16:30:41 <jberkus> #startmeeting fedora_atomic_wg
16:30:41 <zodbot> Meeting started Wed Nov  8 16:30:41 2017 UTC.  The chair is jberkus. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:30:41 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:30:41 <zodbot> The meeting name has been set to 'fedora_atomic_wg'
16:30:49 <jberkus> #topic roll call
16:30:54 <jberkus> .hello jberkus
16:30:56 <zodbot> jberkus: jberkus 'Josh Berkus' <josh@agliodbs.com>
16:31:00 <scollier> .hello scollier
16:31:01 <zodbot> scollier: scollier 'Scott Collier' <emailscottcollier@gmail.com>
16:31:04 <dustymabe> .hello2
16:31:04 <jbrooks> .fas jasonbrooks
16:31:06 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
16:31:08 <zodbot> jbrooks: jasonbrooks 'Jason Brooks' <JBROOKS@REDHAT.COM>
16:31:33 <miabbott> .hello miabbott
16:31:34 <zodbot> miabbott: miabbott 'Micah Abbott' <miabbott@redhat.com>
16:31:38 <ksinny> .hello sinnykumari
16:31:40 <zodbot> ksinny: sinnykumari 'Sinny Kumari' <ksinny@gmail.com>
16:31:56 * ttomecek lurks for a few minutes
16:33:51 <maxamillion> .hello2
16:33:52 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
16:34:03 <puiterwijk> .hello2
16:34:04 <zodbot> puiterwijk: puiterwijk 'Patrick "マルタインアンドレアス" Uiterwijk' <puiterwijk@redhat.com>
16:34:16 <jberkus> #chair ksinny jbrooks jberkus miabbott scollier maxamillion puiterwijk
16:34:16 <zodbot> Current chairs: jberkus jbrooks ksinny maxamillion miabbott puiterwijk scollier
16:35:11 <jberkus> hmmmph.  I thought this meeting time was supposed to be better for the Europeans & Indians?
16:35:38 <jberkus> anyway
16:35:40 <jbrooks> Just a half hour earlier
16:35:53 <dustymabe> jberkus: ksinny voted +1 for this time and sayan said he could make it work
16:36:12 <jberkus> #topic Fedora 27 Atomic schedule
16:36:12 <dustymabe> ttomecek: and eliksa were the other two that said this time wasn't great for them
16:36:31 <dustymabe> unfortunately these times are the ones that had the most votes so..
16:36:51 <jberkus> yah, it's 1.5 hours earlier for me but DST
16:36:56 <jberkus> anyway
16:37:20 <dustymabe> jberkus: previous meeting action items?
16:37:30 <jberkus> So F27 is supposed to release next week.  For that reason, I'm starting with that.
16:37:38 <dustymabe> k
16:38:15 <jberkus> what do we need to do for the release that we haven't done?
16:38:32 <dustymabe> jberkus: as a working group?
16:38:36 <jberkus> the blog posts are in process, and we need to figure out the schedule/placement for them
16:38:50 <dustymabe> jberkus: yes. if we could put those on a calendar that would be great
16:38:50 <jberkus> dustymabe: yah
16:38:57 <dustymabe> assuming we ship next week
16:38:58 <ksinny> +1
16:39:02 <jberkus> talking points are done
16:39:11 <dustymabe> if we don't ship next week then we just move the calendar we create by 1 week
16:39:17 <jberkus> dustymabe: if F27 ships next week, what day?
16:39:26 <dustymabe> tuesday
16:39:29 <dustymabe> always on a tuesday
16:39:33 <jberkus> would they really release the week of American Thanksgiving?
16:39:34 <dustymabe> for major releases
16:39:58 <jbrooks> That's not next week is it?
