16:31:23 <jbrooks> #startmeeting fedora_coreos_meeting
16:31:23 <zodbot> Meeting started Wed Aug 15 16:31:23 2018 UTC.
16:31:23 <zodbot> This meeting is logged and archived in a public location.
16:31:23 <zodbot> The chair is jbrooks. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:31:23 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:31:23 <zodbot> The meeting name has been set to 'fedora_coreos_meeting'
16:31:30 <slowrie> .hello2
16:31:31 <zodbot> slowrie: slowrie 'Stephen Lowrie' <slowrie@redhat.com>
16:31:32 <jbrooks> #topic roll call
16:31:33 <ajeddeloh> .hello2
16:31:37 <zodbot> ajeddeloh: ajeddeloh 'Andrew Jeddeloh' <andrew.jeddeloh@redhat.com>
16:31:37 <jdoss> .hello2
16:31:40 <zodbot> jdoss: jdoss 'Joe Doss' <joe@solidadmin.com>
16:31:41 <strigazi> .hello2
16:31:43 <zodbot> strigazi: strigazi 'Spyros Trigazis' <strigazi@gmail.com>
16:31:54 <jlebon> .hello2
16:31:55 <zodbot> jlebon: jlebon 'None' <jonathan@jlebon.com>
16:31:57 <yzhang> .hello2
16:31:58 <zodbot> yzhang: yzhang 'Yu Qi Zhang' <jzehrarnyg@gmail.com>
16:32:20 <rfairley> .hello2
16:32:21 <zodbot> rfairley: Sorry, but you don't exist
16:32:29 <rfairley> .hello rfairleyredhat
16:32:30 <zodbot> rfairley: rfairleyredhat 'Robert Fairley' <rfairley@redhat.com>
16:32:32 <sanja> .hello2
16:32:33 <dustymabe> .hello2
16:32:34 <geoff-> hi: Geoff Levand <geoff@infradead.org>
16:32:36 <zodbot> sanja: sanja 'Sanja Bonic' <sanja@redhat.com>
16:32:39 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dusty@dustymabe.com>
16:32:40 <dustymabe> kaeso: \o
16:32:42 <mskarbek> .hello2
16:32:43 <zodbot> mskarbek: mskarbek 'None' <redhat@skarbek.name>
16:32:57 <kaeso> .hello2
16:32:58 <zodbot> kaeso: Sorry, but you don't exist
16:33:19 <kaeso> .hello lucab
16:33:21 <zodbot> kaeso: lucab 'Luca Bruno' <lucab@redhat.com>
16:33:41 <jbrooks> #chair slowrie ajeddeloh jdoss strigazi jlebon yzhang rfairley sanja dustymabe geoff- mskarbek kaeso
16:33:41 <zodbot> Current chairs: ajeddeloh dustymabe geoff- jbrooks jdoss jlebon kaeso mskarbek rfairley sanja slowrie strigazi yzhang
16:33:49 <jbrooks> #topic Action items from last meeting
16:33:53 <miabbott> .hello2
16:33:55 <zodbot> miabbott: miabbott 'Micah Abbott' <miabbott@redhat.com>
16:33:56 <bgilbert> .hello2
16:33:57 <bhavin192> .hello2
16:33:58 <zodbot> bgilbert: bgilbert 'Benjamin Gilbert' <bgilbert@backtick.net>
16:34:01 <zodbot> bhavin192: bhavin192 'Bhavin Gandhi' <bhavin7392@gmail.com>
16:34:28 <jbrooks> #chair bgilbert bhavin192
16:34:28 <zodbot> Current chairs: ajeddeloh bgilbert bhavin192 dustymabe geoff- jbrooks jdoss jlebon kaeso mskarbek rfairley sanja slowrie strigazi yzhang
16:34:35 <jbrooks> #chair miabbott
16:34:35 <zodbot> Current chairs: ajeddeloh bgilbert bhavin192 dustymabe geoff- jbrooks jdoss jlebon kaeso miabbott mskarbek rfairley sanja slowrie strigazi yzhang
16:34:43 <jbrooks> dustymabe to update #18 with results of partition layout discussion  and close ticket
16:34:55 <jbrooks> ksinny to update #13 to reflect her, ed, geoff are point team for  aarch64
16:35:23 * dustymabe looks
16:36:05 <dustymabe> I did update #18
16:36:12 <dustymabe> #link https://github.com/coreos/fedora-coreos-tracker/issues/18#issuecomment-409668929
16:36:34 <jbrooks> #info ksinny did update #13
16:36:47 <jbrooks> #link https://github.com/coreos/fedora-coreos-tracker/issues/13#issuecomment-409660564
16:37:02 <jbrooks> dustymabe, Is that ticket ready to close
16:37:31 <dustymabe> which one ?
