where possible
13:40:49 <mattdm> In the long run Anaconda is likely to be one of the biggest beneficiaries of modularity, because you can control change a lot better.
13:41:07 <mattdm> mkolman: maybe? I dunno!
13:41:47 <langdon> mkolman: we hang out in #fedora-modularity..  if you want to see the general chatter
13:41:57 <mkolman> sgallagh: sure, possibly also rvykydal
13:42:12 <mattdm> thanks mkolman!
13:42:14 <mattdm> mkolman++
13:42:14 <zodbot> mattdm: Karma for m4rtink changed to 2 (for the f26 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
13:42:41 <langdon> soo.... release dates vs innovation? ;)
13:42:48 <mattdm> this doesn't solve our immediate question but seems like something we definitely need for the future
13:43:46 <mattdm> langdon, sgallagh How confident are you that this list could be down to a non-scary size in two weeks?
13:43:51 <sgallagh> Straw-man proposal: scrap the F27 cycle, rename the current F28 cycle "F27" and get back to work?
13:44:00 <mattdm> sgallagh: tempting!
13:44:06 * mattdm lights straw on fire
13:44:18 <sgallagh> Scarecrow! NOOOOOO
13:44:21 <langdon> mattdm: 60%?
13:44:25 <langdon> maybe higher
13:44:36 <langdon> the modules are all building.. we just need these pieces put together
13:44:52 <mattdm> 60% is pretty low :(
13:45:16 <langdon> very hard to say without not knowing what we don't know.. we need the compose.. shooting hard for this week
13:45:44 <mattdm> So I think "Keep schedule as is, release with Modular Server Edition" is off the table.
13:45:55 <langdon> really?
13:46:02 <langdon> even with beta very likely to slip anyway?
13:46:11 <mattdm> well, with 60% confidence, yeah
13:46:18 <mattdm> if you said 99.99% I'd have a different opinion
13:46:22 <langdon> ha.. then let's make it 80%
13:46:39 <langdon> i actually think the team thinks we can do it in 2 weeks..
13:46:41 * sgallagh lifts one side of the goalpost
13:46:45 <langdon> but, im conservative
13:47:00 <robyduck> what is the release criteria? What we want to ship for F27GA needs to go also through beta, right? So that shipping traditional server means to slip the whole release and retry for F28?
13:47:17 <langdon> or release it on the side
13:47:21 <langdon> like boltron
13:47:33 <mattdm> robyduck: Yes, but in that case we'd release it as Boltron II: Spouse of Boltron
13:47:47 <robyduck> ok, got it finally :)
13:47:51 <mattdm> :)
13:47:52 <langdon> which sounds "fine" .. but means much less usage
13:47:53 <jkurik> langdon: release it on the side is my personal preference
13:48:19 <robyduck> langdon: but we would have the possibility to test it more, given the long list I saw before
13:48:26 <robyduck> jkurik: ,1
13:48:33 <robyduck> +1 I mean
13:49:18 <langdon> so.. of course we would have more time to test it.. but, I am not sure that is a reason to stop it.. i would point out a number of other things that have shipped over the years that could have stood more testing
13:49:31 <langdon> gnome 3 comes to mind
13:49:41 <langdon> but.. i think we need to be careful
13:50:02 <langdon> as in .. we need to allow for major changes.. and they aren't all gonna be perfect.. it is a fine line of what is ok
13:50:35 <langdon> i just don't think slipping the schedule by a bit, especially considering we are likely to anyway, is that big a tradeoff
13:50:48 <mattdm> So anyway, I guess with 80% we could leave that at the edge of the table and revisit next week
13:50:59 <mattdm> The other options I see are:
13:51:14 <mattdm> - Declare this super-important, slip whole release by about a month
13:51:15 <mattdm> or
13:51:45 <mattdm> - Declare ourselves in an era free from the tyranny of releasing all editions at once
13:52:17 <bexelbie> your second option, while exciting really opens up some broader questions we should consider first, imho
13:52:20 <sgallagh> mattdm: That has rel-eng implication that would probably result in the first alt as well
13:52:21 <mattdm> and ship workstation on GA day, along with atomic build, and then put a "coming soon!" placeholder in place of server
13:52:22 <langdon> there is one more (assuming you were done).. create a new edition, that is "beta" for f27
13:53:43 <mattdm> langdon: how would that look
13:53:44 <mmcgrath> .hello mmcgrath
13:53:45 <mattdm> ?
13:53:45 <zodbot> mmcgrath: mmcgrath 'Mike McGrath' <imlinux+fedora@gmail.com>
13:53:54 <mattdm> oh hi mmcgrath!
13:54:21 <mmcgrath> FWIW, modularity is incredibly important to Red Hat, and anything we can do to ensure some form of modularity makes it into F27 in an official (non-side car) capacity, even if that means delays, would be preferrable
13:54:29 <langdon> just a 4th edition .. managed by the server wg, that for f27 would either release later or would be "beta" with lots of updates after ga
13:54:45 <asamalik> I have a crazy comparison.. what Apple did yesterday.. they released a new iPhone 8 which is just another iteration of iphone... and they also announced iPhone X, a completely redesigned thing, but available in late November.. and they put a fancy picture on the website.. similar to what mattdm said...
