16:05:00 <jberkus> #startmeeting atomic-general-community
16:05:00 <zodbot> Meeting started Mon Dec 11 16:05:00 2017 UTC.  The chair is jberkus. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:05:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:05:00 <zodbot> The meeting name has been set to 'atomic-general-community'
16:05:00 <centbot> Meeting started Mon Dec 11 16:05:00 2017 UTC.  The chair is jberkus. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:05:00 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:05:12 <jberkus> #topic roll call
16:05:16 <jberkus> .hello jberkus
16:05:17 <zodbot> jberkus: jberkus 'Josh Berkus' <josh@agliodbs.com>
16:05:20 <dustymabe> .hello dustymabe
16:05:22 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
16:05:26 <walters> .hello walters
16:05:27 <zodbot> walters: walters 'Colin Walters' <walters@redhat.com>
16:05:46 <ashcrow> .hello
16:05:46 <zodbot> ashcrow: (hello <an alias, 1 argument>) -- Alias for "hellomynameis $1".
16:05:51 <ashcrow> .hello smilner
16:05:52 <zodbot> ashcrow: smilner 'None' <smilner@redhat.com>
16:05:55 <dustymabe> maybe runcom or lsm5 might be around
16:05:59 <dustymabe> #containers
16:06:01 <sanja[m]> .hello sanja
16:06:04 <zodbot> sanja[m]: sanja 'Sanja Bonic' <sanja@redhat.com>
16:06:31 <giuseppe> .hello gscrivano
16:06:35 <zodbot> giuseppe: gscrivano 'Giuseppe Scrivano' <gscrivan@redhat.com>
16:06:36 <runcom> .hello runcom
16:06:38 <zodbot> runcom: runcom 'Antonio Murdaca' <amurdaca@redhat.com>
16:06:44 <dustymabe> \o/
16:06:45 <rubao> .hello rubao
16:06:46 <zodbot> rubao: rubao 'rubao' <rubao.net@hotmail.com>
16:06:51 <jbrooks> .hello jasonbrooks
16:06:53 <zodbot> jbrooks: jasonbrooks 'Jason Brooks' <jbrooks@redhat.com>
16:07:06 <dustymabe> nice crowd
16:07:09 <jberkus> yah
16:07:41 <whenry> .hello whenry
16:07:42 <zodbot> whenry: whenry 'William Henry' <william_henry@me.com>
16:07:44 <jberkus> howdy, everyone!  this is the last atomic-general meeting of the year, since I don't think anyone wants a meeting on christmas day
16:08:04 <ashcrow> :-D
16:08:05 <whenry> +1 :)
16:08:09 <jberkus> I totally failed to call for agenda items because I was at kubecon, so if you have one, please PM me so that I can sort them
16:08:18 <whenry> unless there are presents.
16:08:38 <jlebon> .hello jlebon
16:08:39 <zodbot> jlebon: jlebon 'None' <jlebon@redhat.com>
16:09:05 <jberkus> #topic newer versions of Docker for Fedora / FAH
16:09:08 <jberkus> dustymabe: ?
16:09:36 <dustymabe> jberkus: i've been asked no fewer than 5 times in the past week about newer versions of docker and Atomic HOst
16:09:53 <runcom> dustymabe: jberkus I'm afraid we don't have an answer for that yet
16:09:56 <dustymabe> I really don't know what our plans are there, but would like to have a good answer for the community on that front
16:09:58 <jberkus> what's available and newer?  is there something newer than Docker CE 17?
16:10:10 <jbrooks> I think that's newest
16:10:18 <jberkus> jbrooks has a system container for docker ce 17, I think?
