17:02:45 <dustymabe> #startmeeting
17:02:45 <zodbot> Meeting started Wed Oct 21 17:02:45 2015 UTC.  The chair is dustymabe. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:02:45 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:02:54 <dustymabe> #topic roll call
17:03:01 <scollier> .hello scollier
17:03:02 <zodbot> scollier: scollier 'Scott Collier' <emailscottcollier@gmail.com>
17:03:09 <roshi> sure thing dustymabe
17:03:09 <mattdm> .hello mattdm
17:03:10 <zodbot> mattdm: mattdm 'Matthew Miller' <mattdm@mattdm.org>
17:03:13 <dustymabe> .hellomynameis dustymabe
17:03:16 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dustymabe@redhat.com>
17:03:22 <dustymabe> roshi: there is always one thing i forget to do at this point
17:03:28 <dustymabe> like meeting title or something
17:03:28 <praveenkumar> .hellomynameis kumarpraveen
17:03:29 <zodbot> praveenkumar: kumarpraveen 'None' <kumarpraveen.nitdgp@gmail.com>
17:03:45 <roshi> #meetingname Cloud WG
17:03:45 <zodbot> The meeting name has been set to 'cloud_wg'
17:04:08 <roshi> and typically you do #startmeeting <meeting name>
17:04:14 <dustymabe> ahh ok
17:04:43 <mattdm> note that this seems silly but is important for meetbot
17:04:54 <mattdm> I often read logs at https://meetbot.fedoraproject.org/
17:05:02 <dustymabe> trying to find notes from old meeting now
17:05:54 <mattdm> feel free to copy this wiki page and make it a template
17:05:57 <mattdm> https://fedoraproject.org/wiki/Council_meeting_process#Day_of_meeting
17:06:05 <mattdm> i use that and just copy and paste every time :)
17:06:16 <dustymabe> the search on mote doesn't seem to be working
17:06:22 <mattdm> and it has Silly Wiki Magic to get the date right
17:06:42 <mattdm> wfm
17:06:43 * roshi just goes to meetbot-raw and looks in the room for the date of the last meeting
17:06:54 <dustymabe> ok here we are
17:06:55 <mattdm> everyone is inconsistant with the meeting name
17:06:56 <dustymabe> https://meetbot.fedoraproject.org/teams/cloud_wg/cloud_wg.2015-10-07-17.02.html
17:07:05 <dustymabe> I think
17:07:08 <dustymabe> that is from two weeks ago
17:07:11 <dustymabe> was there a meeting last week?
17:07:16 <mattdm> cloud_wg cloud-wg cloud-workgroup fedora-cloud :)
17:07:40 * roshi for the most part does the same thing every time, he thinks
17:07:50 <maxamillion> .hello maxamillion
17:07:51 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
17:08:01 <roshi> same with blocker review - since I have idle thoughts of parsing more from our meeting minutes - so I want things the same
17:08:18 <dustymabe> can someone please tell me how to search this thing by freenode channel?
17:08:24 <dustymabe> i used to be able to
17:08:27 <mattdm> yeah it mostly all seems to be under cloud_wg
17:08:28 <mattdm> https://meetbot.fedoraproject.org/sresults/?group_id=cloud_wg&type=team
17:08:59 <mattdm> dustymabe: dunno. rfe → https://github.com/fedora-infra/mote/issues
17:09:00 <roshi> https://meetbot-raw.fedoraproject.org
17:09:04 <dustymabe> roshi: did we not have a meeting last week?
17:09:39 <dustymabe> ok here we go
17:09:41 <dustymabe> https://meetbot-raw.fedoraproject.org/fedora-meeting-1/2015-10-14/fedora-meeting-1.2015-10-14-17.02.html
17:09:43 <roshi> honestly, I can't remember
17:09:56 <dustymabe> sigh.. and that is why I send the minutes to the list when I host a meeting
17:10:02 <roshi> see how the meeting name doesn't show up there?
17:10:08 <dustymabe> #topic items from last meeting
17:10:20 <dustymabe> praveenkumar and coolsvap will work on https://fedorahosted.org/cloud/ticket/121
17:10:22 <dustymabe> coolsvap will put in his ideas to https://fedorahosted.org/cloud/ticket/124
17:10:24 <dustymabe> praveenkumar will write the examples for https://fedorahosted.org/cloud/ticket/125
17:10:36 <dustymabe> coolsvap around today?
17:11:18 <dustymabe> praveenkumar: ^^
17:11:29 <praveenkumar> looks like he is not here :(
17:11:44 <dustymabe> praveenkumar: do you have any updates from your end?