13:55:09 <mattdm> mmcgrath: Thanks. That's important to consider.
13:55:33 <mattdm> mmcgrath: What do you think about the official-non-sidecar-but-not-GA idea?
13:55:48 <langdon> mattdm: aka 4th edition?
13:56:14 <mmcgrath> I haven't fully processed what it will mean in terms of user experience, user exposure, and updates
13:56:43 <mattdm> jkurik: we have prioritized bugs meeting in this channel next -- okay if we steal 15 minutes from that?
13:57:02 <jkurik> mattdm: yes
13:57:23 <mattdm> mmcgrath: I think it might actually *increase* user exposure
13:57:41 <mattdm> Because reviews tend to be "Fedora came out today. There are three editions. I focused on Workstation..."
13:57:57 <mattdm> With a staggered release, we could possibly get seperate articles for each
13:58:06 <mattdm> especially with Modularity being Something Diferent
13:58:44 <robyduck> mattdm: then I would prefer to delay just server and not set up a fourth edition, which is impossible to set up for websites right now
13:58:56 <bexelbie> I think decoupling can really provide us with some differentiation and marketing value
13:58:57 <sgallagh> Yeah, I think I'm with robyduck on this.
13:59:02 <mattdm> robyduck: Yeah, that's my thinking too.
13:59:11 <sgallagh> Plus, that will give us a little more leeway with some ongoing FreeIPA issues.
13:59:20 <mattdm> Do, or do not. There is no half do and keep the other thing.
13:59:24 <mattdm> That's from Star Wars
13:59:25 <bexelbie> how does freeipa factor into this?
13:59:40 <mmcgrath> mattdm: at a mminum I like that aspect of it then ;-D
13:59:44 <sgallagh> bexelbie: There are release criteria for Server that require FreeIPA to work
14:00:01 <sgallagh> It's one of the ways we determine if it's acceptable to release to users
14:00:52 <mattdm> robyduck: I'm sure mmcgrath can help make sure that red hat designer resources available to help so that you're not just stuck with a whole ton  of extra work for this
14:01:00 <langdon> but the freeipa issues are unrelated to modularity, IIRC
14:01:17 <sgallagh> langdon: As far as I know, that's correct.
14:01:29 <sgallagh> Actually, let's talk about what this means project-wide
14:01:48 <sgallagh> If we give Server one extra month without doing the same for the others, how do we handle schedule dates?
14:01:57 <sgallagh> Particularly freezes.
14:02:20 <langdon> i think that all stays the same.. this is an exception..
14:02:26 <robyduck> mattdm: ok, but let's see how much time we have to do that
14:02:30 <sgallagh> langdon: It cannot
14:02:33 <langdon> later we can revisit the discussion of true decoupling
14:02:47 <mattdm> sgallagh: I think we'd still have a GA "Everything" release on the regular schedule
14:02:57 <mattdm> "Everything non-modular"
14:03:08 <sgallagh> langdon: We have to have the ability to freeze Server Edition to prepare its Beta and GA
14:03:09 <robyduck> sgallagh: IIRC we also did that for Atomic already?
14:03:27 <robyduck> or was this before it became an edition?
14:03:35 <sgallagh> Though I suppose we can probably do that by using explicit git hashes in the module definitions.
14:03:35 <mattdm> The fact that modularity has a different build system and goes into a different tree gives us some flexibility -- it actually *is* decoupled
14:03:40 <sgallagh> That's a little annoyingly manual though
14:04:18 <sgallagh> mattdm: Sure, but that doesn't answer open questions about traditional Server Edition
14:04:29 <sgallagh> Will that stick to the same schedule as Fedora Workstation, then?
14:04:45 <sgallagh> Or are we asserting we need to drop it?
14:04:52 <bexelbie> sgallagh, I thought it was an "or" we either have an official traditional or an officail modular server
14:04:57 <sgallagh> (I'm just trying to make sure we know what we're getting into)
14:04:57 <mattdm> sgallagh: It'd stick to the same schedule as "Fedora Everything Build"
14:04:59 <bexelbie> the other becomes a lab/spin/something
14:04:59 <pingou> (which comes as, overruling the SIG decision)
14:05:14 <langdon> i guess i meant, from a fedora marketing perspective, the GA will still, generally, fall at the same time.. so the "freezes" are the same.. modularity just lets us decouple when necessary..
14:05:17 <sgallagh> pingou: The SIG decision can be remade if we're told we have an extra month
14:05:28 <mattdm> pingou: "Asking SIG to reconsider in light of high-level project priorities"
14:05:30 <pingou> sgallagh: fair
14:05:34 <mattdm> "and mmcgrath asking nicely" :)
14:05:48 <robyduck> sorry I have to run, if you need my vote to anything: +1 to make a release exception for Server and only for F27 becasue we need to go with modularity. Otherwise if that is not doable, second option is the sidecar option and release traditionally.
14:06:09 <robyduck> will read the minutes. See you.
