19:00:02 <mattdm> #startmeeting Council (2018-04-18)
19:00:02 <zodbot> Meeting started Wed Apr 18 19:00:02 2018 UTC.  The chair is mattdm. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:02 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:00:02 <zodbot> The meeting name has been set to 'council_(2018-04-18)'
19:00:04 <mattdm> #meetingname council
19:00:04 <zodbot> The meeting name has been set to 'council'
19:00:06 <mattdm> #chair mattdm jkurik jwb langdon robyduck bexelbie dperpeet Amita nb dgilmore pbrobinson
19:00:06 <zodbot> Current chairs: Amita bexelbie dgilmore dperpeet jkurik jwb langdon mattdm nb pbrobinson robyduck
19:00:08 <mattdm> #topic Introductions, Welcomes
19:00:21 <mattdm> Good afternoon, evening, and middle of the night, everyone!
19:00:26 <mattdm> possibly also morning
19:00:28 <bexelbie> .hello bex
19:00:29 <zodbot> bexelbie: bex 'Brian (bex) Exelbierd' <bexelbie@redhat.com>
19:00:37 <dgilmore> hola
19:00:45 <mattdm> (Wouldn't a flat earth with one timezone be easier?)
19:01:04 <bexelbie> Yes
19:01:40 <langdon> .hello2
19:01:41 <zodbot> langdon: langdon 'Langdon White' <langdon@redhat.com>
19:02:06 <langdon> mattdm: just use utc all the time
19:02:31 <dgilmore> mattdm: can you use the fedora orbital laser to make that happen?
19:02:33 <mattdm> then I'll be five hours early for everything at home
19:02:37 <mattdm> four hours during winter
19:02:43 <mattdm> anyway.
19:02:45 <dgilmore> mattdm: kill DST
19:02:47 <mattdm> #topic Today's Agenda
19:02:59 <mattdm> This is our (delayed one week) Objectives Update meeting
19:03:23 <mattdm> We have langdon here... and I *think* pbrobinson will join soon
19:03:40 <mattdm> and dperpeet has something but asked it to be not first (I think he has a meeting conflict)
19:03:47 <mattdm> also we have a budget ticket to finalize
19:04:08 <mattdm> I suggest we have Modularity first, and then IoT if Peter shows up, and then CI
19:04:19 <dgilmore> works for me
19:04:31 * langdon thought the whole idea of CI was to do it first </snark>
19:04:36 <mattdm> and either put budget at the end, or stick it in the middle if we need filler :)
19:04:56 <mattdm> langdon: _continuously_ actually
19:04:56 <dgilmore> poor bexelbie
19:05:05 <mattdm> anyway, langdon, you ready?
19:05:09 <langdon> sure
19:05:23 <mattdm> #topic Fedora Modularity Update from Landon
19:05:37 <langdon> #undo
19:05:37 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x18283190>
19:05:42 <langdon> #topic Fedora Modularity Update from Langdon
19:06:32 <langdon> ok.. so .. i am having a bit of trouble with "what to update"
19:06:44 <langdon> however.. we have modules, we have defaults, and we have some bugs
19:07:00 <mattdm> (sorry, the g key on my keyboard has some issues, because thinkpad)
19:07:13 <langdon> so.. re: bugs.. we really need https://bodhi.fedoraproject.org/updates/FEDORA-2018-183f616405 to go through (im about to karma it after meeting)
19:07:34 <langdon> mattdm: ha.. no worries.. but that is how people end up calling me "michael"
19:08:07 <mattdm> Let's maybe start higher level. For people who haven't been following closely, what's the picture for modularity in F28 release?
19:08:32 <langdon> we would like to get a few more modules in .. and have candidates... but have had some chicken/egg questions re: the new arch.. e.g.. people thought they were done already but the new arch required some changes which they didn't realize etc
19:08:44 <langdon> ok.. higher level
19:09:52 <langdon> everything repo just as it has always been, adding a new repo with modules for alternative versions, which is purely optional. We are enabling it for server (but it can be disabled).. on workstation, we don't recommend it because it doesn't play well with packagekit / gnome-software yet
19:10:27 <langdon> you can use it in a container (after installing dnf assuming you use fedora-minimal which only has microdnf)
19:10:28 <dgilmore> langdon: but on server cockpit updates use packagekit right?
