fedora_coreos_meeting
LOGS
16:29:56 <dustymabe> #startmeeting fedora_coreos_meeting
16:29:56 <zodbot> Meeting started Wed Jan 15 16:29:56 2020 UTC.
16:29:56 <zodbot> This meeting is logged and archived in a public location.
16:29:56 <zodbot> The chair is dustymabe. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:29:56 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:29:56 <zodbot> The meeting name has been set to 'fedora_coreos_meeting'
16:30:00 <dustymabe> #topic roll call
16:30:19 <bgilbert> .hello2
16:30:20 <zodbot> bgilbert: bgilbert 'Benjamin Gilbert' <bgilbert@backtick.net>
16:30:46 <slowrie> .hello2
16:30:47 <zodbot> slowrie: slowrie 'Stephen Lowrie' <slowrie@redhat.com>
16:30:47 <kaeso[m]> .hello lucab
16:30:50 <zodbot> kaeso[m]: lucab 'Luca Bruno' <lucab@redhat.com>
16:31:01 <mnguyen_> .hello mnguyen
16:31:04 <zodbot> mnguyen_: mnguyen 'Michael Nguyen' <mnguyen@redhat.com>
16:31:14 <dustymabe> .hello2
16:31:15 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dusty@dustymabe.com>
16:31:38 <miabbott> .hello miabbott
16:31:39 <zodbot> miabbott: miabbott 'Micah Abbott' <miabbott@redhat.com>
16:31:43 <jlebon> .hello2
16:31:44 <zodbot> jlebon: jlebon 'None' <jonathan@jlebon.com>
16:32:02 <dustymabe> #chair bgilbert slowrie kaeso[m] mnguyen_ jlebon miabbott
16:32:02 <zodbot> Current chairs: bgilbert dustymabe jlebon kaeso[m] miabbott mnguyen_ slowrie
16:33:08 <red_beard> .hello redbeard
16:33:09 <zodbot> red_beard: redbeard 'Brian 'redbeard' Harrington' <bharring@redhat.com>
16:33:14 * bgilbert waves at red_beard
16:33:17 <dustymabe> #chair red_beard lorbus
16:33:17 <zodbot> Current chairs: bgilbert dustymabe jlebon kaeso[m] lorbus miabbott mnguyen_ red_beard slowrie
16:33:20 <red_beard> sup doggy
16:33:33 * dustymabe waves at all :)
16:33:51 <dustymabe> #topic Action items from last meeting
16:33:57 * mnguyen_ waves
16:34:18 <dustymabe> * miabbott to help us get the azure image upload/boot tested
16:34:20 <dustymabe> * jlebon to ask more about the contstraints in #338
16:34:28 <dustymabe> I think miabbott was able to delegate that task
16:35:03 <miabbott> aye, passed it off to darkmuggle.  but he ran into troubles with his account/access, so i don't believe the task is complete
16:35:29 <dustymabe> miabbott: fun all around!
16:35:57 <jlebon> #info jlebon asked more about the constraints in https://github.com/coreos/fedora-coreos-tracker/issues/338#issuecomment-572225679
16:36:05 <miabbott> something in the RH account that was used seemed to cause the uploaded FCOS images to be deleted
16:36:13 <miabbott> I'll continue to chase it down
16:36:33 <dustymabe> +1
16:36:44 <dustymabe> thanks jlebon
16:37:05 <dustymabe> I'll jump into topics now.. mostly one big one
16:37:16 <dustymabe> #topic stable stream updates
16:37:47 <dustymabe> So we have created a stable stream and done two builds on that stream (thanks to all who helped get it all working)
16:38:07 <dustymabe> verified upgrades worked and plumbed everything through the update service
16:38:16 <dustymabe> thanks jlebon, and kaeso[m] for all the hard work there
16:38:53 <kaeso[m]> yay!
16:38:57 <jlebon> it was really a team effort (and we're not done yet!)
16:39:06 <dustymabe> At this point I think it's fair for us to start talking about what's important for us to do next
16:39:09 <miabbott> jlebon++
16:39:09 <zodbot> miabbott: Karma for jlebon changed to 2 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:39:11 <miabbott> dustymabe++
16:39:14 <miabbott> kaeso[m]++
16:39:14 <zodbot> miabbott: Karma for dustymabe changed to 3 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:39:21 <miabbott> bgilbert++
16:39:21 <zodbot> miabbott: Karma for bgilbert changed to 2 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:39:21 <dustymabe> #cookieparty
16:39:24 <miabbott> abai++
16:39:24 <zodbot> miabbott: Karma for abai changed to 1 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:40:03 <miabbott> you folks worked so hard to get to this...i'm incredibly proud and happy for you 👍
16:40:23 <dustymabe> Thanks miabbott!