14:06:22 <mattdm> robyduck: thanks!
14:06:36 <sgallagh> mattdm: Could you phrase something like you are going to propose it to the Server SIG for agreement?
14:06:46 <sgallagh> I just want to know precisely what we're asking for.
14:07:03 <mattdm> yeah hold on let me compose
14:07:06 <sgallagh> Thanks
14:07:52 <langdon> pingou: and we would only over rule a WG decision ;)
14:08:53 <pingou> langdon: :s
14:09:03 <mattdm> Proposal: Modularity is an important Fedora priority, and having it be a non-sidecar option is important to Red Hat as a stakeholder. Since Modularity allows us to decouple the Modular Server release from the GA date, we'd like to take advantage of that and ask the Server SIG and Modularity WG to come up with an alternate schedule landing about a month later.
14:09:44 <mattdm> wait that's not all because we have to get "as the official edition" in there
14:09:49 <langdon> s/SIG/WG
14:09:54 <mattdm> and that
14:10:00 <sgallagh> langdon: The Server WG no longer exists.
14:10:02 <sgallagh> Leave it as SIG
14:10:22 <mattdm> Proposal: Modularity is an important Fedora priority, and having it be a non-sidecar option is important to Red Hat as a stakeholder. Since Modularity allows us to decouple the Modular Server release from the GA date, we'd like to take advantage of that and ask the Server SIG and Modularity WG to come up with an alternate schedule landing as an official Edition about a month later.
14:10:23 <sgallagh> We revised the governance rules yesterday to be more open
14:10:33 <sgallagh> I need to update the governance page today
14:11:02 <langdon> contyk: threebean: any thoughts on ^^ ?
14:11:19 <langdon> sgallagh: i saw all that.. missed WG->SIG
14:11:19 <mattdm> Looking at the time: Let's make a ticket for this.
14:11:35 <mattdm> langdon, can you open the ticket starting with what I just wrote with any rephrasing?
14:11:38 <pingou> should the SIG try to write down their own schedule and requirements then?
14:11:44 <sgallagh> Well, it's more WG->Shed-painters ;-)
14:12:01 <mattdm> sgallagh: shedpainters sounds like WG to me :)
14:12:08 <langdon> mattdm: yeah.. although i don't understand the "rephrasing" part
14:12:17 <bexelbie> mattdm, Does this mean that we will have F26 server still on the site for 1.5 months?
14:12:20 <bexelbie> 1 month, I mean
14:12:26 <mattdm> langdon: like, anything I wrote that sounds stupid, right it differently.
14:12:42 <mattdm> bexelbie: I think we have a lovely "Coming Soon!" graphic in its place
14:12:49 <langdon> liek right->write ?? ;)
14:13:06 <pingou> or liek -> like? :D
14:13:07 <mattdm> with a small "Looking for F26 server now? Go here..." bit
14:13:13 <langdon> look, we are actually making robyduck's job easier! ;)
14:13:14 <mattdm> langdon, pingou yes, those
14:13:20 <bexelbie> I presume it is too much work to get an F27 traditional server out and then replace it with the new out of band releases a month later?
14:13:39 <langdon> bexelbie: we talked about that .. but we think it would be bad for users
14:13:50 <bexelbie> but it makes it feel less like a slip and more like a plan :)
14:14:10 <langdon> we don't think we could, cleanly, do it without a reinstall.. so there is that
14:14:12 <sgallagh> bexelbie: If nothing else, it means doubling the work on QA, which isn't fair.
14:14:19 <mattdm> I think it's more *confusing* than too much work.
14:14:23 <bexelbie> hence my presumption of too much work :)
14:14:25 <mattdm> but also too much work :)
14:14:44 <mattdm> okay, let's wrap this up and move to prioritized bugs meeting
14:14:50 * bexelbie isn't sure about the confusion part - that is a messaging issue, but yes
14:14:51 <mattdm> because we've neglected that for most of the summer
14:15:04 <contyk> langdon, mattdm: so you'd like to postpone Server, effectively?
14:15:15 <langdon> contyk: right
14:15:23 <contyk> ok
14:15:44 <langdon> contyk: but let's take it to #f-modularity  and/or the ticket i am gonna make if you have qs
14:15:49 <langdon> so the other meeeting can get started
14:15:53 <mattdm> contyk: No, I'd like to ship it on time. But, given reality, I'd like to ship it *well*.
14:16:47 <mattdm> okay, ending this meeting in 60 seconds unless someone has something really exciting to add
14:17:09 <mattdm> let's continue discussion in to-be-filed-soon ticket
14:17:11 <jkurik> the fact that we do not know how to deal with the server edition at the moment is a blocker for the Go/N-Go meeting tomorrow IMO
14:17:47 <mattdm> jkurik: Yeah. But https://qa.fedoraproject.org/blockerbugs/milestone/27/beta/buglist already says to me that it's an academic blocker
14:18:07 <jkurik> right
14:18:25 <mattdm> okay, moving on. thanks everyone. thanks mmcgrath for the rh perspective
14:18:30 <mattdm> #endmeeting