16:37:40 <dustymabe> #18 ?
16:37:48 <jbrooks> Yeah, part of the action was to close it
16:37:58 <dustymabe> ahh I see
16:38:34 <ajeddeloh> relating to closing issues: do we want to create "rolling design doc" that we consolidate the results from the discussion in the issues into
16:39:05 <jbrooks> ajeddeloh, That sounds like a good idea, is there an example we can follow for that
16:39:08 <dustymabe> jbrooks: closed
16:39:10 <bgilbert> ajeddeloh: I could see that going either way
16:39:19 <bgilbert> there's value in having separate tickets open for outstanding issues
16:39:35 <ajeddeloh> not that I know off the top of my head, I can PR one today
16:39:35 <bgilbert> but we should certainly do one or the other
16:39:44 <ajeddeloh> bgilbert: I mean for issues where we've come to a conclusion
16:39:52 <dustymabe> yeah. It would probably be worth having a "decisions" .md file in the repo
16:39:54 <bgilbert> we've come to a conclusion but we haven't exactly done the thing yet :-)
16:40:01 <dustymabe> somewhere in https://github.com/coreos/fedora-coreos-tracker
16:40:03 <jbrooks> #action ajeddeloh to PR rolling design doc for comment
16:40:13 <jbrooks> And then we can chat in the PR comments
16:40:30 <ajeddeloh> sgtm
16:40:33 <jbrooks> Cool
16:40:45 <jbrooks> On to issues labeled meeting
16:40:49 <kaeso> do we have docs-related repo already? we can cut a "dev" or "design" subdir in there
16:41:18 <jbrooks> I think the tracker is the first we've created as a project?
16:41:27 <ajeddeloh> kaeso: lets discuss that in the PR
16:41:38 <kaeso> ack
16:41:52 <jbrooks> #topic arm64 / aarch64 support for Fedora CoreOS
16:42:00 <sanja> kaeso I'm doing the docs repo
16:42:04 <sanja> this week
16:42:04 <jbrooks> #link https://github.com/coreos/fedora-coreos-tracker/issues/13
16:42:10 <sanja> will update the issue with the link then
16:42:19 <jbrooks> sanja, should I action you on that
16:42:24 <sanja> yes please
16:42:29 <jbrooks> #action sanja to create docs repo this week
16:42:33 <dustymabe> do we have ed-packet or geoff- here today :)
16:42:38 <geoff-> hi
16:42:46 <dustymabe> I know sinny is AFK today I believe
16:43:26 <geoff-> I don't have any thing arm64 specific today
16:43:52 <jbrooks> It looks like one pending action from the ticket was to request arm64 hw
16:44:03 <jbrooks> #link https://github.com/coreos/fedora-coreos-tracker/issues/13#issuecomment-411151165
16:44:33 <jbrooks> OK, we can chat about this issue next time
16:44:38 <jbrooks> On to open floor
16:44:45 <jbrooks> #topic Open Floor
16:45:00 <jbrooks> dustymabe, Any news from Flock?
16:45:08 <geoff-> ed-packet and I have been working to bring up fedora on the Cavium ThunderX servers at Packet
16:45:39 <jbrooks> geoff-, Cool, it might be worth dropping a note in the ticket for anyone who's following along
16:45:56 <dustymabe> #info benjamin gilbert and I gave a talk at FLock about Fedora CoreOS and how we got here
16:46:08 <jlebon> doesn't seem like videos are up yet on the Fedora youtube channel
16:46:08 <dustymabe> #link https://dustymabe.fedorapeople.org/2018-08_Building-Fedora-CoreOS.pdf
16:46:16 <dustymabe> ^^ there is a link to the slides
16:46:23 <strigazi> Has the discussion on customizing the host started somewhere? We can collect use cases there
16:46:28 <dustymabe> no videos yet. when there are videos I'll send a mail to the list
16:46:57 <kaeso> #info lucab gave a lightning talk at Flock about ignition, no slides but good reception/questions
16:47:00 <dustymabe> strigazi: most of our discussions are in tickets
16:47:13 <dustymabe> https://github.com/coreos/fedora-coreos-tracker/issues
16:47:17 <ajeddeloh> strigazi: have anything in mind in particular?