16:40:51 <jlebon> awwww :)
16:41:04 <abai> miabbott++
16:41:04 <zodbot> abai: Karma for miabbott changed to 1 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:41:14 <dustymabe> There has been a lot of collaboration. It'll be nice to now settle into a normal release cadence and tick off the important things that are left
16:41:31 <bgilbert> +1
16:41:46 <dustymabe> Items that I see that are important to work on next from an infrastructure/release tooling perspective:
16:42:18 <dustymabe> - automated ostree imports (get rid of manual releng involement during releases): https://github.com/coreos/fedora-coreos-tracker/issues/199
16:42:37 <dustymabe> ^^ That one is on me and I plan to work on it as soon as I'm back from devconf
16:42:59 <dustymabe> - getting our update services running in Fedora Infra's prod openshift (i.e. not stg)
16:43:14 <dustymabe> ^^ kaeso[m]: was there a ticket for that?
16:43:59 <kaeso[m]> dustymabe: I don't think so
16:44:24 <dustymabe> Maybe let's make one in the tracker and work together on it?
16:44:27 <dustymabe> WDYT?
16:44:40 <jlebon> also critical to me: better testing story, esp. updates testing
16:45:08 <kaeso[m]> dustymabe: on pagure infra, yes
16:45:44 <dustymabe> jlebon: it would certainly be nice if we could read the "graph" of possible upgrade paths and execute common ones or something
16:46:27 <miabbott> that sounds like the kind of testing that is done for OCP
16:46:27 <dustymabe> any other pieces that need to happen first after this milestone?
16:46:38 <jlebon> dustymabe: yeah, that sounds plausible.  i've been meaning to look into this, even if it's just the basic case of N-1 -> N for now
16:47:27 <miabbott> will the release tasks for the next stable release be as significant as this initial release?  i think i counted 35 tasks on the tracker ticket
16:47:53 <dustymabe> miabbott: i'm not sure what you mean exactly?
16:48:05 <jlebon> dustymabe: i think we can focus on those for now.  there's a lot of other important things, but somewhat less so than those
16:48:13 <miabbott> meaning, do we have another 35 tasks to run through the next time we bump stable
16:48:18 <dustymabe> I think we plan to just do time based updates
16:48:25 <jlebon> yeah, that's one of the slightly less important task i want to tackle: automate more of the release process :)
16:48:46 <jlebon> the ostree importer thing itself will help a lot there
16:49:08 <dustymabe> miabbott: i.e. the steps in https://github.com/coreos/fedora-coreos-tracker/issues/199 ?
16:49:18 <miabbott> https://github.com/coreos/fedora-coreos-streams/issues/47
16:49:27 <dustymabe> oops meant that link^^ yeah
16:49:50 <dustymabe> miabbott: yep, we want to minimize that list
16:50:10 <dustymabe> the importer will be a big part of it
16:50:13 <miabbott> ah i should read the final line `NOTE: In the future, most of these steps will be automated and a syncer will push the updated metadata to S3.`
16:50:16 <jlebon> the checklist includes hints/guidelines on how we'll reduce those steps in the future
16:50:18 <miabbott> ¯\_(ツ)_/¯
16:50:29 * miabbott goes back to lurking
16:50:47 <dustymabe> bgilbert: can you think of anything from an infrastructure/release process perspective that we should concentrate on first?
16:51:13 <dustymabe> if not, how about from a feature/bug perspective within the released artifacts?
16:51:52 <bgilbert> eh, my only major concern with the release process is the manual ostree import
16:51:59 <bgilbert> other than that, whatever's low-hanging fruit
16:52:20 <dustymabe> +1
16:52:36 <dustymabe> ok real quick. right now we just did a testing and a stable release
16:52:38 <bgilbert> re the distro itself, continuing work to add platforms
16:53:00 <dustymabe> we should get these both on a preferred release cadence
16:53:13 <bgilbert> and docs.  so many docs.
16:53:15 <dustymabe> which I think means releasing testing, waiting some time, then releasing stable based on testing
16:53:32 <bgilbert> +1
16:53:46 <dustymabe> What should that cadence be? Do we have that written down already?
