16:02:47 <jberkus> #startmeeting atomic-community
16:02:47 <centbot> Meeting started Mon Oct 16 16:02:47 2017 UTC.  The chair is jberkus. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:47 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:02:47 <zodbot> Meeting started Mon Oct 16 16:02:47 2017 UTC.  The chair is jberkus. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:02:47 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:02:47 <zodbot> The meeting name has been set to 'atomic-community'
16:03:08 <jberkus> #topic roll call
16:03:17 <jberkus> .hello jberkus
16:03:19 <zodbot> jberkus: jberkus 'Josh Berkus' <josh@agliodbs.com>
16:03:26 <walters> .hello walters
16:03:27 <dustymabe> .hello dustymabe
16:03:27 <zodbot> walters: walters 'Colin Walters' <walters@redhat.com>
16:03:30 <walters> (but i'll be back in 5 mins)
16:03:30 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
16:03:34 <ashcrow> .hello smilner
16:03:36 <zodbot> ashcrow: smilner 'None' <smilner@redhat.com>
16:03:41 <jbrooks> .hello jasonbrooks
16:03:43 <zodbot> jbrooks: jasonbrooks 'Jason Brooks' <JBROOKS@REDHAT.COM>
16:03:46 <jlebon> .hello jlebon
16:03:47 <zodbot> jlebon: jlebon 'None' <jlebon@redhat.com>
16:04:05 <dwalsh> .hello dwalsh
16:04:06 <zodbot> dwalsh: dwalsh 'Daniel J Walsh' <dwalsh@redhat.com>
16:04:11 <tsweeney> .hello tsweeney
16:04:12 <zodbot> tsweeney: Sorry, but you don't exist
16:04:22 <dwalsh> :^)
16:04:30 <tsweeney> lol, I just went poof!
16:05:55 <jberkus> #chair walters dustymabe tsweeney jlebon jbrooks dwalsh smilner
16:05:55 <centbot> Current chairs: dustymabe dwalsh jberkus jbrooks jlebon smilner tsweeney walters
16:05:55 <zodbot> Current chairs: dustymabe dwalsh jberkus jbrooks jlebon smilner tsweeney walters
16:06:13 <jberkus> #topic new hardware ideas for blades
16:07:18 <jzb> .hellomynameis jzb
16:07:19 <zodbot> jzb: jzb 'Joe Brockmeier' <jzb@redhat.com>
16:07:23 <jberkus> just wanted to raise a quick request for ideas.  OSAS Infra offered Fedora Atomic some blade server hosting  ~~ 10 months ago.  The HW was delayed, but is finally ready.
16:07:48 <jberkus> however, the original plan for these doesn't seem applicable anymore
16:08:20 <jberkus> the blades need to be used for something which involves Fedora, Atomic, and Origin in some way
16:08:42 <jzb> jberkus: is that no longer useful?
16:08:58 <jberkus> jzb: well, the original goal was to get Origin working on Fedora Atomic
16:09:05 <jberkus> which it is
16:09:25 <jzb> jberkus: well, get it working + keep it working :-)
16:09:38 <jberkus> yah.  so test cluster is a possible use
16:09:42 <jbrooks> I'm still interested in that
16:09:43 <jzb> IIRC the proposal/idea was to have a testbed for Atomic + Origin + containers
16:09:52 <jzb> where contributors could also test things out
16:09:59 <ashcrow> Test cluster makes sense to me
16:10:25 <jberkus> the question is, who's going to operate it?
16:10:33 <dustymabe> I just don't think we're going to really use it
16:10:45 <dwalsh> Adding support for CRI-O and other stuff would be helpful.  We would love to be able to see Origin running on CRI-O.
16:10:48 <dustymabe> and when we do it won't be updated
16:11:26 <walters> give to centos ci?
16:11:34 <dustymabe> jberkus: do you have any numbers on how many of these things we have?
16:11:45 <walters> (and our jobs get preferred access to it?)
16:12:25 <jberkus> dustymabe: hmmm... lemme nudge the team about stats
16:12:39 <jberkus> I had them once upon a time but that's buried in 10-mo old email
16:12:42 <jbrooks> walters, does centos ci live in that same cage?
16:13:42 <walters> jbrooks, no idea, if there's vague agreement about this let's loop in the cci guys (bstinson etc)
16:14:36 <walters> this is a different path from "test cluster" which I think people mean like "has DNS and a way for people to use kubectl/oc and run apps"
16:14:37 <jberkus> I don't think there's an agreement about this yet
16:15:10 <dustymabe> yeah
16:15:24 <dustymabe> one other thing to note is that centos ci is hardware rich, where other parts of our chain are not
16:15:40 <jbrooks> I liked the idea of the FOSP
16:15:44 <jbrooks> Or FOOT
16:15:46 <jbrooks> heh
16:16:00 <ashcrow> dustymabe: What is an area that isn't HW rich that could use more power?