19:10:41 <mattdm> #info "Everything" repo as it has always been, but we now have a new "Modular" repo which has modules.
19:11:00 <mattdm> #info This new repo is optional, and enabled by default on Server
19:11:04 <langdon> dgilmore: ughh.. i suppose they do.. so, that is icky
19:11:06 <dgilmore> langdon: and microdnf will not support modules ever?
19:11:31 <langdon> dgilmore: microdnf, packagekit, etc should all support modules as soon as the code is pushed from dnf -> libdnf
19:11:33 <mattdm> #info Not recommended on Workstation because compatiblity with PackageKit is not done
19:11:48 <dgilmore> langdon: okay :)
19:11:50 <langdon> which is the "problem" .. (in quotes because it isn't a bug just work not done yet)
19:12:00 <mattdm> #info Also same problem with microdnf and Cockpit compatibility
19:12:08 <mattdm> #info not a bug, per se, just work not done yet
19:12:15 <dgilmore> langdon: there was arequest filed with fesco to get an exception for it
19:12:16 <mattdm> langdon: is there a bug for this, though?
19:12:24 <mattdm> as in, a bugzilla entry?
19:12:28 <dgilmore> langdon: though it was only basic support
19:12:31 <mattdm> where not all bugs are _bugs_ :)
19:12:35 <langdon> dgilmore: excpetion for the bodhu update?
19:12:44 <dgilmore> not supporting installing modules but to support updating things correctly
19:12:53 <langdon> mattdm: rfe for libdnf? or bug re: the bodhi update above?
19:13:00 <mattdm> langdon: rfe for libdnf
19:13:16 <dgilmore> langdon: https://pagure.io/fesco/issue/1883 to allow a rebase of dnf/libdnf
19:13:28 <langdon> dgilmore: gotcha..
19:13:33 <mattdm> the bodhi update I'm not worried about because it looks like there's already an accepted freeze exception
19:13:42 <langdon> mattdm: /me searches.. i thought i had seen one
19:14:42 <mattdm> langdon: and, does that fesco issue cover *addressing* the issue, or is that *working around it until propre fix can land*?
19:14:49 <mattdm> *proper
19:15:03 <langdon> mattdm: actually there are a bunch of them.. /me looks for tracker bug
19:15:21 <langdon> mattdm: i think we are talking about 2 different things
19:15:40 <langdon> ohh.. i am talking about a different thing
19:15:44 <mattdm> ha
19:15:55 <mattdm> I may be talking about as many as THREE different things
19:16:02 <langdon> i thought that was a fesco exception for the bodhi update..
19:16:26 <mattdm> that's not needed -- that's a QA-process freeze exception
19:16:37 <mattdm> and adam says there already is one
19:16:45 <langdon> so .. we have 1) dnf update to make anything work related to defaults that is in bodhi now 2) we have to modify libdnf to support modules so all pkgmgmt apps that use it get modular support
19:16:58 <langdon> mattdm: right..
19:17:03 <langdon> thats 1) ^^
19:17:22 <langdon> 2) i believe the plan is "real support" in libdnf
19:17:43 <mattdm> and is 2 the fesco ticket, or just a stopgap?
19:17:55 <langdon> the reason it isn't there already is because they only wanted to write it once in c .. so i don't think it would make sense to do a workaround version
19:17:55 <dgilmore> mattdm: I think it is a stopgap
19:18:10 <langdon> well.. it is "partial support"
19:18:17 <dgilmore> right
19:18:19 <langdon> so not a hack.. but not all modular support
19:18:36 <dgilmore> it will cause thiings to not be broken if repos with modules are enabled
19:18:59 <dgilmore> but it will not support installing modules, that will take more work
19:19:18 <langdon> i am hoping for another exception after that one that will be full support.. but I don't have an eta on that
19:19:24 <mattdm> so you can enable modules with dnf on, say, Workstation, and updates from Software will work?