16:10:26 <jbrooks> Sort of
16:10:46 <jbrooks> At kubecon I quickly hacked up our docker sys container to make it run docker-ce
16:10:50 <dustymabe> right, there are some newer features that people are interested in, but I think mostly they just look and see 'oh that's xx months old, let's go use something else'
16:11:05 <jbrooks> I just swapped the repo and tore out some stuff unique to us
16:11:08 <dustymabe> so for a brand new user evaluating the different options it could make a difference
16:11:42 <jbrooks> The package layering route is pretty smooth too
16:12:10 <jbrooks> I like the idea of offering more than one ref for things like this, too
16:12:19 <jbrooks> Depending on how burdensome that is
16:12:26 <jbrooks> But those are three different options
16:12:30 <giuseppe> does Docker ce 17 require a new client as well?
16:12:39 <dustymabe> jbrooks: yeah, i guess I just want to know what our plans are. we typically don't have these problems because we follow upstream pretty close
16:13:06 <dustymabe> and when I say upstream I'm referring to all upstreams in Fedora land
16:13:16 <dustymabe> not a particular upstream
16:13:36 <jberkus> jbrooks: when we use a system container to overide the built in docker, how is that in terms of user experience?
16:13:37 <giuseppe> we should start probably copying /usr/bin/docker from the system container to the host (/usr/local/bin/docker) so that the same docker client can be used from the host
16:13:40 <jbrooks> Do we have ppl from the team that maintains the PA branched docker here?
16:13:53 <jbrooks> jberkus, I didn't do more than "docker info"
16:13:55 <dustymabe> jbrooks: yes, runcom
16:14:05 <jbrooks> but it was install, docker restart, docker info, it was updated
16:14:16 <jbrooks> And the reverse reversed it
16:15:07 <jberkus> jbrooks: ok, so more testing required
16:15:16 <dustymabe> so summary of problem: this hasn't been a problem in the past because we've been pretty close to upstream, but now it's getting a little old and people are starting to ask and it's going to get worse
16:15:39 <dustymabe> so i want to press on us "having a plan" here
16:15:51 <walters> this is all related to https://pagure.io/atomic-wg/issue/360
16:15:54 <jberkus> dustymabe: yes, especially because when CE 18 comes out, it's be tied to a specific Kubernetes version, which won't work for OpenShift
16:16:10 <ashcrow> walters: agreed
16:16:31 <dustymabe> walters: yes and no. we're still going to have the default of what is "in fedora"
16:16:32 <jberkus> related to this: docker survey results later this week, since I'll have time to do some analysis
16:17:21 <ashcrow> jberkus: sounds good
16:17:33 <jbrooks> Is strigazi here?
16:18:08 <jbrooks> I think he's said he's done something w/ docker in a system container, too
16:18:26 <dustymabe> jbrooks: yeah they are using docker in a system container I believe
16:18:57 <jbrooks> dustymabe, Because I'm seeing now that my docker-ce container isn't working quite right -- may be related to the client part
16:19:22 <dustymabe> so.. long story short, who can take an action item on this front
16:19:23 <jbrooks> Anyway, we should at least document the ways one can replace the built-in docker
16:20:01 <sanja[m]> seems like something I can do
16:20:13 <ashcrow> sanja[m]++
16:20:39 <walters> `override` finally being stable helps with that
16:20:48 <dustymabe> sanja[m]: with guidance from the rest of us
16:20:50 <walters> (when it also percolates to all the variants)
16:20:53 <jbrooks> sanja[m], this can help some https://jebpages.com/2017/11/08/install-docker-ce-on-fedora-atomic-host-if-thats-what-youre-into/
16:21:37 <sanja[m]> thanks, everyone - you can expect questions :D
16:22:41 <dustymabe> should probably create a tracker issue for this in pagure.io/atomic-wg
16:22:54 <dustymabe> we really need engagement from the container team on this one, though
16:23:11 <runcom> +1 yeah
16:23:32 <jberkus> "container team"?
16:23:33 <dustymabe> ashcrow: do you think you can help sanja[m] lead this discussion?
16:23:51 <dustymabe> jberkus: yeah, people that contribute to projectatomic/docker
16:23:55 <dustymabe> and others
16:24:23 <jberkus> ah, ok.