16:39:59 <dustymabe> jberkus: i think you are a week ahead
16:40:13 <dustymabe> turkey day is on the 23rd
16:40:14 <jberkus> if the release slips by a week
16:40:15 <davdunc> .hello davdunc
16:40:16 <zodbot> davdunc: davdunc 'David Duncan' <davdunc@amazon.com>
16:40:31 <dustymabe> ahh yes. that is right. if the release slips a week then it would be tday week
16:40:35 <dustymabe> and the answer is 'no'
16:40:39 <dustymabe> they wouldn't ship
16:40:44 <dustymabe> so it's important we don't miss this date
16:40:48 <jberkus> #chair ksinny jbrooks jberkus miabbott scollier maxamillion puiterwijk davdunc
16:40:48 <zodbot> Current chairs: davdunc jberkus jbrooks ksinny maxamillion miabbott puiterwijk scollier
16:41:00 <jberkus> so let's assume next week
16:41:03 <puiterwijk> jberkus: hint: you only need to #chair the new person
16:41:12 <dustymabe> :)
16:41:13 <puiterwijk> It will automatically retain all of the old chair's
16:41:30 <jberkus> question: of the current blog posts, should we try to push one of them onto the Fedora Community Blog?  If so, which one?
16:42:33 <dustymabe> jberkus: I know some of them are planned to be released on fed mag and atomic blog
16:43:09 <jberkus> dustymabe: we don't have any of them in process for fedora magazine
16:43:24 <dustymabe> define 'in process' ?
16:44:18 <jberkus> submitted to the magazine, being reviewed
16:44:35 <ksinny> I propose mine for getting published on Fedora Magazine as well https://lists.fedoraproject.org/archives/list/magazine@lists.fedoraproject.org/thread/66QZHX6BHQCE4WK6WLTIGVWPW2VAVANG/
16:44:47 <ksinny> Didn't receive any response so far
16:45:04 <dustymabe> yeah ksinny did propose hers and the one on azure can really publish whenever
16:46:02 <jberkus> cool.  ksinny, do you need any of the rest of us to do anything for that?
16:46:33 <jberkus> dustymabe: I proposed the feature round-up for the Fedora Community Blog, but I'm having 2nd thoughts, just because historically CommBlog has lagged 1-2 weeks behind
16:46:47 <ksinny> jberkus: I am good for now.
16:46:50 <dustymabe> jberkus: also it might be nice to publish "Article For F27: Fedora 27 Atomic Features" on fed mag instead of comm blog
16:47:01 <dustymabe> or maybe not
16:47:10 <dustymabe> i don't know how many people read comm blog vs fed mag
16:47:17 <ksinny> I will ping tomorrow on ML if I don't get any response
16:47:32 <dustymabe> i'm think fed mag is a more general audience and comm blog is more heavily fedora contributors
16:49:20 <jberkus> I'm pretty sure fedmag gets a lot more views
16:49:39 <jberkus> but ... the F27Atomic feature post isn't really magazine material
16:49:43 <maxamillion> yeah, fedmag does pretty good on the traffic front
16:49:50 <maxamillion> pretty well*
16:49:55 <maxamillion> words are hard yo
16:50:06 <dustymabe> ehh. I've done f2X test day posts on fedmag before
16:50:26 <dustymabe> which maybe not the place to do it, but I wouldn't worry too much about it being too good for our post :)
16:51:10 <puiterwijk> Just submit it, and the magazine editors will help you further
16:51:21 <jberkus> ok
16:51:49 <jberkus> #action jberkus to post target publication, calendar of posts later today to mailing list
16:51:58 <dustymabe> jberkus: thanks!
16:52:15 <jberkus> puiterwijk: one thing we *don't* have on the list of posts is an updated post about Atomic Workstation
16:52:26 <jberkus> which would be nice
16:52:31 <dustymabe> there might be some 'ordering issues' to work out because we want to link to some posts from others
16:52:40 <dustymabe> but we can resolve that when the schedule comes out
16:52:50 <puiterwijk> jberkus: fun fact: I'm staring at the F27AW installation right this moment
16:53:06 <puiterwijk> So let me try it out and then I'll see if I can use it to write a post :)
16:53:12 <jberkus> puiterwijk: thanks
16:53:37 <dustymabe> jberkus: do you want to add an issue for writing an AW post ?