19:19:25 <dgilmore> so stopgap in teh sense that it is partial
19:19:39 <dgilmore> mattdm: thats how I read the FESCo issue
19:20:00 <langdon> mattdm: right.. basically pkg mgrs won't honor the "virtual repo-ness" of the modular content without the change
19:20:11 <dgilmore> you will have to use dnf to install and select modules
19:20:18 <langdon> so they will walk nvrs just like they always do.. which, with modules, you aren''t supposed to
19:20:25 <langdon> dgilmore: correct
19:20:32 <mattdm> #info there is an update in the near future (shortly after F28 GA) which will make cockpit/microdnf/PackageKit *not cause horrible trouble* when modules are enabled, but not to allow those tools to work with modules directly.
19:20:46 <mattdm> #info We're expecting full support in the F29 timeframe
19:21:11 <mattdm> and I'd really like a bugzilla bug for that so I can propose it as an F29 blocker :)
19:21:52 * langdon hopes before f29
19:22:02 <langdon> mattdm: i can go back to digging
19:22:05 <dgilmore> langdon: how many modules are we expecting at GA?
19:22:22 <mattdm> langdon: after meeting, maybe?
19:22:28 <mattdm> dgilmore: good next question :)
19:22:36 <langdon> dgilmore: we have ~15 now
19:22:49 <langdon> i am not sure how many more i can get to land before GA
19:23:10 <mattdm> #info we have about 15 modules now -- hoping for more at GA but number uncertain
19:23:21 <langdon> obviously.. some are dupes.. or alt streams.. and a couple are "enables" .. e.g. go-ecosystem
19:23:27 <langdon> *enablers
19:23:28 <dgilmore> langdon: got to start with 1 :D
19:23:42 <mattdm> "dupes" meaning "multiple streams are available"? Because that seems like a feature, not a bug :)
19:24:03 <dgilmore> like nodejs 6 7 and 8?
19:24:10 <dgilmore> seems like a feature to me
19:24:25 <langdon> mattdm: ohh yes.. definitely a feature.. but technically "not countable as modules"... as it is the same module, different streams.. but i think the number you really want is the 15 not the 10 or whatever
19:25:16 <langdon> so.. the real q/a is "how many streams will we have" which is "15" .. but that gets euphemistically turned in to "how many modules do we have" .. i believ
19:25:56 <langdon> make sense?
19:26:19 <mattdm> yeah. may make sense to count differently somehow but I'll have to think about it
19:26:20 <dgilmore> yep
19:26:33 <bexelbie> Can we do outreach to the maintainers in the recent mag article to see if that gets us twomore?
19:26:56 <langdon> bexelbie: which mag article? did i miss something?
19:27:20 <langdon> i have queries from people about 1-2 .. that i need to figure out and respond to..
19:27:33 <mattdm> there was an article where the authors say "this is in fedora, but if you want the newer version, try copr"
19:27:42 <mattdm> that seems like an ideal case for modular streams
19:27:42 <bexelbie> Article that referenced newer versions in COPR. I’m on a phone so it isn’t handy.
19:27:54 <langdon> bexelbie: ahh .. yes.. i saw that
19:27:58 <mattdm> https://fedoramagazine.org/4-try-copr-april-2018/
19:28:00 <langdon> i had grand plans to reply..
19:28:02 <langdon> but life
19:28:16 <mattdm> life? what's that? :)
19:28:36 <mattdm> anyway, we should wrap up so dperpeet has time and pbrobinson if he shows...
19:28:43 <langdon> ack
19:28:46 <mattdm> it'd be awesome if you could ping those authors
19:29:01 <mattdm> and quickly though, what's the chicken and egg problem you referenced?
19:29:13 <mattdm> and can we cook it into an omlet somehow?
19:30:02 <langdon> mattdm: hmmm...
19:30:41 <langdon> ohh..
19:31:10 <langdon> so.. someone did a module for f27, it worked, everything was fine.. no errors on rebuilds because it doesn't ref f28.. but.. it won't work on f28 .. so they don't know they need to do an update..
19:31:35 <langdon> i filed some bugs.. but that was a problem for awhile because we didn't know how we were doing bugs either..