16:24:25 <ashcrow> dustymabe: absolutely
16:25:27 <ashcrow> dustymabe: I'll create the tracker ticket for it as well
16:25:49 <sanja[m]> I'll write you a separate email then afterwards, ashcrow
16:25:53 <ashcrow> sanja[m]: k
16:25:59 <dustymabe> sanja++ ashcrow++
16:25:59 <zodbot> dustymabe: Karma for sanja changed to 1 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:26:09 <dustymabe> sanja[m]: FYI you just got a cookie!
16:26:44 <jberkus> ok, can we do more on this in this meeting?
16:26:55 <dustymabe> jberkus: probably not
16:27:10 <jberkus> ok
16:27:11 <whenry> jberkus: are you looking for someone?
16:27:23 <jberkus> #topic open floor
16:28:03 <jberkus> walters: you had a new tool on the list?
16:28:03 <dustymabe> jberkus: summary of kubecon? any key takeaways for pa?
16:28:49 <walters> jberkus:  I'd just note that jigdo ♲📦 is going well on the code side (more info at https://lists.projectatomic.io/projectatomic-archives/atomic-devel/2017-December/msg00003.html ) though at the moment there hasn't been a lot of releng feedback
16:29:36 <walters> but i think it'll help a lot and then we can focus more on content issues like we were just discussing
16:30:48 <ashcrow> I'm finishing up the https://github.com/ashcrow/image-helpgen/ tool this week. Some clean up before releasing.
16:31:09 <jberkus> ashcrow++
16:31:30 <jbrooks> We had a very minor monthly cah release: https://seven.centos.org/2017/12/centos-atomic-host-7-1711-available-for-download/
16:31:45 <jberkus> jbrooks: anything special in it?
16:32:08 <jbrooks> No, just a few bugfixes from the prior month
16:32:14 <jbrooks> No changes to any core components
16:33:08 <dustymabe> jberkus: you maybe missed this:
16:33:10 <dustymabe> jberkus: summary of kubecon? any key takeaways for pa?
16:33:26 <ashcrow> jberkus: (after dustymabe's question) can we get sanja[m] added to atomic-wg group?
16:33:27 <jberkus> dustymabe: folks were talking about other things
16:33:33 <jbrooks> dustymabe, Did we have a container release on the last atomic release day?
16:33:39 <dustymabe> jberkus: +1
16:33:43 <jberkus> ashcrow: oh!  I'll take care of that right away
16:34:05 <dustymabe> jbrooks: I don't know actually
16:34:24 <dustymabe> I feel like there isn't very good ownership of the container releases in Fedora land right now
16:34:40 <dustymabe> i've intentionally tried to focus on FAH, and jbrooks has CAH
16:34:48 <dustymabe> could use some leadership in that area
16:35:01 <walters> i'd be quite interested in a rollup post about post-kubecon blogs/notes or so
16:35:18 <whenry> +q
16:35:36 <jberkus> so, some takeaways from Kubecon: I didn't spend a lot of time talking about atomic host there, so I don't have a lot of feedback on it.  The one thing I'll say is that there's a lot of interest in availability on ARM, although how much of that is gadget factor I don't know
16:35:52 <jberkus> jdetiber now has the overdrive 1000 and has committed to testing new AH releases on ARM
16:36:45 <dustymabe> jberkus: so people were just interested in talking about pure kube/openshift?
16:36:51 <jberkus> for other Atomic projects: CRI-O was big and well-accepted there, with a busy talk and BOF.  But what was most popular about CRI-O was the knit caps we gave away, since it snowed in Austin
16:37:16 <whenry> :)
16:37:41 <jberkus> dustymabe: more than anything, folks at Kubecon largely don't care about the "lower layers" of the stack from their perspective.  So while there's some general interest in Atomic Host et. al., there's not a lot of specific interest
16:37:50 <jberkus> in any features.
