19:02:01 <mattdm> #startmeeting Council (2018-08-29)
19:02:01 <zodbot> Meeting started Wed Aug 29 19:02:01 2018 UTC.
19:02:01 <zodbot> This meeting is logged and archived in a public location.
19:02:01 <zodbot> The chair is mattdm. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:02:01 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:02:01 <zodbot> The meeting name has been set to 'council_(2018-08-29)'
19:02:03 <mattdm> #meetingname council
19:02:03 <zodbot> The meeting name has been set to 'council'
19:02:05 <mattdm> #chair mattdm jkurik jwb langdon robyduck bexelbie dperpeet Amita dgilmore pbrobinson tyll bcotton
19:02:05 <zodbot> Current chairs: Amita bcotton bexelbie dgilmore dperpeet jkurik jwb langdon mattdm pbrobinson robyduck tyll
19:02:07 <mattdm> #topic Introductions, Welcomes
19:02:07 <bexelbie> .hello ex
19:02:09 <mattdm> hi everyone!
19:02:09 <zodbot> bexelbie: Sorry, but you don't exist
19:02:09 <dgilmore> hola
19:02:10 <bexelbie> .hello bex
19:02:12 <zodbot> bexelbie: bex 'Brian (bex) Exelbierd' <bexelbie@redhat.com>
19:02:14 * pbrobinson o/
19:02:15 <mattdm> this is an objectives update meeting
19:02:17 <langdon> .hello2
19:02:18 <zodbot> langdon: langdon 'Langdon White' <langdon@redhat.com>
19:02:23 <mattdm> so I'm glad to see pbrobinson and langdon here :)
19:02:33 <langdon> im even prepared
19:02:35 <bcotton> .hello2
19:02:36 <zodbot> bcotton: bcotton 'Ben Cotton' <bcotton@redhat.com>
19:02:42 <mattdm> langdon: nice :)
19:02:47 <tyll> .hello till
19:02:48 <zodbot> tyll: till 'Till Maas' <opensource@till.name>
19:02:52 <mattdm> I think dperpeet still can't make this time.
19:03:04 <mattdm> which brings us to the first topic :)
19:03:16 <mattdm> which I thinkw e'll just jump into because it looks like most of us are here already
19:03:20 <mattdm> #topic Meeting time!
19:03:20 <tyll> He is from Germany, too, isn't he? It is 9pm here
19:03:27 <mattdm> tyll: yeah :(
19:03:31 <mattdm> so, bcotton ....
19:03:59 * bexelbie is amazed at being at this meeting while it is daylight outside (/me is traveling)
19:04:01 <bcotton> #info the general consensus is that 1400 UTC on Wednesdays is the best time
19:04:10 <pbrobinson> mattdm: should I be scared?
19:04:20 <bcotton> #info meetings will be held at 1400 UTC on Wednesdays in #fedora-meeting-1 beginning next week
19:04:26 <mattdm> pbrobinson: probably :)
19:04:28 <bcotton> #action bcotton to send an email to council list with new meeting times
19:04:34 <bcotton> EOF
19:04:46 <mattdm> #info this will include _all_ of the Council meetings, not just the Objectives updates
19:04:59 <mattdm> This is the one time that *almost* works for everyone
19:05:04 <mattdm> sorry tyll and jwb :)
19:05:38 <mattdm> In seriousness, I know it's basically impossible to get a time that works for everyone. we'll try to make it so major decisions have tickets and can get input from everyone
19:05:45 <langdon> does it follow the daylight savings?
19:05:49 <mattdm> langdon: yes.
19:05:58 <mattdm> at least, that's what we did last several years
19:06:01 <langdon> mattdm: where? :)
19:06:13 <mattdm> langdon: North Korea
19:06:15 <mattdm> wait, no
19:06:42 <mattdm> Normally, we've followed US DST
19:06:56 <mattdm> it's only a few weeks of non-overlap, right?
19:06:58 <langdon> well eastern :)
19:07:05 * mattdm wonders when the world will end this insanity
19:07:21 <mattdm> fortunately at least all us dst ends and begins on the same day
19:07:30 <bcotton> mattdm: if only the world has listened to Indiana
19:07:31 * pbrobinson doesn't think we'll solve DST here so can we move on?