16:53:47 <jberkus> yah
16:54:15 * dustymabe is certainly happy we seem to be more organized about 'communications' for f27 than we were for f26
16:54:20 <jberkus> btw, y'all will be happy to know that we're "officially" calling it Atomic Workstation now
16:54:36 <puiterwijk> jberkus: cool. good to know :)
16:54:38 <jberkus> mostly because nobody could think of a different name
16:54:53 <davdunc> jberkus: will that be reflected in the cloud image names?
16:54:57 <dustymabe> jberkus: i missed that discussion, but happy to hear we're going with a name that is pretty obvious
16:55:11 <dustymabe> davdunc: I don't think we plan to put out cloud images for that (at least right now)
16:55:18 <puiterwijk> So can I abbreviate it as F27AW and F27AH then?
16:55:22 <jberkus> note that, for the time being, our downstream has no plans to take up AW
16:55:26 <davdunc> jberkus: got it.
16:55:38 <dustymabe> puiterwijk: I have been
16:55:42 <jberkus> davdunc: AW is pretty much just for laptops
16:55:59 <dustymabe> jberkus: but we don't discriminate against desktops either
16:56:05 <dustymabe> equal opportunity here
16:56:18 <jberkus> what's a desktop?  one of those 18" laptops?
16:56:32 <dustymabe> jberkus: have you ever seen walters laptop :)
16:56:37 <dustymabe> that's a desktop in a bag
16:56:39 <jberkus> ok, other stuff: are all updates to getfedora.org queued up?  including changes to cloud provider links?
16:57:13 <dustymabe> jberkus: nope.
16:57:21 <maxamillion> was that an action item assigned to anyone?
16:57:29 <dustymabe> i usually work with robyduck on that - need to get to it
16:57:32 <jberkus> not AFAIK
16:57:41 <dustymabe> would love if someone else wanted to pick it up this time
16:58:03 <dustymabe> it's kind of a complicated process, unfortunately
16:58:12 <jberkus> yes.  especially the cloud provider links
16:58:28 <jberkus> but we need davdunc's new images on there, as well as DO
16:58:32 <dustymabe> ehh, the cloud provider links are just s/AMI/AMI
16:58:39 <dustymabe> jberkus: yeah, that's feature work
16:58:54 <dustymabe> i'm not volunteering for that unfortunately
16:59:03 <dustymabe> i keep the lights on
16:59:15 <dustymabe> which basically means the exact page you see today with updated information on it
17:00:01 <jberkus> what needs to change other than the new links & versions?  is it just DO?
17:00:21 <davdunc> jberkus: I have another hoop to jump through before the AWS Marketplace images are out.
17:00:23 <dustymabe> adding a section on aws marketplace would be new
17:00:26 <davdunc> should not be long.
17:00:30 <dustymabe> adding a section on DO would be new
17:00:31 <maxamillion> I've dug into the website code a few times in the past (I have commit rights as a side effect) but starting in a couple weeks I'm going to have very few spare cycles so I'm not sure I can pick it up
17:00:46 <dustymabe> maxamillion: yeah I have commit access as well
17:00:54 <kurushiyama> Good evening
17:00:57 <dustymabe> kurushiyama: hiya
17:00:59 <maxamillion> dustymabe: +1
17:01:46 <ashcrow> Hey there kurushiyama!
17:01:49 <dustymabe> jberkus: if you open a ticket tag it with websites
17:01:51 <dustymabe> https://pagure.io/atomic-wg/issues?status=Open&tags=websites
17:02:10 <jberkus> dustymabe: we already have a ticket open for DO, it's been open for ~~ 4mo
17:02:21 <jberkus> I'll just add to that
17:02:24 <dustymabe> jberkus: sure. ok
17:02:29 <jberkus> but clearly we need someone to take that on as their task
17:02:40 <dustymabe> i'm all tasked out
17:02:40 <jberkus> preferably someone who's not already doing 7 other things ...