16:53:49 <bgilbert> 2 weeks
16:54:04 <bgilbert> https://github.com/coreos/fedora-coreos-tracker/blob/master/Design.md#production-refs
16:54:43 <dustymabe> bgilbert: right. so we theoretically do a testing and a stable release at the same time
16:54:49 <dustymabe> testing is based on latest content
16:54:55 <dustymabe> stable is based on last testing, correct?
16:55:06 <jlebon> that's my understanding
16:55:10 <bgilbert> for in-cycle releases, yes
16:55:14 <jlebon> basically the reverse of what we did this release :)
16:55:14 <bgilbert> not for out-of-cycle ones
16:55:51 <dustymabe> so to get us off the ground (and to not wait a month until the next stable) how about we do a testing release next week and then start the "every two weeks releases" the week after that
16:56:23 <dustymabe> so next week: testing release
16:56:35 <dustymabe> following week: testing and stable release based on the previous week's testing
16:57:05 <bgilbert> is that a newly branched testing?
16:57:07 <bgilbert> for week 1
16:57:26 <dustymabe> it would be content from latest testing-devel
16:57:40 <dustymabe> so approximately content from 01/21
16:58:07 <jlebon> seems reasonable to me
16:58:10 <bgilbert> +1
16:58:14 <dustymabe> cool
16:58:26 <dustymabe> glad we got that worked out. I'll create tracker tickets for those two releases
16:58:38 <dustymabe> bgilbert: would you like to execute that testing release?
16:58:43 <bgilbert> sure
16:58:44 <dustymabe> to get in the rotation
16:58:46 <dustymabe> cool
16:58:48 <dustymabe> sounds great
16:59:08 <dustymabe> and maybe we can get kaeso[m] (or any other volunteers) for the stable the week after
16:59:38 <bgilbert> one tip based on our CL experience:
16:59:46 <bgilbert> when doing simultaneous releases, have the same operator for all of them
16:59:52 <bgilbert> the checklists ensure the operator won't get confused
16:59:58 <bgilbert> and there are economies of scale
17:00:36 <dustymabe> seems fine to me. @jlebon and I ran these last two releases of both testing and stable in parallel
17:00:44 <jlebon> yeah, agreed. it's not 2x the work to do both in parallel
17:00:55 <dustymabe> +1
17:01:05 <dustymabe> ok I'll move on to next topic now
17:01:18 <dustymabe> #topic GitHub project boards
17:01:48 <dustymabe> We have the stable release board: https://github.com/orgs/coreos/projects/84
17:01:57 <dustymabe> and the papercuts board: https://github.com/orgs/coreos/projects/83
17:02:11 <dustymabe> anything we want to do with those at this time?
17:02:35 <bgilbert> I gardened them a bit the other day
17:02:53 <miabbott> so i have a radical proposal that i'm sure is going to go over well
17:02:57 * dustymabe inserts garden emoji here
17:03:04 <bgilbert> ISTM the larger question is: do we want a new project board for post-preview-exit?
17:03:35 <bgilbert> I'm not sure they've been working super-well in practice
17:03:35 <dustymabe> bgilbert: or no project board at all, or more project boards
17:03:49 <miabbott> the RHT folks that work on FCOS have been using Jira to track work internally.  but recently we've switched to a Jira instance that is public facing.
17:03:56 <bgilbert> dustymabe: right, I'm assuming the existing ones are basically done
17:04:01 <bgilbert> (maybe not the papercuts one)
17:04:30 <miabbott> in order to duplicate some paperwork across boards, i'd like to explore using the public Jira instance for FCOS work
17:04:31 <dustymabe> i'm cool with whatever we want to do with the project boards, but figured it'd be good to bring it up
17:04:42 * miabbott puts on flame retardant gear
17:05:08 <jlebon> miabbott: you mean deduplicate, right? ;)
17:05:23 <miabbott> right :facepalm:
17:05:27 <dustymabe> miabbott: i'm not sure how much using a public jira would help, but i'm not super opposed to trying it
17:05:38 <bgilbert> so, those of us who work for the entity will be using JIRA _anyway_
17:06:04 <bgilbert> and I'm not sure the GH board has been useful enough to maintain separately
17:06:55 <bgilbert> so long as we're using JIRA only for work tracking, and keeping discussion/decisions in GH, I'm +1 to public JIRA
17:07:09 <dustymabe> I think it would depend on the details of how things are set up
17:07:24 <dustymabe> miabbott: would you be willing to add some details to the proposal?