19:07:33 <langdon> it does?
19:07:36 <mattdm> pbrobinson: fair
19:07:42 <mattdm> okay to business
19:07:56 <mattdm> let's let peter go first since it's evening there?
19:08:03 <mattdm> langdon yell if you object.
19:08:10 <langdon> no.. thats good
19:08:16 <mattdm> #topic Objective Update: Internet of Things
19:08:18 <tyll> I can usually make it at leastevery other Wednesday and if you take long I might be able to join
19:08:29 <pbrobinson> we're moving forward with the base image
19:08:37 <pbrobinson> it's almost looking good
19:08:50 <pbrobinson> and people/companies are starting to use it
19:09:05 <langdon> \o/
19:09:23 <bexelbie> woot!
19:09:24 <mattdm> "almost"?
19:09:36 <mattdm> My ears are finely tuned for words like that :)
19:09:42 <pbrobinson> we should have ARMv7 RSN, all the bits have landed upstream now so I'm hoping tomorrow or Friday I'll be able to do a full end to end test of that
19:10:01 <mattdm> awesome.
19:10:06 <tyll> what is RSN?
19:10:08 <mattdm> how are things going for bringing more people in?
19:10:12 <dgilmore> pbrobinson: any challenges we can help with
19:10:16 <mattdm> tyll: "real soon now"
19:10:17 <pbrobinson> mattdm: it's not perfect as I wanted it, I'm going to need to reach out to the sytemd people about some of the lockdown features
19:10:32 <bexelbie> will we be doing any end-to-end testing in an automated way?
19:10:33 <tyll> mattdm: thx
19:10:45 <mattdm> tyll: it means "I thought it would be done by now but it isn't so ... sooon, but I can't really promise when and it might again take longer than I hope"
19:10:52 <mattdm> which is a lot of weight for three little letters :)
19:11:04 <pbrobinson> dgilmore: off the top of my head not sure but I will be reaching out, the anaconda stack bits have landed and I expect a build today or so I'm told
19:11:39 <pbrobinson> with those I can finish end to end testing for armv7, I'm sure there will be bugs there
19:12:14 <pbrobinson> other things are the PRD that I need to review in the next few days so bcotton and I can make that official
19:12:31 <dgilmore> pbrobinson: okay, feel free to let me know if you need anything
19:12:33 <pbrobinson> and I've seen no updates on details about the website and other design bits
19:12:36 <mattdm> cool, yeah.
19:13:00 <mattdm> pbrobinson: yeah, with robyduck on hiatus we're kind of in a bind with websites
19:13:01 <pbrobinson> I'd like to have a general iot.fp.o landing site but I'm not sure the design team have registered my request there
19:13:15 <mattdm> pbrobinson: is there a ticket?
19:13:36 <pbrobinson> mattdm: yes, let me dig it out
19:14:09 <pbrobinson> https://pagure.io/design/issue/596
19:14:19 <mattdm> pbrobinson: cool. I'll help shepard it through
19:14:24 <pbrobinson> https://pagure.io/fedora-websites/issue/788
19:14:52 <pbrobinson> https://pagure.io/design/issue/586
19:15:06 <mattdm> thanks
19:15:18 <mattdm> Let me see what I can do there
19:15:21 <pbrobinson> thanks
19:15:46 <pbrobinson> and from there we're awaiting the multiarch support for the container build pipeline
19:16:11 <dgilmore> pbrobinson: is that on track?
19:16:14 <mattdm> Cool. I'm reviewing https://fedoraproject.org/wiki/Objectives/Fedora_IoT and things seem basically on track with the schedule there
19:16:21 <mattdm> presuming f30 edition
19:16:22 <pbrobinson> I managed to wrangle another two aarch64 servers which should be on their way inbound to the DC and the infra team knows of the details
19:16:52 <pbrobinson> yes I think F-30 for edition should be achievable
19:16:58 <mattdm> neat. are those for general build or for other things?