16:16:00 <jbrooks> I thought it was on hold till hw, I'm not clear on when it died
16:16:06 <dustymabe> jbrooks: yeah, i just don't have the time to maintain it
16:16:10 <jberkus> anyway, I didn't necessarily want to brainstorm this live
16:16:17 <dustymabe> ashcrow: basically anything in Fedora
16:16:19 <jbrooks> Was that the plan, that dustymabe was going to maintain it
16:16:32 <jberkus> just say: look for an email about this and please respond to it if you care about HW
16:16:32 <dustymabe> jbrooks: i think it was the collective "we"
16:16:43 <ashcrow> jberkus: sounds good
16:17:04 <jberkus> #topic upcoming events
16:18:07 <jberkus> first, some planning: for 2018/2019, I'm going to send around a worksheet for the various Atomic teams to nominate and rate events based on how much they care about them
16:19:14 <jberkus> particularly, it would be good to hear from contributors on events there's no Project Atomic presence at that we *ought* to go to
16:19:31 <jberkus> ok
16:19:34 <jberkus> specific events
16:19:46 <jberkus> DockerCon EU is right now, and Runcom is there
16:20:11 <jberkus> OSS Prague: I know two members of the Cockpit team are there, is anyone else?
16:21:12 <dustymabe> jberkus: not sure
16:22:16 <jberkus> ok, I'll take that as a no ... if you know of other project atomic folks going (surely there are some from Brno?) please let me know after the meeting
16:23:16 <jberkus> KubeCon: we have a bunch of folks there.  Goal for this conference is to have a ton of booth demos.  If you or a member of your team are going and have not contacted me about demos, please do so, we want to get you scheduled.
16:23:47 <jberkus> ATO: Charlie Drage is going, anyone else?
16:23:55 <dustymabe> i'll be at ATO
16:24:10 <dustymabe> since it's in raleigh
16:24:15 <jberkus> yah
16:24:32 <jberkus> there;s a booth
16:24:52 <jberkus> well, a spot -- for Atomic stuff
16:25:17 <jberkus> ok, what events am I missing in Oct/Nov/Dec?
16:26:08 <jberkus> anything?
16:26:21 <dustymabe> jberkus: holidays holidays holidays :)
16:26:34 <jberkus> yah, but there's no atomic community booth for Hanukkah
16:27:07 <jberkus> OK: DevConf/FOSDEM
16:27:36 <jberkus> lemme get FOSDEM out of the way: the Containers Devroom at FOSDEM is being run by some other org, Atomic is not involved
16:27:41 <jberkus> don't know who at this stage
16:27:41 <walters> i'll be at devconf, still need to submit a talk
16:28:23 <jberkus> we don't have a specific presence planned for FOSDEM itself, but the CentOS team has suggested that a full CentOS Atomic track at the CentOS Dojo in Brussels would be welcome
16:28:28 <dustymabe> FYI: I created a ticket to track submissions for devconf: https://pagure.io/atomic-wg/issue/351
16:28:35 <jberkus> dustymabe: great!
16:28:42 <ashcrow> dustymabe++
16:29:27 <jberkus> for DevConf, we'll want talks, a couple of BOFs, and we'll probably have half a table
16:29:57 <jligon> .hello jligon
16:29:59 <zodbot> jligon: jligon 'Jeff Ligon' <jligon@redhat.com>
16:30:11 <jberkus> so if you're going to DevConf I'm going to hit you up for a couple hours of table duty
16:30:36 <jberkus> anyway, please track your talk submissions on Dusty's ticket, even if those submissions aren't about Fedora Atomic
16:30:59 <jberkus> anything I'm missing on events?
16:31:58 <jberkus> ok, next topic
16:32:10 <jberkus> #topic CRI-O 1.0 release
16:32:31 <jberkus> so first contratulations to dwalsh runcom etc.
16:32:58 <jzb> dwalsh: +1
16:33:00 <jzb> runcom: +1
16:33:06 <dustymabe> congrats!
16:33:07 <ashcrow> Great work!
16:33:08 <jberkus> please echo the release announcement on social media if you do that kind of thing
16:33:22 <jberkus> https://medium.com/cri-o/cri-o-1-0-is-here-d06b97b92a98
16:33:49 <jberkus> dwalsh: anything special planned for CRI-O release?
16:35:05 <jberkus> dwalsh: ?
16:35:13 <jberkus> ok, next topic then.
16:35:21 <jberkus> #topic open floor
16:35:36 <jzb> quick question
16:35:40 <dwalsh> jberkus, Nothing. Sorry stayed away.
16:35:44 <jberkus> additional items?  since email ate my call for agenda items
16:35:50 <dwalsh> Would like to get CRI-O Containers out.
16:35:56 <dwalsh> Sorry stepped away.
16:35:59 <jberkus> aha
16:35:59 <jzb> jbrooks: I think you are/were working on a cri-o post w/Origin ?
16:36:05 <jberkus> #topic back to CRI-O
16:36:24 <jbrooks> jzb, yes I am
16:36:30 <jberkus> dwalsh: CRI-O images are the same format as OCI, no?
16:36:37 <dustymabe> dwalsh: can we get some of the containers peeps working more in the atomic working group
16:36:38 <jbrooks> I'm working on a post using the cri-o system containers
16:36:38 <dwalsh> Right, we would like to see about getting origin to default to CRI-O, which is the way we hope to see OpenShift going in the future.