19:32:13 <langdon> so.. now, with stream expansion and a settled arch this shouldn't happen in the future.. but .. tracking back those older modules has been a bit confusing both for the authors and for me
19:32:15 <dgilmore> can we do bugs now?
19:32:18 <langdon> yes
19:32:33 <langdon> rather.. people know to expect bugs
19:33:03 * robyduck here
19:33:09 <langdon> and.. not to mention.. XYZ version was interesting in f27 because it hadn't landed yet.. but isn't in f28 because it is mainline in f28
19:33:48 <langdon> so.. i also need to see if the maintainer wants to do X-1 or X+1 version in f28 ..
19:33:56 <mattdm> *nod*
19:34:00 <mattdm> okay, so any last words?
19:34:05 <langdon> so.. just a situation that is made for confusion in all parties
19:34:20 <langdon> sounds dire..
19:34:23 <mattdm> can we fix that?
19:34:23 <langdon> but .. im good
19:34:35 <mattdm> okay. let the rest of us know if you need anything!
19:34:41 <mattdm> dperpeet: you ready?
19:34:44 <langdon> mattdm: slowly? but, honestly, i wouldn't stress it as this is a one off
19:34:46 <dperpeet> yes!
19:34:55 <mattdm> langdon: cool. more things to not stress are good
19:35:14 <mattdm> #topic Fedora CI Objective from Dominik
19:35:30 <dperpeet> Hey... so when we spoke last, I think I promised more progress on the "test all packages" aspect of CI. Well, good news. :)
19:35:43 <dperpeet> While that testing isn't in production yet, it's looking good.
19:35:54 <dperpeet> #link https://jenkins-continuous-infra.apps.ci.centos.org/view/Fedora%20All%20Packages%20Pipeline/
19:36:14 <dperpeet> The CI Pipeline folks (thank you jbieren) and CentOS CI (thank you bstinson) have put in serious work to get us this far. We run tests on koji-builds, if there are tests present in dist-git for that package. As we speak, jbieren is finishing up the pull request testing that will use koji for scratch builds, giving us test feedback on pull requests in Pagure.
19:37:05 <dperpeet> If you look at the fedora 28 pipeline in the above link, you can see passing and failing tests (green vs yellow)
19:37:08 <dperpeet> #link https://jenkins-continuous-infra.apps.ci.centos.org/view/Fedora%20All%20Packages%20Pipeline/job/upstream-fedora-f28-pipeline/
19:37:18 <mattdm> #info "all packages" testing not in production yet, but looking good
19:37:37 <mattdm> that is a LOVELY hostname that I can *totally* remember off the top of my head :)
19:37:44 * mattdm bookmarks
19:37:50 <dperpeet> :D
19:37:58 <dperpeet> well, I hope you don't have to
19:38:05 <dperpeet> we're also working on a monitoring system
19:38:11 <dperpeet> a bot to monitor the pipeline
19:38:15 <dperpeet> a bot overseer, so to speak
19:38:31 <dperpeet> which will probably post a dashboard like page to the wiki
19:38:45 <mattdm> so as a packager, my interaction will not be through the jenkins ui?
19:39:00 <bexelbie> Not wiki please :)
19:39:01 <dperpeet> and once this is in production, we'll add the info to https://fedoraproject.org/wiki/CI
19:39:04 <dgilmore> Problem accessing /view/Fedora%20All%20Packages%20Pipeline/job/upstream-fedora-f28-pipeline. Reason:
19:39:07 <dgilmore> Not Found
19:39:18 <dperpeet> bexelbie, I feel bad about not migrating to docs :/
19:39:25 <bexelbie> Will status post to pagure?
19:39:47 <bexelbie> Neither is good for a dashboard ...
19:39:52 <dperpeet> the test results themselves will go to Pagure
19:40:03 <dperpeet> and other notifications that developers opt into
19:40:05 <langdon> dperpeet: stupid q.. couldn't you just make a bit.ly in the meantime?
19:40:05 <dperpeet> e.g. e-mail
19:40:14 <dperpeet> langdon, we probably should
19:40:20 <dperpeet> the link not working is probably due to spaces
19:40:26 <bexelbie> Will there be a “stoplight” on the screen?