16:37:52 <jbrooks> Our booth was totally openshift-focused, so, that directed a lot of the focus
16:38:12 <jberkus> I did not get asked to compare it with CoreOS, which I did a lot last year, but I don't know that that change means
16:38:29 <jbrooks> I had one person ask me if atomic was still a thing, since he didn't hear mention of it at the openshift commons gathering
16:38:57 <jberkus> jbrooks: yah, that seems like an omission, I'm sure unintentional
16:39:28 <jberkus> since I may be helping run OpenShift Commons in copenhagen, I can make sure it's mentioned
16:39:33 <whenry> jberkus: +q  should I add that new Buildah tutorial to the atomic/blog
16:39:40 <jberkus> whenry: yes, please!
16:39:46 <whenry> ack
16:39:58 <jberkus> whenry: is anyone working on integrating buildah with ansible-container?
16:40:17 <whenry> jberkus: will try to do today. Also going to start on libkpod
16:40:41 <whenry> jberkus: that was re blog and not re integrating buildah with ansible container :D
16:40:47 <jberkus> whenry: CfP for KubeCon.EU is open, hint, hint
16:41:23 <whenry> jberkus: +1 I will do the CFP today too. Buildah or Kpod?
16:41:49 <whenry> jberkus: and I don't know about buildah and ansible-container. I'll look into that.
16:42:07 <jberkus> I'm not sure if KubeVirt is considered an Atomic project or not, but we had a big focus on that at KubeCon, including a virtualization salon where we tried to get all of the VM-on-kube projects into a joint effort
16:42:22 <jberkus> whenry: I suggest one of each
16:42:28 <whenry> ack
16:43:16 * whenry is dropping off now. couple of things to do before next meeting. bbl
16:43:30 <cverna> jberkus: I think ttomecek was looking at buildah and ansible-container
16:43:57 <cverna> https://github.com/ansible/ansible-container/pull/790
16:43:58 <jberkus> finally, I had an interesting conversation with a member of the Kubernetes Dashboard team.  Apparently they're having issues because their webapp code doesn't really support having optional Kube modules extend the UI
16:44:16 <jberkus> so I suggested that they take a look at Cockpit, which they took pretty seriously
16:44:25 <jberkus> stefw, petervo ^^^
16:44:38 <jberkus> cverna: thanks!
16:44:55 <jberkus> that's it from me.  jbrooks, you were doing system containers demo.  anything come out of that?
16:45:33 <jbrooks> Not really, the people I demoed to were interested, but they weren't really familiar w/ atomic host to begin with
16:48:15 <jberkus> yah, getting in a session in Kubecon on Atomic Host for Kubernetes would be good for AH, but teh Kubecon committee doesn't seem interested
16:48:40 <ttomecek> cverna, still working on it, but getting closer
16:48:57 <jbrooks> There were a ton of non-kube sessions, too, they have a lot to fit in w/ the while cncf slate of projcets
16:49:11 <dustymabe> jberkus: what is there interest level in coreOS? i.e. is it more than AH or same?
16:49:24 <dustymabe> their*
16:49:57 <jbrooks> I don't think I heard a peep about coreos container linux there
16:50:04 <jberkus> dustymabe: I heard very little conversation about CoreOS contianer linux, and didn't see anything on the schedule
16:50:15 <dustymabe> interesting
16:50:16 <sanja[m]> dustymabe: from my experience so far, coreOS much bigger interest in general, at least in Europe
16:50:19 <cverna> ttomecek++
16:50:20 <zodbot> cverna: Karma for ttomecek changed to 1 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:50:20 <jberkus> for that matter, Intel announced both Katacontianers *and* their own container linux
16:50:25 <sanja[m]> mostly because awareness is better and it's more present on cloud services
16:50:37 <jberkus> and there was a bunch about kata, and pretty much nothing about their linux
16:50:43 <dustymabe> sanja[m]: yeah we really need to get our stuff in more cloud providers
16:51:22 <jberkus> dustymabe: we also need to advertise the cloud providers we do have better
16:51:29 <jberkus> ok, anything else?
16:51:40 <jberkus> #endmeeting