17:00:28 <dustymabe> #startmeeting fedora_atomic_wg
17:00:29 <zodbot> Meeting started Wed Aug  9 17:00:28 2017 UTC.  The chair is dustymabe. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:29 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:00:29 <zodbot> The meeting name has been set to 'fedora_atomic_wg'
17:00:33 <maxamillion> .hello maxamillion
17:00:34 <dustymabe> #topic roll call
17:00:34 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
17:00:40 <dustymabe> .hello dustymabe
17:00:41 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
17:00:51 <ksinny> .hello sinnykumari
17:00:52 <zodbot> ksinny: sinnykumari 'Sinny Kumari' <ksinny@gmail.com>
17:00:55 <kushal> .hellomynameis kushal
17:00:56 <jbrooks> .hello jasonbrooks
17:00:57 <zodbot> kushal: kushal 'Kushal Das' <mail@kushaldas.in>
17:01:00 <zodbot> jbrooks: jasonbrooks 'Jason Brooks' <JBROOKS@REDHAT.COM>
17:01:58 <bowlofeggs> .hello bowlofeggs
17:02:03 <zodbot> bowlofeggs: bowlofeggs 'Randy Barlow' <randy@electronsweatshop.com>
17:02:34 <_ari_> .hello dustymabe
17:02:36 <zodbot> _ari_: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
17:03:45 <_ari_> .hello _ari_
17:03:46 <zodbot> _ari_: Sorry, but you don't exist
17:04:00 <dustymabe> #chair kushal maxamillion ksinny jbrooks _ari_ bowlofeggs
17:04:00 <zodbot> Current chairs: _ari_ bowlofeggs dustymabe jbrooks ksinny kushal maxamillion
17:04:11 <dustymabe> #chair nzwulfin
17:04:11 <zodbot> Current chairs: _ari_ bowlofeggs dustymabe jbrooks ksinny kushal maxamillion nzwulfin
17:04:18 <dustymabe> nzwulfin: :)
17:04:26 <dustymabe> #topic previous meeting action items
17:04:29 * nzwulfin waves
17:04:39 <dustymabe> lets start with my action items
17:04:46 <dustymabe> * dustymabe to create ticket to track RFE for rawhide based containers
17:04:49 <dustymabe> to be made available from registry.fedoraproject.org
17:04:52 <dustymabe> #info dusty created https://pagure.io/atomic-wg/issue/311 to track RFE for rawhide based containers to be made available from registry.fedoraproject.org
17:04:58 <dustymabe> * dustymabe to create a ticket for experimenting with modularity -
17:05:01 <dustymabe> where we can track progress on an atomic host module POC
17:05:04 <dustymabe> #info dusty created https://pagure.io/atomic-wg/issue/312 for experimenting with modularity
17:05:13 <dustymabe> * jbrooks dustymabe to come up with list of questions for modularity
17:05:13 <dustymabe> folks and invite them to our atomic weekly meeting in 1 or 2 weeks
17:05:15 <dustymabe> * dustymabe to create a ticket for questions for them for us to curate
17:05:17 <dustymabe> with a meeting tag
17:05:24 <dustymabe> #info dusty created https://pagure.io/atomic-wg/issue/313 for open questions for modularity team and jbrooks is helping create a list of questions
17:05:51 <dustymabe> any questions?
17:05:57 <dustymabe> #chair sayan
17:05:57 <zodbot> Current chairs: _ari_ bowlofeggs dustymabe jbrooks ksinny kushal maxamillion nzwulfin sayan
17:05:59 <sayan> .hello sayanchowdhury
17:06:03 <zodbot> sayan: sayanchowdhury 'Sayan Chowdhury' <sayan.chowdhury2012@gmail.com>
17:06:24 <dustymabe> ok other people action items:
17:06:32 <maxamillion> dustymabe: nope, all looks good. thanks for taking on all of that
17:06:32 <dustymabe> * jberkus strigazi to continue moving kube issues to new kube-sig
17:06:34 <dustymabe> tracker
17:06:50 <dustymabe> i think jberkus is AFK today
17:06:57 <dustymabe> is strigazi around?