16:47:43 * ajeddeloh cares deeply about Ignition
16:47:52 <strigazi> dustymabe: I was thinking something like the survey for the cri and docker option we did for fedora atomic
16:48:30 <dustymabe> strigazi: i.e. to see who is using what ?
16:48:36 <strigazi> dustymabe: yes
16:48:39 <dustymabe> vs cri-o podman and docker ?
16:49:05 <bgilbert> strigazi: we know we'll need to support a range of options there
16:49:09 <strigazi> yes, but also syscontainers
16:49:17 <bgilbert> the percentages will change between now and launch
16:49:49 <jdoss> dustymabe: is there a video of this talk?
16:49:59 <jdoss> derp just read up sorry
16:50:23 <dustymabe> jdoss: :)
16:50:43 <dustymabe> strigazi: some discussion of runtimes is in this ticket: https://github.com/coreos/fedora-coreos-tracker/issues/14
16:50:49 <geoff-> has the topic of rkt support in FCOS come  up?
16:50:52 <bgilbert> since it's come up a few times, should we have an explicit ticket for discussing system containers?
16:51:06 <dustymabe> bgilbert: probably
16:51:15 <strigazi> bgilbert: dustymabe if you have already and indication of what people is using sounds good to me. I thought a small survey will allow more passive user to be heard?
16:51:41 <dustymabe> strigazi: would you possibly like to work with sanja to see what that survey would look like ?
16:51:49 <bgilbert> strigazi: I'm fine doing it if there's a purpose, just don't want to cause survey fatigue
16:52:06 <jbrooks> strigazi, There's no fedora coreos to use yet, would it be a survey of cl and atomic users
16:52:06 <bgilbert> current thinking is: multiple docker versions, podman, multiple cri-o and kubelet versions
16:52:08 <bgilbert> no rkt
16:52:17 <strigazi> dustymabe: sure, we can discuss it and do it only if it makes sense
16:52:24 <strigazi> bgilbert: ^^
16:52:31 <kaeso> geoff-: yes, FCOS likely won't carry it
16:52:34 <dustymabe> strigazi: +1, real quick though
16:52:43 <dustymabe> what bgilbert said above
16:52:49 <dustymabe> 'current thinking is: multiple docker versions, podman, multiple cri-o and kubelet versions'
16:53:04 <dustymabe> that doesn't cover system containers, or torcx
16:53:15 <dustymabe> but does that answer the "runtime" question?
16:53:58 <strigazi> the runtime yes, it almost answers, people are experimenting with kata too
16:54:35 <dustymabe> strigazi: fun :)
16:54:38 <strigazi> it doesn't make sense to ship it in the host, but with syscontainers we managed to run it
16:54:47 <bgilbert> anyone with opinions on system containers want to file that ticket?
16:54:55 <ajeddeloh> I'm also a little worried about relying too much on surveys since *i think* they skew towards the existing fedora base. The CL user base isn't as involved per user. We have a lot of users that never talk to us. We put out a CT/Ignition/Cloudinit survey and got like 25 responses. 10 were from CoreOS employees or something like that
16:55:34 <strigazi> bgilbert: I will summarize in an issue
16:55:49 <dustymabe> ajeddeloh: i get that argument, but poking around in the dark can be frustrating too
16:55:56 <jbrooks> #action strigazi to file ticket for system containers discussion
16:55:59 <ajeddeloh> Not saying that's a bad thing either, it's great that it works as a "launch and forget" OS, but it means that if everythings working they don't really see a need to get involved
16:56:19 <ajeddeloh> dustymabe: agreed, just don't want to put *too* much weight on them
16:57:01 <jlebon> nice slides -- so how was the reception for the FCOS talk?