17:03:55 * dustymabe notes we have 30 minutes left
17:03:58 <jberkus> ok, anything else for F27 release?
17:04:18 <dustymabe> jberkus: mostly a few chess pieces that need to move behind the scenes
17:04:32 <dustymabe> me puiterwijk and mboddu mostly will be taking care of those
17:04:40 <jberkus> #topic action items from last meeting
17:05:21 <jberkus> ashcrow to schedule container-runtimes meetings for discussions around cri-o/docker
17:05:40 <jberkus> I've had some email with ashcrow about this, but I don't know that we've had a meeting
17:05:44 <jberkus> ashcrow: ?
17:05:54 <jberkus> one of the things on my plate is the survey
17:06:06 <ashcrow> jberkus: we have not had the meeting. I have a card to set it up if needed as well, but want to wait until we have some data first.
17:06:13 <jberkus> ok
17:06:23 <jberkus> I'll get a survey out today so it doesn't get swamped by f27
17:06:48 <jberkus> ashcrow: I have some anecdotal data from LISA ... apparently for downstream it *does* matter that Docker is included
17:07:07 <jberkus> s/anecdotal data/anecdotes
17:07:17 <jberkus> #action jberkus to get out survey about container runtimes
17:07:21 <ashcrow> jberkus: interesting. Good to know.
17:07:36 <dustymabe> jberkus: we also have the feedback from kurushiyama last meeting
17:07:50 <jberkus> ashcrow: regulatory.  if Docker is part of RHAH, then agencies don't need to get separate permission for it
17:07:58 <jberkus> dustymabe: I didn't make last meeting, is that in the logs?
17:08:07 <dustymabe> jberkus: yes, it shoudl be
17:08:31 <ashcrow> hrm, I wonder if that extends to other runtimes too :-/
17:08:38 <jberkus> jbrooks jberkus to test asciibinder rpm
17:08:44 <jberkus> I've done nothing on this, jbrooks ?
17:08:45 <ashcrow> anyway, we'll circle back when we get some of the survey data in as well
17:08:50 <jbrooks> jberkus, me neither
17:09:02 <jberkus> #action jbrooks jberkus to test asciibinder rpm
17:09:20 <jberkus> strigazi jbrooks to document the steps they took for f27 kube upgrade
17:09:25 <jberkus> that's done, no?
17:09:34 <jbrooks> jberkus, Yeah, I'll file a pr today for a blog post
17:09:52 <jberkus> maxamillion to announce to the world that we are planning to remove fedora docker layered images from docker hub and start fresh
17:10:44 <maxamillion> yeah, I need to get the automatic rebuild work done before I can do that or else we won't have anything to sync right away
17:10:52 <maxamillion> I'm hoping to finish that up Friday
17:10:58 <jberkus> maxamillion: re-action?
17:11:01 <maxamillion> jberkus: yes please
17:11:09 <jberkus> #action maxamillion to announce to the world that we are planning to remove fedora docker layered images from docker hub and start fresh
17:11:13 <jberkus> ok
17:11:14 <kurushiyama> ashcrow Basically, it applies to other runtimes. However, since using docker is pretty much the reason why one chooses AH,...
17:11:27 <jberkus> #topic Issues tagged meeting
17:11:44 <jberkus> https://pagure.io/atomic-wg/issues?status=Open&tags=meeting
17:11:59 <jberkus> 372: include firewalld in atomic host
17:12:09 <jberkus> walters?  dustymabe ?
17:12:35 <dustymabe> jberkus: mostly it's a proposal to include firewalld in the host
17:12:46 <jberkus> yah, but someone wanted to discuss it at the meeting
17:12:50 <dustymabe> there has been a decent amount of discussion in the ticket
17:12:59 <ashcrow> For firewalld it looks like we have mainly +1's there
17:13:00 <dustymabe> basically would be nice if we can formally decide
17:13:24 <ashcrow> We did ask for folks to give their $0.02 so we could decide by this meeting.