16:36:46 <jberkus> dustymabe: +1
16:36:47 <jbrooks> With kube and origin
16:36:54 <dwalsh> jberkus, Yes CRI-O is a runtime that runs OCI Images.
16:36:59 * gholms raises an eyebrow
16:37:13 <dustymabe> gholms: body language?
16:37:22 <dwalsh> dustymabe, Sure they should have more time freeing up now that we have a release out the door.
16:37:23 <jberkus> dwalsh: is there any reason why existing containers in FLIBS and CP would not work under CRI-O?  Has anyone tested?
16:37:38 <jbrooks> It should work
16:37:47 <dwalsh> Next major release is for kuberneted 1.8 (CRI-O 1.8) which should be what Origin is working with.
16:37:48 <jbrooks> The guestbook app I always use to test works fine
16:38:27 <dwalsh> jberkus, No reason at all they should not work.  If they work with origin/kubernetes/docker they should work with origin/kubernetes/crio
16:38:33 <jzb> excellent
16:38:33 <dwalsh> Only better :^)
16:38:44 <tsweeney> no BFD?
16:38:45 <jberkus> ;-)
16:39:32 <jbrooks> We need crio 1.0 in koji and in centos' cbs
16:39:35 <jberkus> dwalsh: I think the FAH team could use help bundling CRI-O for FAH.  for one thing, we need to decide installed vs. system container
16:40:08 <jbrooks> I think messagingwise, we should be baking it in
16:40:29 <jbrooks> Otherwise, we suggest that docker ought to be the default
16:40:30 <dwalsh> Well I would love to see AOH to move away from any of these daemons installed by default.  I had customers begging me for that last week in the three letter agencies in DC.
16:40:53 <jbrooks> Well that's another option, bake in none
16:40:59 <jberkus> jbrooks: I thought the plan was to move docker to syscontainer
16:41:14 <jberkus> jbrooks: esp. because of 1.12 vs. 1.13+ issues
16:41:19 <jbrooks> In atomic 28?
16:41:23 <dustymabe> i really don't think 'bake in none' is a great option for new users who are learning
16:41:24 <dwalsh> I believe AOH should become Kernel/Init/Logging and everything else is added, but sshd is tough to get rid of
16:41:44 <jberkus> dwalsh: well, we need systemd
16:41:44 <jbrooks> runc would need to be there
16:41:50 <jbrooks> right?
16:41:56 <dwalsh> jberkus, systemd==init
16:41:56 <jberkus> yes
16:42:11 <dwalsh> yes runc is required.  I am talking about daemons.
16:42:23 <jbrooks> I have concerns about FLIBS
16:42:28 <jbrooks> about relying on it
16:42:39 <jbrooks> It needs to get more automated
16:42:45 <jbrooks> More predictable
16:43:12 <jberkus> jbrooks: OK.  can you put that on the agenda for Wednesday?
16:43:23 <jbrooks> jberkus, Yes
16:44:41 <jberkus> dustymabe: that's a useful argument to have.  of course, one could say the same about not baking in kube/Origin/etc.  where should we hash out "new user experience" issues?
16:45:02 <walters> dwalsh, there's also the linkage of selinux-policy ➭ container-selinux ➭ [container runtime]
16:45:36 <dwalsh> Sure container-selinux needs to be installed by default.  But we should do that as a dependency of runc.
16:45:42 <ashcrow> dwalsh: +1
16:45:52 <dustymabe> jberkus: typically with kube/origin the user would use some sort of 'installer'
16:46:04 <dustymabe> which will handle where to grab the software
16:46:31 <dustymabe> for stand-alone container cases i think we should have something for that
16:46:41 <jberkus> dustymabe: right, but I don't think we can hash this out during this meeting.  do you want to have another meeting about this, or a ticket discussion?
16:46:54 <dustymabe> jberkus: sure. i was just mentioning it
16:47:46 <jberkus> dustymabe: well, it's something we need to settle also for the CRI-O integration
16:48:04 <jberkus> because if docker is built-in, then CRI-O should be also
16:48:08 <jberkus> and contrawise
16:48:17 <jberkus> dustymabe: let's open a ticket
16:48:53 <jberkus> hey, given that we're running down the clock, did anyone have a different topic they were waiting to introduce?
16:49:50 <jberkus> #action jberkus to open a ticket about container runtime inclusion
16:50:18 <jberkus> dwalsh: if you and/or another member of the CRI-O team could attend Wednesday's FAH meeting, it would help
16:50:53 <dwalsh> Time?
16:51:25 <jberkus> 1pm EDT
16:51:30 <dustymabe> dwalsh: 1pm eastern, though we are working on a new time
16:51:53 <jbrooks> not for this week, though
16:51:59 <dustymabe> correct
16:53:20 <jberkus> as in, the time will not change this week.  it's 1pm EDT
16:53:33 <jberkus> ok, any other items?
16:53:56 <jberkus> ok, thanks all!
16:54:00 <jberkus> #endmeeting