17:12:03 <praveenkumar> Anyway I was working on #121 and put comments about progress.
17:12:26 <scollier> this ticket can be close: https://fedorahosted.org/cloud/ticket/124  the CI one.
17:12:26 <dustymabe> #info praveenkumar was working on #121 and put comments about progress.
17:12:38 <dustymabe> scollier: ok. do you have the rights to do that?
17:12:52 <scollier> dustymabe, i'll try
17:12:53 <dustymabe> if you don't just comment in the ticket and I will close
17:13:16 <dustymabe> # scollier to close https://fedorahosted.org/cloud/ticket/124
17:13:21 <dustymabe> #action scollier to close https://fedorahosted.org/cloud/ticket/124
17:13:28 <dustymabe> ok moving on to tickets
17:13:29 <praveenkumar> scollier: is it resolved now?
17:13:41 <scollier> dustymabe, i can't close it.
17:13:46 <dustymabe> I am going to prioritize some issues
17:13:53 <dustymabe> scollier: ok just comment and I will close
17:14:00 <scollier> praveenkumar, well, there will be no CI on these if we are moving to dist-git-dockerfiles
17:14:15 <praveenkumar> ah right.
17:14:22 <scollier> praveenkumar, let me rephrase, I think the CI should be done after the move.
17:14:23 <dustymabe> #topic python3 only means ansible won't work
17:14:30 <mattdm> presumably that solution will have its own ci, right?
17:14:34 <dustymabe> https://fedorahosted.org/cloud/ticket/126
17:14:39 <mattdm> +1 to scollier's rephrase :)
17:14:47 <dustymabe> mattdm: I am interesting in your thoughts on $topic
17:14:50 <praveenkumar> scollier: Agreed.
17:15:01 <mattdm> dustymabe: so, there's this:
17:15:03 <mattdm> http://blog.oddbit.com/2015/10/15/bootstrapping-ansible-on-fedora-23/
17:15:15 <dustymabe> mattdm: yeah I saw that
17:15:18 <mattdm> which is a somewhat complicated but functional bootstrapping approach
17:15:34 <mattdm> maybe we could package that up in in some way?
17:16:05 <mattdm> The other idea would be to somehow use a super-privleged container, but I'm not sure how that'd work exactly in order to given ansible access to the host in a way that wouldn't need to be weirdly customized
17:16:35 <dustymabe> mattdm: SPC would only be realistic for atomic host since that is the only place docker is already installed
17:16:44 <walters> i vote adding the python 2 version of python-selinux
17:17:01 <walters> and maybe python-docker-py though honestly the way we expect docker containers to get deployed generally is via kube
17:17:02 <mattdm> dustymabe: yes, true
17:17:04 <dustymabe> walters: thanks fro joining
17:17:15 <dustymabe> walters: mattdm: ok let's separate our discussion
17:17:23 <dustymabe> first let's talk about cloud base image
17:17:49 <dustymabe> so for cloud base we don't want to add python2 ?
17:17:52 <dustymabe> is that accurate?
17:18:46 <dustymabe> did we lose mattdm?
17:18:47 <mattdm> I'd say that's accurate, yeah.
17:18:51 <mattdm> No I was thinking
17:18:52 <dustymabe> ok
17:18:58 <mattdm> I'm depressed about the image size overall :)
17:19:00 <maxamillion> +1 to python-docker-py
17:19:33 <roshi> speaking of the base image, is there thoughts/talks of making our cloud base image more like server?
17:19:40 <dustymabe> roshi: bad
17:19:46 <dustymabe> let's focus for now
17:19:51 <mattdm> bigger different discussion, yeah :)
17:19:51 <roshi> it was brought up earlier in the week
17:20:01 <dustymabe> ok so we will do with workarounds for cloud base for now
17:20:08 <roshi> well, i thought it fit right in there with "should we bring in python2"
17:20:09 <dustymabe> and try to publicize alternatives
17:20:25 <mattdm> +1 to Solve it with Docs!