17:13:28 <dustymabe> and then maybe discuss the implications of enabling it by default or not. my thoughts are that we would initially not enable it by default
17:13:28 <ashcrow> (if I remember correctly)
17:14:04 <maxamillion> yeah +1 here for firewalld
17:14:08 <jbrooks> +1
17:14:11 <jberkus> it looks like there's some technical work to be done first
17:14:12 <walters> yeah, i'd agree; it'd be quite traumatic to enable by default
17:14:21 <jberkus> how much MB does it add to the image?
17:14:34 <dustymabe> #link https://pagure.io/atomic-wg/issue/372
17:14:37 <walters> the other concern I had was that adding it wouldn't actually enable ansible to work as it requires extra modules on the host
17:14:39 <kurushiyama> dustymabe Why not. It does not do any harm, and helps a lot.
17:14:39 <jbrooks> rhelah isn't doing this, are they?
17:14:48 <jbrooks> Seems like they should
17:15:14 <ashcrow> walters: are you saying there are more packages needed for firewalld to be helpful for openshift?
17:15:30 <walters> for openshift-ansible i think probably yes
17:15:40 <jberkus> has anyone done a size/dependency check, in case firewalld drags in a bunch of dependencies?
17:15:50 <dustymabe> kurushiyama: it would be a change. to the defaults and this change won't make f27 release. I'd prefer for us to not change the default until later I think. Unless there is really good reason to
17:15:59 <walters> https://github.com/ansible/ansible-modules-extras/blob/6c7d63b15c77126b4d6a8a7668545555578469c5/system/firewalld.py#L497
17:16:11 <jberkus> so I'm +1 pending issues not yet known per the questions above
17:16:13 <dustymabe> jberkus: I can do that, but I don't think it pulled much in when I installed it last time
17:16:26 <walters> ideally ansible would just talk dbus directly...
17:16:35 <jberkus> ok, folks, let's continue discussion on the ticket, sounds like folks are positive to firewalld, but some technical work needs doing
17:16:44 <ashcrow> jberkus: +1
17:16:57 <jberkus> 363: WG membership and quorum rules
17:17:04 <jberkus> https://pagure.io/atomic-wg/issue/363
17:17:05 <dustymabe> walters: let's make sure we include anything needed for ansible+firewalld to work
17:17:11 <dustymabe> because that is one of the biggest reasons for doing it
17:17:26 <jberkus> for membership, I've gotten pretty much zero feedback on people to add/remove
17:17:28 <walters> ok, then we're pulling in python3-firewall too
17:17:51 <walters> but at this point we already have so much python2 and 3 stuff it'll just be lost in the noise
17:18:31 <jberkus> however, I would like to update the quorum rules
17:18:35 <maxamillion> walters: I don't know how to do that but would be willing to migrate to talking to dbus directly, I just use the python bindings because that's what I knew how to implement
17:18:46 <jberkus> currently valid quorum is 51% of all listed members
17:18:52 <dustymabe> jberkus: so you want to talk about who to remove/add?
17:19:05 <jberkus> which was fine when there were 6 listed members, but not when there's 20
17:19:14 <walters> maxamillion, oh it's understandable, the module long predates Atomic Host and containers etc
17:19:29 <jberkus> I'd like to reduce the quorum to "5 members, or 51%, whichever is less"
17:19:31 <walters> at some point though we'll ideally make progress on having ansible use a "tools" container
17:19:42 <maxamillion> walters: but yeah, I'm totally open to change
17:19:45 <jberkus> because the alternative is restricting membership
17:19:53 <dustymabe> jberkus: yeah it's hard to comment on who to add/remove unless we know if we're restricting membership or not
17:19:55 <maxamillion> walters: ah +1
17:20:04 <maxamillion> walters: that'd be interesting
17:20:15 <kurushiyama> jberkus I see a bit of a problem in general. Just from my noob view in this group, I guess there are two types of decision
17:20:40 <kurushiyama> Day to day decisions, for which a majority of _present_ members should be enough.