19:17:33 <pbrobinson> we already have two for copr/cloud etc which with luck should be going live  RSN
19:18:01 <pbrobinson> these two I believe they'll put in core infra for build/container pipeline and other related stuff
19:18:45 <pbrobinson> once we have container pipeline we can put some of the utils into a "IoT dev container" and slim the base OS right down.
19:19:23 <pbrobinson> one of the users is also writing an article (which I think will be actually 2-3) for Fedora magazine too which should help spin up interest
19:19:42 <mattdm> pbrobinson: awesome!
19:19:53 <mattdm> that in itself is a success metric ++++
19:20:01 <pbrobinson> agreed
19:20:22 <pbrobinson> actually writing this all down actually makes me realise I've achieved quite a bit since the last report :)
19:20:28 <langdon> ha
19:20:45 <dgilmore> pbrobinson: :) reflection is good
19:20:54 <bexelbie> +1 to "yes you have" :P
19:21:09 <mattdm> pbrobinson++
19:21:16 <pbrobinson> we're also having quite a regular and reasonable attendance to the meetings, and a lot of useful flock/devconf discussions
19:21:46 <pbrobinson> which has some quite exciting things in the pipeline, some while will be a little way out but none the less I think we're getting people thinking about stuff :)
19:22:34 <mattdm> this is nice to hear. anything else?
19:22:50 <pbrobinson> nothing from me unless anyone has any questions
19:22:56 <bexelbie> will we be doing any end-to-end testing in an automated way?
19:23:08 <bexelbie> if that was answered earlier, I missed it
19:23:16 <pbrobinson> yes! My plan is to do as much as possible
19:23:28 <pbrobinson> we're just starting to investigate that
19:23:31 <bexelbie> are there any non-IoT blockers there?
19:23:33 <bexelbie> ahh, ok
19:23:44 <bexelbie> wanted to illuminate anything there as I know this is a big topic for Fedora
19:23:51 <pbrobinson> and yes, I'm waiting for some changes and improvements in the CI pipeline there
19:24:18 <pbrobinson> and also an answer to what is the replacement to autocloud, or how we'll end up testing those sorts of things
19:24:36 <pbrobinson> I've not had a concrete answer to any of those questions
19:25:18 <langdon> arggh... need to step away for 5m
19:25:24 <pbrobinson> I had a call with mohan and bstinson over CI bits the other day and bstinson has some improvements there which will help some things (in general too)
19:25:30 <dgilmore> does the CI pipeline support non x86?
19:25:30 <mattdm> pbrobinson: maybe this is something that could be an infra / qa / iot / ci hackfest
19:25:37 <dgilmore> last I looked it didn't
19:25:52 <bexelbie> I wonder if we should ask FESCo (or whomever owns this) to update us and we can help amplify their status?
19:25:52 <dgilmore> but what mattdm said
19:25:54 <pbrobinson> dgilmore: I thought it did now
19:26:27 <dgilmore> pbrobinson: hopefully we can have that clarified
19:26:28 <bexelbie> as my understanding is this is different from the CI Objective ..
19:26:34 <langdon> back
19:26:43 <pbrobinson> I know there's a Coreos tool that might be useful as a replacement to autocloud
19:27:07 <pbrobinson> basically I plan to test as much of the base image as possible in virt
19:27:55 <pbrobinson> then work out how we can test the delta of that on actual HW remotely in an automated fashion
19:28:33 <pbrobinson> so to answer bexelbie's question, yes as much as possible automated, a lot of the issues there is no different to that of coreos etc
19:28:50 <mattdm> bexelbie: yeah, different from CI Objective but probably doable in CentOS CI
19:29:04 <bexelbie> pbrobinson, ty
19:29:16 <mattdm> I'm definitely feeling like a hackfest around this may be in order if we can get the right people from the aforementioned groups and ...
19:29:20 <pbrobinson> I know they do some rpm-ostree testing in CI, I was going to look at that when I get a moment
19:29:21 <bexelbie> mattdm, I am wondeirng if we need some kind of reporting here - I hear lots of side conversations but see little output of information
19:29:26 <mattdm> ... get something that serves both IoT and CoreOS out of it
19:29:26 <bexelbie> it may be that I am missing it though
19:29:48 <mattdm> bexelbie: on CI Objective, or CI in general?