16:57:12 <dustymabe> seemed to go over pretty well I thought
16:57:20 <bgilbert> yeah
16:57:23 <strigazi> jlebon: it was a full room
16:57:35 <dustymabe> no one stormed out or called us crazy or anything
16:57:35 <jlebon> nice
16:57:44 <jbrooks> dustymabe, any juicy pointed q's
16:57:45 <jbrooks> :)
16:58:17 <dustymabe> i don't think so.. there was a questions about provisioning on bare metal with static IPs
16:58:27 <bgilbert> we had one "why are automatic updates good"
16:58:35 <bgilbert> but it wasn't especially pointed :-)
16:58:55 <jlebon> as in... they *didn't* want automatic updates?
16:59:32 <bgilbert> my read was more that they wanted us to be explicit about it
16:59:36 <dustymabe> jlebon: probably just a very foreign concept to someone who is risk averse
17:00:00 <jlebon> gotcha
17:00:11 <jbrooks> Anything else from Flock, or other open floor topics?
17:00:23 <ajeddeloh> So for Ignition
17:00:29 <dustymabe> jbrooks: i'm preparing the PRD for an update based on our "use cases" discussion from last time
17:00:37 <dustymabe> will send that out once it's updated
17:00:39 <ajeddeloh> Do we want FCOS to accept 2.x spec configs
17:01:12 <bgilbert> ajeddeloh: maybe that'd go better in a ticket?
17:01:19 <jbrooks> ajeddeloh, Sounds like a ticket is in order, do we have one for ignition?
17:01:31 <ajeddeloh> There's some problems (will link in a sec) with 2.x that we want to fix with 3.x, but that means everyone transitioning from CL will have another thing that needs to change
17:01:48 <ajeddeloh> jbrooks: +1 will create
17:01:54 <bgilbert> "another thing that needs to change": we'll want a translator anyway
17:02:04 <jbrooks> #action ajeddeloh to file ticket regarding ignition and spec versions
17:02:21 <dustymabe> ooh ooh.. FYI - ignition in Fedora recently learned about SELinux
17:02:22 <ajeddeloh> https://github.com/coreos/ignition/issues/608 < issue if anyone is curious
17:02:27 <dustymabe> #link https://github.com/coreos/ignition/pull/569
17:02:31 <dustymabe> nice work @jlebon
17:03:16 <jlebon> :)
17:03:26 <jlebon> thanks ajeddeloh for reviewing
17:04:23 <dustymabe> btw thanks strigazi for being an involved member of the community
17:04:41 <dustymabe> it was good to see you at flock, I think you met kaeso and bgilbert too ?
17:04:53 <kaeso> yup
17:04:57 <strigazi> I really like this community :)
17:05:02 <bgilbert> :-)
17:05:05 <strigazi> yeap, both
17:05:25 <dustymabe> strigazi: soon enough those thousands of atomic host nodes under openstack magnum will be Fedora CoreOS nodes :)
17:05:43 <jbrooks> Awesome
17:05:51 <jbrooks> Anything else for this week?
17:05:53 <geoff-> please let us know of any Flock video presentations of interest, maybe post to the  ML
17:06:15 <dustymabe> geoff-: yep. when they're up i plan to send something out
17:06:18 <strigazi> dustymabe: yes, the sooner the better
17:07:17 <bgilbert> apologies in advance for the combination of audio problems and having to watch us deal with audio problems
17:07:38 <dustymabe> haha..
17:07:53 <dustymabe> a few of us are going to devconf.us this weekend right ?
17:08:11 <dustymabe> i know walters was planning on a talk
17:08:14 <dustymabe> maybe jlebon
17:08:22 * miabbott raises hand
17:08:31 <dustymabe> woot for micah
17:08:37 <jlebon> yes, I'm attending, not talking \o/
17:08:38 <jbrooks> #link https://devconfus2018.sched.com/
17:09:37 <dustymabe> cool. cool. i don't know if they'll have live streaming (they have at some devconfs I think)
17:09:49 <dustymabe> will try to catch those videos later too
17:11:05 <jbrooks> OK, I'm going to close
17:11:15 <jbrooks> 3
17:11:16 <jbrooks> 2
17:11:17 <jbrooks> 1
17:11:20 <jbrooks> #endmeeting