17:07:20 <dustymabe> looks like no
17:07:33 <dustymabe> #action jberkus strigazi to continue moving kube issues to new kube-sig tracker
17:07:58 <dustymabe> ok we don't have anything with the meeting tag but I do want to bring a few things up
17:08:06 <dustymabe> #topic questions for the modularity team
17:08:16 <dustymabe> #link https://pagure.io/atomic-wg/issue/313
17:08:35 <dustymabe> We want to invite the modularity team to join us next week for this meeting
17:09:00 <maxamillion> +
17:09:01 <maxamillion> +1
17:09:02 <dustymabe> we'd like to have some questions for them about building atomic host from a module
17:09:27 <dustymabe> do we want to discuss any other questions here at this time or can we come up with more questions for them in the ticket before next week
17:09:47 <dustymabe> the idea is to create a FAQ from the resulting discussion so future people can get answers to the same questions they might have
17:10:23 <dustymabe> also petr from the modularity team is working on a proof of concept for creating an atomic host out of a module (https://pagure.io/atomic-wg/issue/312#comment-451920)
17:10:38 <dustymabe> so he can possibly present that to us so that we can get a more concrete feel for what it all really means
17:10:54 <jbrooks> That'll be good
17:11:13 <jbrooks> Based on what I've been hearing, it seems that it'd be pretty straightforward
17:11:29 <dustymabe> jbrooks: yeah - only one way to find out :)
17:11:49 <jbrooks> I'm tempted to bust out w/ a test myself
17:12:02 <jbrooks> Sort of tempted
17:12:04 <jbrooks> heh
17:12:08 <dustymabe> jbrooks: go for it - then we would have a lot more questions to add to the issue
17:12:23 <dustymabe> always better for someone who has never done it before, because there will be many questions along the way
17:12:53 <dustymabe> jbrooks: want me to action you??? :)
17:13:00 <dustymabe> anyone else have any questions about this?
17:13:01 <jbrooks> dustymabe, sure
17:13:22 <dustymabe> #action jbrooks to look at building a module himself, this will help us come up with even more questions before our meeting next week
17:13:47 <dustymabe> #action dustymabe to invite petr sabata to the atomic host meeting next week to answer questions and to present a POC atomic host from a module
17:13:59 <ksinny> dustymabe: One question, does this effect only building AH using module or anythingelse as well in project Atomic?
17:14:48 <dustymabe> ksinny: mostly atomic host itself, because that is where all of the subcomponents are brought together
17:15:01 <ksinny> dustymabe: ok, thanks :)
17:15:10 <dustymabe> the other pieces of project atomic (docker skopeo atomic cli) will be included in the module
17:15:16 <dustymabe> and will be tested more by CI
17:15:24 <dustymabe> so possibly more detection of breakage
17:15:35 <ksinny> sounds good
17:15:54 <dustymabe> ok moving on to next topic
17:16:17 <maxamillion> mmmm CI
17:16:38 <dustymabe> #topic containerized kubernetes in f26 atomic host
17:16:56 <dustymabe> so we currently don't have kubernetes in our rawhide atomic host and I don't want to add it back
17:17:02 <dustymabe> for f27
17:17:09 <maxamillion> +1
17:17:12 <jbrooks> Yeah
17:17:15 <jbrooks> +1
17:17:20 <maxamillion> is k8s available as a system container?
17:17:24 <dustymabe> what steps do we need to take to make sure that happens
17:17:29 <dustymabe> maxamillion: yes
17:17:33 <maxamillion> rocking
17:17:34 <dustymabe> jbrooks: has worked very hard on that
17:17:38 <dustymabe> jbrooks++
17:17:39 <maxamillion> jbrooks++
17:17:39 <zodbot> maxamillion: Karma for jasonbrooks changed to 2 (for the f26 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:17:58 <jbrooks> The kube containers in f26 are usable now as system containers or regular ones
17:18:08 <maxamillion> awesome
17:18:11 <jbrooks> But there's also etcd
17:18:19 <maxamillion> ah yes
17:18:22 <jbrooks> And gluster/ceph/flannel?
17:18:42 <jbrooks> I think we should consider adding the containernetworking-cni package to the host
17:18:56 <maxamillion> I forget all that stuff is separate, openshift just kind of throws it all together
17:19:02 <jbrooks> It's an enabler for various different plugins
17:19:14 <dustymabe> so here is a question - what is needed by k8s that isn't needed by openshift?