19:40:29 * langdon owns fed-mod.org for this problem :)
19:40:31 <mattdm> wiki isn't the _worst_ thing as an interim place to dump results
19:40:49 * dgilmore watches bexelbie cringe
19:40:54 <dperpeet> bexelbie, "stoplight" as in red / yellow / green status?
19:41:38 <bexelbie> Yes
19:41:45 <dperpeet> yeah
19:41:50 * langdon prefers a cowbell
19:42:05 <dperpeet> langdon can of course fork our code and produce a page with a cowbell
19:42:07 <dperpeet> :)
19:42:09 <dperpeet> open source wins
19:42:29 * jwb is actually here today but just hasn't had much to say
19:42:30 <dperpeet> I will make sure there is a better link for the production page :)
19:42:40 <mattdm> jwb: but have something to say about cowbell now?
19:43:04 <mattdm> anyway, not to derail over the link. It's cool that this is working. :)
19:43:08 <dperpeet> so that side of CI is going well, but there have been some concerns about gating
19:43:20 <dperpeet> In the last two weeks there have been discussions around disabling gating. At this point I ask that we don't conflate the issues of gating in principle and certain tests and or workflows being unreliable. Disabling gating would in my opinion send a wrong message, catering to the "we're in a hurry, let's properly test it later" mindset which I sadly know very well... :)
19:43:29 <dperpeet> Let's address the issues we can and try to stand by the vision of testing and gating.
19:43:56 <jwb> i agree dperpeet
19:44:22 <mattdm> #info Dominik asks us to not mix up the issue of gating in principle with problems with specific tests/workflow problems
19:44:38 <dperpeet> thank you for the concise summary :)
19:44:42 <bexelbie> I agree. Let’s fix the tests/system not ignore them.
19:44:46 <mattdm> #info Thi would send a "we're in a hurry, let's properly test it later" message
19:44:58 <jwb> irony
19:44:59 <dgilmore> dperpeet: indeed, we should enable gating, and make some less than extremely reliable tests non gating. but keep gating enabled
19:45:07 <jwb> #undo
19:45:07 <zodbot> Removing item from minutes: INFO by mattdm at 19:44:46 : Thi would send a "we're in a hurry, let's properly test it later" message
19:45:18 <jwb> #info This would send a "we're in a hurry, let's properly test it later" message
19:45:23 <mattdm> thanks jwb :)
19:45:33 <jwb> I HAVE CONTRIBUTED TODAY!
19:45:37 <dperpeet> :)
19:45:41 <dperpeet> jwb++
19:45:41 <zodbot> dperpeet: Karma for jwboyer changed to 3 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
19:45:44 <robyduck> yay :)
19:46:09 <dperpeet> so moving forward we want to make sure testing is enabled + visible for all fedora 28 packages
19:46:15 <dperpeet> and that we get pull request testing working properly
19:46:26 <dperpeet> i.e. at the minimum, feedback on Pagure pull requests against dist-git
19:46:29 <mattdm> nice -- that's super-exciting
19:46:36 <mattdm> what's the liklely timeframe for that?
19:46:58 <dperpeet> <jbieren> dperpeet, it triggers on commits. i am switching it to PRs literally as we speak
19:47:19 <dgilmore> dperpeet++
19:47:19 <zodbot> dgilmore: Karma for dperpeet changed to 3 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
19:47:30 <dperpeet> realistically I hope for 2-4 weeks until it's reliable
19:47:31 <mattdm> jbieren++
19:47:31 <zodbot> mattdm: Karma for jbieren changed to 1 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
19:47:38 <mattdm> dperpeet++
19:47:42 <dgilmore> jbieren++ also
19:47:42 <zodbot> dgilmore: Karma for jbieren changed to 2 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
19:47:50 <mattdm> so basically right around the F28 release
19:48:00 <dperpeet> yes, I think by then we should be good
19:48:05 <mattdm> Do we have documentation lined up for the casual Fedora packager?