19:29:52 <pbrobinson> A wiki page might be useful to log requirements and issues with current options
19:30:28 <bexelbie> mattdm, the bits of CI that are being worked on for things like this end-to-end testing that aren't part of hte CI objective and so won't be talked about today
19:30:35 <dgilmore> +1 to a hackfest on this. in the UK would work well.
19:30:56 <mattdm> bexelbie: we could do an update from that team -- bstinson and co
19:31:18 <mattdm> well, there is https://fedoraproject.org/wiki/CI
19:31:34 <pbrobinson> when I refer to end-to-end testing I fully see this as a number of different components
19:32:14 <bexelbie> mattdm, that would work well ... as it would also share out who "and co" is and maybe get them some help
19:33:24 <mattdm> #action mattdm to ask bstinson for presentation on CentOS CI and Fedora
19:33:53 * mattdm hopes maybe bstinson will see this in irc and he won't even have to email :)
19:34:09 <mattdm> pbrobinson: are you listing components now are are we ready to move on? :)
19:35:20 <pbrobinson> move on
19:35:39 <pbrobinson> just copy me in on any useful discussions here so I don't miss out
19:35:43 * bstinson peeks into this meeting
19:36:00 <mattdm> pbrobinson: sounds good
19:36:03 <pbrobinson> bstinson: I'd run ;-)
19:36:25 <dgilmore> bstinson++
19:36:31 <mattdm> bstinson: see above. can you do a presentation for us on CentOS CI?
19:36:48 <mattdm> Next open slot is Sept 19
19:37:00 <mattdm> (Or we could reschedule for a different Wednesday.)
19:37:07 <mattdm> I can get you details offline.
19:37:27 <bstinson> yep, i can put something together for then
19:37:48 <langdon> probably worth outlining what we want to know
19:37:56 <langdon> status? usage? usage by fedora?
19:38:02 <langdon> something else?
19:38:09 <langdon> i am not 100% clear myself
19:38:57 <mattdm> 1. What's availble 2. How subprojects and others can use it. 3. Who to talk to. 4. Future plans. 5. Help you could use
19:39:00 <mattdm> That kind of stuff
19:39:08 <mattdm> but anyway now, let's switch to ...
19:39:28 <mattdm> #topic Modularity Objective
19:39:44 <langdon> ok.. so this is a bunch of bullets.. but let's see how it plays
19:39:58 <mattdm> ok :)
19:40:04 <langdon> on the good side:
19:40:06 <langdon> f29: 32 uniq modules, 45 including alt streams
19:40:12 <langdon> f28: 27 uniq modules, 33 including alt streams
19:40:29 <mattdm> that's amazing.
19:40:35 <langdon> #link Jonathon Dieter wrote a nice article derived from him live implementing a module at the experts panel at Flock: https://www.jdieter.net/posts/2018/08/11/building-a-module/  Jonathon Dieter wrote a nice article derived from him live implementing a module at the experts panel at Flock
19:40:39 <mattdm> is there an easy way to get that number at a glance?
19:40:41 <langdon> #undo
19:40:41 <zodbot> Removing item from minutes: <MeetBot.items.Link object at 0x7f56bb71d490>
19:41:12 <langdon> dnf module list > out.txt, edit for junk lines, cut -d ' ' -f 1 out.txt | uniq | wc -l
19:41:21 <langdon> and cut -d ' ' -f 1 out.txt | wc -l
19:41:35 <mattdm> on a system running that version, right?
19:41:39 <langdon> there is a space between those '' which i can't tell if it came through
19:41:41 <langdon> mattdm: right
19:41:48 <mattdm> this is what I've been doing, with a docker image for f28
19:41:57 <mattdm> i just wondered if there were something less annoying :)
19:42:03 <bexelbie> this smells like a great dashboard number :D
19:42:21 <langdon> ha
19:42:34 <langdon> mattdm: exactly how i did it
19:42:41 <langdon> and for f29
19:42:51 <mattdm> ok :)
19:42:55 <mattdm> bexelbie: yes
19:42:55 <langdon> podman run -it --rm fedora:29 /bin/bash
19:43:01 <langdon> ok.. next?