17:19:17 <jbrooks> I should open an issue about that
17:19:24 <dustymabe> i.e. is all that stuff needed in the host for openshift too?
17:19:26 <jbrooks> etcd
17:19:31 <jbrooks> openshift rolls etcd in
17:19:45 <maxamillion> openshift rolls basically everything in that it can
17:19:51 <dustymabe> ok
17:19:55 <jbrooks> kube needs some binaries for networking
17:20:10 <maxamillion> the networking/SDN is rolled in, etcd is rolled in, docker registry, and some other bits
17:20:13 <dustymabe> so I guess we need to decide where it makes sense to draw the line for kube
17:20:14 <maxamillion> all rolled in
17:20:15 <jbrooks> We include flannel now, but we could include containernetworking-cni and enable flannel plus others through that
17:20:57 <dustymabe> jbrooks: if we're using a system container for kube, what can we put in there (or other system containers) vs in the host?
17:21:19 <maxamillion> I don't know, I'm not super well versed in "vanilla k8s" so I'm likely of little help
17:21:36 <dustymabe> maxamillion: was mostly asking jbrooks I think
17:21:43 <jbrooks> dustymabe, was have an etcd system container now
17:21:46 <maxamillion> +1
17:21:46 <jbrooks> in the registry
17:22:06 <jbrooks> You can use the system containers to copy, like, anything, onto the host
17:22:31 <jbrooks> I'm not sure about gluster / ceph
17:22:40 <jbrooks> We might want to think about leaving them in
17:22:58 <jbrooks> I think it'd be fine to take flannel out, but I'd like that cni pkg in
17:23:14 <jbrooks> So it'd be gluster, ceph, cni in the image, etcd and kube in system containers
17:23:38 <dustymabe> do people use the gluster stuff today?
17:24:01 <dustymabe> I guess that's a question for a different topic/day
17:24:04 <jbrooks> Probably
17:24:07 <kushal> :)
17:24:10 <jbrooks> I keep meaning to
17:24:17 <dustymabe> ok so jbrooks we have a bunch of tickets: https://pagure.io/atomic-wg/issues?status=Open&tags=remove-kube
17:24:25 <jbrooks> But I don't do much w/ persistent volumes at the moment
17:24:42 <jbrooks> dustymabe, I'll take an action item to go through those
17:24:44 <dustymabe> can you go through and update them
17:24:47 <dustymabe> ok
17:25:00 <dustymabe> #action jbrooks to go through 'remove-kube' tickets and update them
17:25:18 <dustymabe> jbrooks: also we need some sort of strategy for when we do remove kube - what to tell people
17:25:32 <dustymabe> if we need to tell them to re-install, that's fine
17:25:38 <dustymabe> just as long as we have a clear message
17:25:47 <dustymabe> and the earlier we know that the more we can talk about it
17:25:51 <dustymabe> on the blog
17:26:07 <jbrooks> Yeah, I can work on that -- it'll basically be -- before you upgrade, run these system containers install commands
17:26:16 <jbrooks> And then it'll just work
17:26:24 <dustymabe> wow - nice
17:26:25 <jbrooks> These are drop-in replacements
17:26:29 <dustymabe> have you tested that all out?
17:26:45 <jbrooks> parts of it
17:26:55 <jbrooks> I'll test it all -- they bind mount the conf dirs
17:27:11 <jbrooks> And they override the systemd units by going into /etc
17:27:17 <dustymabe> so the ultimate test is to get a cluster up and running on f26 and then run your commands and then rebase to rawhide
17:27:24 <jbrooks> So they work the same as the built-in binaries
17:27:32 <dustymabe> since rawhide doesn't have k8s
17:27:36 <jbrooks> Yeah
17:27:54 <jbrooks> Did we do our container release?
17:28:07 <maxamillion> no
17:28:10 <maxamillion> I haven't had time
17:28:13 <dustymabe> ok next topic
17:28:17 <dustymabe> #topic open floor
17:28:24 <maxamillion> it's about a full day of pain for me to get a container release out the door right now
17:28:30 <dustymabe> I have one item
17:28:45 <maxamillion> dustymabe: shoot
17:29:24 <dustymabe> FYI: How Do We Create OSTree Repos and Artifacts in Fedora
17:29:30 <dustymabe> #link https://lists.projectatomic.io/projectatomic-archives/atomic-devel/2017-August/msg00025.html
17:29:42 <dustymabe> might be some cool bedtime reading
17:30:01 * ksinny marks it for reading
17:30:14 <kushal> Do we have a ticket which has all the plans for Flock (from Atomic WG)?