19:48:23 <dperpeet> we're keeping the wiki up to date
19:48:33 <dperpeet> but I think we'll want at least another magazine article with a demo
19:48:36 <dperpeet> once it works
19:48:47 <mattdm> #link https://fedoraproject.org/wiki/CI
19:49:00 <mattdm> dperpeet: yeah, a magazine article would be great
19:49:22 <dperpeet> one of the biggest blockers there is that due to technical issues on the Fedora Infra side, non-"packagers" can't open pull requests in Pagure
19:49:32 <bexelbie> And then docs :) maybe with the new version of the packaging guidelines.
19:49:55 <mattdm> bexelbie yes.
19:50:01 <robyduck> bexelbie++
19:50:13 <dperpeet> Fedora Infra is aware of the issue and working on it
19:50:15 <dperpeet> #link https://pagure.io/fedora-infrastructure/issue/6361
19:50:16 <dgilmore> bexelbie: docs.....?????
19:50:17 <mattdm> dperpeet: yeah, we need to figure out a resolution there
19:50:33 * mattdm looks at clock -- we should probably move to budget
19:50:46 <mattdm> dperpeet this is looking really great
19:50:51 <bexelbie> Said in the voice of Shatner dgilmore ;)
19:50:57 <mattdm> as I said to langdon, any last words?
19:51:07 <dperpeet> thanks... and I promise if you point me at the right place once this works somewhat, I'll make sure we contribute docs to that right place :)
19:51:15 <dgilmore> bexelbie: :)
19:51:20 <dperpeet> #info Dominik promises docs
19:51:21 <bexelbie> Consider yourself pointed.
19:51:26 <dperpeet> done.
19:51:42 <mattdm> okay, cool
19:51:48 <mattdm> #topic FY19 budget
19:52:17 <mattdm> #info reminder: Our fiscal year starts in March, so we are in FY19 *already*
19:52:27 <langdon> details
19:52:37 <mattdm> bexelbie, got anything to say? :)
19:53:05 <mattdm> I think we can go ahead and make https://pagure.io/Fedora-Council/tickets/issue/200 public
19:53:05 <bexelbie> I’d like is to consider unblocking the budget ticket and passing it if we have consensus.
19:53:34 <mattdm> I saw no opposition and it's been open for a week
19:53:39 <bexelbie> We do not have our final budget from RH however o encourage us to move forward assuming we will receive a similar allocation.
19:53:52 <mattdm> sorry, open for *two* weeks
19:54:05 <bexelbie> I’ll come back if there is a variance
19:54:08 <mattdm> It'd be nice to have a +1 from nb
19:54:32 <mattdm> I suggest we make the ticket public and call for consensus by meeting next week
19:54:42 <bexelbie> +1
19:54:46 <mattdm> that'll allow some public comment and also give nick time
19:56:20 <robyduck> bexelbie: any idea when we will have the final budget? we are going towards may now...
19:56:38 <bexelbie> I expect end of the month. Like last year.
19:56:41 <mattdm> Since pbrobinson couldn't make it, let's use the beginning of next week's Open Floor meeting for IoT
19:56:56 <mattdm> #info Final budget numbers from RH expected by end of month. We hope. No promises.
19:57:04 <mattdm> Also, a reminder:
19:57:04 <dgilmore> +1
19:57:08 <bexelbie> This is why I’m trying to move is forward. I’d like to unblock us and set the idea that we make plans and adjust.
19:57:22 <mattdm> #info Next week's meeting slot is 3 hours earlier than this one
19:57:39 <dgilmore> mattdm: will it be video or IRC?
19:57:45 <mattdm> irc
19:57:50 <dgilmore> just checking
19:58:01 <mattdm> because there was pretty strong feeling that one video meeting a month was all we could handle :)
19:58:35 <mattdm> okay, I need to go on to another meeting
19:58:42 <dgilmore> cheers all
19:58:48 <mattdm> bexelbie can you note the thing about one more week in the budget ticket?
19:59:03 <bexelbie> Not easily at the moment :(
19:59:10 * bexelbie is on a phone
19:59:20 <mattdm> ok I will
19:59:24 <bexelbie> Thank you.
19:59:42 <mattdm> thanks everyone#
19:59:44 <mattdm> thanks everyone!
19:59:48 <mattdm> #endmeeting