19:43:05 <mattdm> yes next
19:43:08 <langdon> #link https://www.jdieter.net/posts/2018/08/11/building-a-module/ Jonathon Dieter wrote a nice article derived from him live implementing a module at the experts panel at Flock
19:43:33 <langdon> I CANNOT type today
19:43:44 <langdon> #link https://gitlab.com/gitlab-org/gitlab-development-kit/blob/master/doc/prepare.md#fedora Neal Gampa mentioned the use of modules in the installation instructions for developing gitlab
19:43:56 <bexelbie> Gompa, I think
19:44:09 <langdon> #link https://docs.fedoraproject.org/en-US/modularity/making-modules/defining-modules/ Nils Philippsen updated the how to for making a modulemd to use fedmod
19:44:15 <langdon> bexelbie: oh really?
19:45:14 <langdon> you are correct per the zodbot
19:45:39 <langdon> #link https://gitlab.com/gitlab-org/gitlab-development-kit/blob/master/doc/prepare.md#fedora Neal Gompa mentioned the use of modules in the installation instructions for developing gitlab
19:46:06 <langdon> ok.. thats the good.. thought the neal part was particularly interesting.. someone recommending using a module in installation instructions
19:46:33 <langdon> questions? or bad stuff now?
19:46:47 <mattdm> I guess bad stuff :)
19:46:50 <langdon> ha
19:46:52 <langdon> ok.
19:46:52 <bcotton> make us sad, langdon
19:46:54 <langdon> F29 branching revealed a number of issues in tools & processes related to modules
19:47:03 <langdon> ohh.. it has been WAY worse
19:47:07 <langdon> this is pretty mild
19:47:14 <langdon> dnf still has a number of open issues (see blockers list for f29)...
19:47:25 <langdon> 1) dnf doesn't follow default profile for an enabled non-default stream
19:47:34 <langdon> 2) dnf doesn't record modular metadata in a local database
19:47:45 <langdon> 3) DNF update fails with "cannot install the best update candidate for package"
19:48:17 <dgilmore> langdon: so its all dnf?
19:48:29 <langdon> 2 is particularly yucky.. it is basically that the module metadata is not being recorded like rpm metadata.. so.. dnf clean will wipe it out.. there by making the installed/enabled modules misbehave if you are offline
19:48:33 <tyll> In the FESCo ticket it said 2) might take long to fix
19:48:58 <langdon> tyll: yeah.. we had a discussion about it today.. there may be a shorter workaround
19:48:59 <mattdm> It seems like... kind of a big oversight
19:49:26 <tyll> 2) is also a problem when temporarily installing modules from a testing repository according to the FESCo ticket
19:49:26 <langdon> its the "virtual repo" euphemism.. which makes you think you don't need the metadata that "much" .. in my opinion
19:50:14 <langdon> tyll: AIUI .. kinda.. it is more about the wiping out the testing/renabling normal.. basically you dump metadata you need to recover and things get messy
19:50:56 <langdon> just imagine the rpm db being corrupted .. but it only happens when a) no cache b) offline (so no cache)
19:51:11 <langdon> sorry not a or b, a AND b
19:51:54 <mattdm> We've only got 8 minutes left so let's not get too big in the technical details
19:52:00 <langdon> basically when this scenario occurs, dnf treats modular rpms as ursine/bear/bare rpms... which starts causing ptroblems
19:52:08 <langdon> yeah.. thats all i got though
19:52:10 <langdon> :)
19:52:34 <bexelbie> I'd like to ask one question if we have time before the next objective
19:52:53 <langdon> bexelbie: i promise to give more than one answer
19:53:16 <bexelbie> langdon, can you all consider either amplifying your docs that you have when replying to devel questions or doing more docs driven development so we can see answers and theories to things like the orphaning release/stream branches questions thought out earlier?