17:30:19 <dustymabe> ksinny: yes! - that should be useful to learn how this beast works right now
17:30:39 <maxamillion> dustymabe: +1
17:30:43 <dustymabe> kushal: I don't think we have a "plan" ticket - we did have the original ticket for submitting flock talks
17:30:48 <ksinny> dustymabe: yeah, I am interested in understanding it
17:31:02 <kushal> dustymabe, maybe we create a plan ticket now.
17:31:18 <dustymabe> kushal: maybe - feel free to do that
17:31:23 <kushal> dustymabe, Okay
17:31:26 <dustymabe> i typically lean on josh for some of that stuff
17:31:40 <dustymabe> #action kushal to create ticket for planning our activities around FLOCK
17:31:45 <maxamillion> +1
17:31:46 <kushal> ah, you are fast
17:31:49 <dustymabe> ok maxamillion you might be able to help me with this
17:31:52 <dustymabe> https://lists.projectatomic.io/projectatomic-archives/atomic-devel/2017-August/thread.html
17:31:54 * kushal was typing in one hand anyway :)
17:32:02 <maxamillion> dustymabe: ?
17:32:15 <dustymabe> I don't see the two week atomic release announcement on that page
17:32:46 <dustymabe> i went in and added noreply@fedoraproject.org as a subscriber last week
17:33:20 <maxamillion> dustymabe: we'll need to ask fedora infra then, I'm not sure why it's not going out ... the email makes it to the other lists just fine
17:33:33 <dustymabe> yeah - i suspect it's on the pa.io side
17:33:38 <dustymabe> it's just freakin weird though
17:33:42 <maxamillion> dustymabe: it is
17:33:58 <dustymabe> and also means we are not reaching our primary audience :(
17:34:06 <dustymabe> at least my re: makes it to the list
17:34:12 <maxamillion> dustymabe: that being said, it's been 6+ years since I admin'd a mail server in any serious capacity so I'm not sure
17:34:30 <maxamillion> we have differing opinions about who our primary audience is, but that's fine
17:34:51 <dustymabe> maxamillion: oh?
17:35:09 <maxamillion> we make things for the Fedora Community do we not?
17:35:27 <maxamillion> actually, it doesn't matter ... let's move on
17:35:31 <dustymabe> sure, but we just decided to use atomic-devel as our main form of communication
17:35:40 <maxamillion> yeah, we did
17:35:45 <maxamillion> I don't agree with it, but that's fine
17:35:50 <maxamillion> we can move on
17:35:56 <dustymabe> i actually think we should add fedora devel list to the release announcement
17:36:06 <dustymabe> it's significant enough now
17:36:14 <maxamillion> sounds good
17:36:30 <dustymabe> ok any other open floor items?
17:36:40 <dustymabe> other than the fact that we miss roshi this week :(
17:36:42 <dustymabe> :tear:
17:36:48 <maxamillion> +1
17:37:03 <maxamillion> I hope to get a container release out tomorrow
17:37:16 <dustymabe> #info maxamillion hopes to get a container release out tomorrow
17:37:20 <maxamillion> as soon as the new architecture for osbs stuff is done, I'll be improving that release process so it's not quite as painful
17:37:21 <jbrooks> cool
17:37:25 <dustymabe> maxamillion: how close is the automation stuff?
17:37:31 <maxamillion> dustymabe: it's not at all
17:37:32 <dustymabe> can you see the light at the end of the tunnel
17:37:34 <dustymabe> :(
17:37:36 <maxamillion> dustymabe: still waiting on various upstreams
17:37:41 <dustymabe> roger
17:37:48 <dustymabe> ok anything else before we close?
17:37:55 <maxamillion> not here
17:38:45 <maxamillion> dustymabe: thanks for hosting!
17:39:09 <dustymabe> np
17:39:11 <dustymabe> #endmeeting