17:07:49 <dustymabe> my concern: now I'm managing two JIRA boards vs one
17:07:53 <miabbott> dustymabe: yeah, this is just the initial "i have this idea".  it's going to need more careful thought, which i haven't done yet
17:08:07 <dustymabe> cool. we'll wait to hear more from you :)
17:08:44 <dustymabe> #topic open floor
17:08:49 <x3mboy> !
17:08:51 <dustymabe> anyone with anything for open floor?
17:08:52 <miabbott> thanks.  the lack of pitchforks is encouraging
17:09:00 <x3mboy> .hello2
17:09:01 <zodbot> x3mboy: x3mboy 'Eduard Lucena' <eduardlucena@gmail.com>
17:09:06 <x3mboy> Hello guys
17:09:08 <dustymabe> welcome x3mboy
17:09:14 <x3mboy> Sorry for intercepting your meeting
17:09:25 <x3mboy> I have a point I like to bring
17:09:36 <x3mboy> I'm talking in behalf of the marketing team, as always
17:10:03 <x3mboy> I would love to have some feedback from you on this design ticket: https://pagure.io/design/issue/642
17:10:40 <dustymabe> #info x3mboy and the marketing team would like some feedback on this design ticket: https://pagure.io/design/issue/642
17:10:42 <x3mboy> The intention is to have an informative infographics published in our social media accounts about the things we do on the Fedora Project
17:11:31 <dustymabe> looks like you've previously worked with bgilbert in the ticket a bit
17:11:37 <x3mboy> So, basically, the intention of this first CoreOS inforgraphic is to bring people attention on what is FCOS and some highlights of the most interesting facts about it
17:11:57 <x3mboy> Yes, I was, but 6 month ago we lost track of it
17:12:12 <jlebon> +1  with stable out, it seems a more appropriate time to focus on this now
17:12:24 <dustymabe> bgilbert: do you care to pick back up with x3mboy?
17:13:11 <bgilbert> yup, can do
17:13:42 <dustymabe> +1, lean on the rest of us too. Would it be worth a video chat at some point to help have a discussion about it?
17:13:47 <dustymabe> if so we can try to schedule something like that
17:13:50 <x3mboy> Thanks
17:13:53 <bgilbert> #action bgilbert to work with x3mboy on FCOS infographic
17:13:54 <bgilbert> sure
17:13:59 <x3mboy> bgilbert++
17:13:59 <zodbot> x3mboy: Karma for bgilbert changed to 3 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:14:03 <dustymabe> bgilbert++
17:14:04 <x3mboy> dustymabe++
17:14:04 <zodbot> x3mboy: Karma for dustymabe changed to 4 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:14:17 <dustymabe> anyone with anything else for open floor?
17:14:36 <fos> I could throw better Btrfs support in the round ;-)
17:14:55 <bgilbert> fos: in Ignition?
17:15:03 * red_beard grabs popcorn for this one
17:15:06 <fos> Yes ;-)
17:15:08 <bgilbert> ah yeah
17:15:26 <bgilbert> fos: ajeddeloh left RH, if you didn't know
17:15:43 <fos> Oh, I didn't - thanks for the notice!
17:15:50 <bgilbert> so responsibilities are getting reshuffled a bit
17:16:23 <bgilbert> my own opinion is that it'd be good functionality to have, but of course it's not a great fit for the existing schema so we'd have to figure that out
17:16:45 <fos> There's no hurry, we implemented a poor man's solution for now, but I guess it would be a good feature in the future.
17:16:52 <bgilbert> yeah
17:16:58 <bgilbert> what was the short-term fix?
17:17:46 <fos> Just extending the key to allow multiple mounts of subvolumes. This of course won't create new subvolumes, but at least existing ones can be used.
17:18:17 <bgilbert> ah, okay
17:19:01 <fos> Basically https://github.com/bmwiedemann/openSUSE/commit/ca57b9df63b73b4c00f4040380e2b6025b83f836#diff-d14249dbd91ae48e59cd13beaa78930c
17:20:17 <fos> As I said, there's no hurry - and now I know where there hasn't been any response.
17:20:22 <dustymabe> +1
17:20:32 <dustymabe> anything else for open floor?
17:21:43 <dustymabe> #endmeeting