17:21:14 <kurushiyama> And decisions with a long term or high impact, for which the group as a whole should agree upon.
17:21:55 <jberkus> kurushiyama: well, realistically we don't actually tally votes for the first
17:22:00 <jberkus> we just do lazy consensus
17:22:28 <jberkus> kurushiyama: but even for the 2nd, for the currently listed WG members, we have folks who attend maybe 1 out of 5 meetings
17:22:31 <ashcrow> kurushiyama: and it can be very hard to get 100% agreement on things ... there is usually at least one person who may have a different point of view.
17:22:56 <ashcrow> as long as it isn't a hard -1 I feel like it shouldn't block
17:23:01 <kurushiyama> So, as far as I understand, we rather need a clear definition on what kind of decisions a majority of the group as a whole should be involved in, no?
17:23:25 <jberkus> currently for a 51% quorum, we would need to have 11 WG members attending the meeting for it to be able to make decisions.  we seldom have that many
17:23:27 <kurushiyama> ashcrow "As a whole" was meant as "quorum of the whole group"
17:23:35 <ashcrow> kurushiyama: ah ok
17:23:50 <dustymabe> jberkus: i think it would be reasonable to have 10 official members
17:23:53 <jbrooks> We have 20 members?
17:24:05 <jberkus> jbrooks: look at the wiki page
17:24:17 <kurushiyama> jberkus Hm. If important decisions come up, make an announcement. Then, reduce it to the quorum of the present members.
17:24:34 <jberkus> so we *either* restrict the the quorum, or the membership
17:25:15 <jberkus> kurushiyama: so a split between pre-announced votes and ones which are not scheduled?
17:25:39 <kurushiyama> Wikipedia (in Germany) does it this way: open a voting page. Everybody can vote during a certain period. Period over and no vote= bad luck.
17:25:47 <kurushiyama> jberkus Basically.
17:26:19 <dustymabe> yeah that's my thoughts for discuss ticket at meeting on week 1, leave ticket open for week for votes, close out on week 2
17:26:30 <jberkus> ok
17:26:33 <ashcrow> dustymabe: yeah, that makes sense
17:26:44 <jberkus> #action jberkus to propose new quorum rules based on scheduled online votes
17:27:00 <jberkus> kurushiyama: I'll ping you about those when I have a draft
17:27:07 <dustymabe> jberkus: do we want to talk about people to remove from that page?
17:27:20 <jberkus> dustymabe: not in the meeting
17:27:27 <dustymabe> ok
17:28:10 <jberkus> 354: the help.md files in FLIBS images
17:28:23 <jberkus> this is a bit of a mess due to us changing the rules several times
17:28:50 <jberkus> which phracek points out
17:29:56 <jberkus> does that ticket have a new set of rules we think are clearer and we can stop changing?
17:30:06 * ashcrow looks
17:30:07 <dustymabe> i'm willing to yield decision on this to whomever thinks they have a good grasp of the problem
17:31:19 <jberkus> hmmm
17:31:20 <ashcrow> Since I updated it to clarify the proposal I'll abstain from +1/-1ing it ... however, the basic idea behind the original request is that ther help behind these images are free form and sometimes missing
17:31:31 <ashcrow> And they may be in one of two files
17:31:39 <jberkus> I don't think its settled, I just had a comment about a different type of contianer
17:31:40 <ashcrow> this proposal makes more firm requirements
17:32:20 * dustymabe notes we are over time
17:32:46 <jberkus> yes
17:32:52 <jberkus> ok, thanks all, sorry for no open floor
17:32:56 <jberkus> lots to do today
17:33:00 <jberkus> #endmeeting