19:53:25 <bexelbie> didn't mean to make that leading
19:53:52 <mattdm> bexelbie: lol sure you didn't
19:53:54 <langdon> bexelbie: apologies.. but i have no idea what you mean
19:54:15 <langdon> what development do you mean?
19:54:19 <bexelbie> basically, I am seeing lots of questions asked around process and it seems the answers are "we haven't considered that" not "we have this theory" that concerns me
19:54:20 <langdon> dnf?
19:54:28 <bexelbie> admitably I am a casual follower
19:54:35 <bexelbie> langdon, of modularity and workflows
19:54:41 <bexelbie> I guess more the workflows
19:54:50 <langdon> ohh... i would never have called that "docs driven dev" ..
19:54:52 <bexelbie> I feel like this needs more workflow than tech - dnf issues aside
19:55:00 <bexelbie> but it is an awesome term
19:55:02 <langdon> ha
19:55:14 <langdon> we have it already.. we call it biz-driven dev
19:55:18 <langdon> ok..
19:55:54 <bexelbie> fair -- and it was a really badly worded question so we can move on
19:55:58 <bexelbie> if I get it better, I'll email it
19:56:02 <langdon> so.. i feel like the problem is.. a) there are lots of docs that people havent read b) there does not seem to be strong consensus on where process should live
19:56:14 <bexelbie> how can we fix a and b?
19:56:15 <bexelbie> fesco?
19:56:17 <langdon> but the "addiing a module" i linked to above is a step by step walktrhough
19:56:23 <bexelbie> more articles? workshops?
19:56:44 <langdon> i think it is articles/workshops but "amplification" of what we have is key
19:56:46 <bexelbie> do you mean the article by jdieter?
19:57:00 <langdon> no.. asamalik's doc
19:57:00 * bexelbie is always leary of articles as docs as they aren't long term findable
19:57:04 <langdon> 2nd one
19:57:07 <bexelbie> ahh, cool - those tend to be awesome
19:57:23 <bexelbie> should we get a virtual presentation together where someone walks a real package through this?
19:57:24 <asamalik> what have I done?
19:57:27 <langdon> bexelbie: the article is also good.. but not docs
19:57:31 <bexelbie> something packagers can relate too?
19:57:43 <bexelbie> langdon, I read jdieter 's article and liked it a lot!
19:57:46 <bexelbie> :D
19:57:46 <langdon> asamalik: ha.. just the doc we discussed earlier
19:58:04 <bexelbie> or focus groups?
19:58:14 <bexelbie> that could also include ambassador/contributor education
19:58:14 <langdon> bexelbie: we could.. that is what jdieter did in the expert session.. but he was the only non-expert there.. (aside from me of course ;) )
19:58:24 * mattdm needs to go.
19:58:28 <langdon> bexelbie: yes.. alll those things would help
19:58:36 <bexelbie> I look forward to seeing them :D
19:58:38 <mattdm> I think dperpeet is not here for CI update -- hopefully new time works out better
19:58:51 <langdon> bexelbie: i thought you were volunteering :)
19:58:55 <mattdm> if you guys want to keep going, can you please turn off the lights when you are done? :)
19:58:57 <bexelbie> if you need logistics help let me know - my package is a terrible volunteer one
19:59:07 <langdon> now that devconf is over.. and i can pay attention again.. ill see what i can do
19:59:15 <bexelbie> something by an experienced packager for experienced packagers is where I think we should start
19:59:36 <sdgathman> Magazine articles are not so much docs as "how tos".  I'm not volunteering, but there could be a howto index linking to such articles.
19:59:43 <langdon> i think we are done.. but we should probably have some folow up on amplification
19:59:48 <bexelbie> basically, if I understand it the first time through without searching for details on basic packaging concepts, we failed :P
19:59:54 <sdgathman> E.g. todays would be linked under "how to scale images"
20:00:16 <bexelbie> sdgathman, it'd be great to see them converted to quickdocs for that ... which gives more durability too
20:00:18 <stickster> OK magazine time guys
20:00:28 <bexelbie> and should be relatively easy from a writing perpsective
20:00:30 <bexelbie> thank you all
20:00:31 <bexelbie> #endmeeting