17:20:26 <dustymabe> walters: for atomic we still have python2
17:20:31 <dustymabe> but we need supporting libraries
17:20:42 <mattdm> for atomic for now, the approach is to bring all the stuff in for now
17:20:53 <mattdm> And if walters and atomic team are for that, I'm supportive
17:21:26 <dustymabe> mattdm: ok we will add back some things that get us most of the functionality
17:21:38 <dustymabe> it might not make release though
17:21:43 <dustymabe> depending on when exactly that happens
17:21:54 <dustymabe> but with 2-week-atomic that shouldn't matter much
17:22:00 <mattdm> s'okay. Gives us something to talk about at the next 2 week point :)
17:22:03 <walters> i'm testing http://fpaste.org/282019/54481151/ locally
17:22:41 <dustymabe> walters: I think we need a few others
17:23:03 <dustymabe> we can discuss that in a ticket - I will open a new ticket specifically for fedora cloud base
17:23:19 <mattdm> +1 ticket
17:23:51 <dustymabe> ok another important one:
17:23:59 <dustymabe> #action dusty to open ansible on atomic 23 ticket
17:24:06 <dustymabe> #topic make docker archived image get imported with lowercase tag
17:24:15 <dustymabe> #link https://fedorahosted.org/cloud/ticket/131
17:25:12 <dustymabe> has anyone looked at this before?
17:25:17 <dustymabe> mattdm ^^
17:25:25 <dustymabe> basically it's just a bad user experience for the user
17:25:53 <mattdm> new to me (sorry!)
17:26:02 <dustymabe> ahh sorry
17:26:12 * dustymabe gives you time to read
17:26:20 <dustymabe> feel free to ask me questions
17:26:22 <mattdm> Is there a corresponding releng ticket?
17:26:35 <dustymabe> mattdm: not currently
17:26:54 <dustymabe> so I discussed this with Dennis and I don't think anything can be done for F23
17:27:03 <dustymabe> so I haven't chased it down further
17:27:13 <mattdm> because, too late?
17:27:20 <dustymabe> yeah.. too late
17:27:45 <mattdm> :-/ we have got to get this process changed so little fixes are possible somehow
17:28:14 <mattdm> with us spinning a hopeful rc _right now_, I see the point of 'too late', but 5 days ago....
17:28:15 <dustymabe> so I think there will be some doc work to account for this?
17:28:19 <mattdm> but annnyway.
17:28:37 <mattdm> possibly Common Bugs?
17:28:47 <mattdm> does this affect people pulling from the docker registry?
17:29:03 <dustymabe> mattdm: it shouldn't affect anyone pulling from docker registry
17:29:17 <dustymabe> only if you pull our image down locally and try to use it
17:29:33 <dustymabe> the docker registry won't allow you to upload anything with an invalid name
17:29:43 <dustymabe> so it would never make it into the registry if it did
17:29:59 <mattdm> *nod*
17:30:12 <mattdm> What's the status on putting out updated docker images?
17:30:26 <dustymabe> good question? I really don't know who runs that?
17:30:34 <maxamillion> mattdm: base images or layered images?
17:30:40 <mattdm> maxamillion: base images
17:30:47 <mattdm> with security fixes
17:31:36 <maxamillion> mattdm: that's me, I can update anytime we like, but it's a manual process where I have to file a github pull request with an image that's been converted from the koji build ... the kicker there is that none of that stuff is tested
17:32:04 <dustymabe> maxamillion: have we been testing the f23 docker images at all?
17:32:10 <dustymabe> I've downloaded it a couple of times
17:32:12 <dustymabe> but not much
17:32:36 <mattdm> so, an updated image could theoretically be much better tested than the one we are going to GA with
17:32:47 * mattdm is not sure what emoticon should go on that last line
17:32:55 <mattdm> probably :-/
17:32:57 <dustymabe> mattdm: define "updated image"
17:32:58 <mattdm> that's my go-to
17:32:58 <maxamillion> I don't know, I've just been given perms by the docker library upstream folks to push images ... I'm "officially" the fedora docker base image maintainer for the hub as of a couple months ago
17:33:15 <dustymabe> maxamillion: \o?
17:33:19 <maxamillion> ?
17:33:19 <dustymabe> \o/
17:33:23 <dustymabe> type
17:33:26 <dustymabe> typo
17:33:28 <maxamillion> alright
17:33:31 <mattdm> dustymabe: we ask releng to do a respun f23 docker image post-release, with security updates rolled in
17:33:47 <mattdm> do a test day, and then have maxamillion push it to the hub and rel-eng to the mirrors
17:34:02 <roshi> maxamillion: I can help you with the testday
17:34:03 <dustymabe> maxamillion: what does our tagging look like now?
17:34:05 <maxamillion> basically, I don't know if anyone is actually testing any of it and I don't want to make a terrible assumption that Fedora QA is doing it because they already test the whole damn world
17:34:20 <mattdm> and put "automate this whole process a la two week atomic" on the Big Backlog List
17:34:22 <maxamillion> dustymabe: "our tagging" ... ?
17:34:28 <maxamillion> mattdm: can't be done
17:34:28 <roshi> if it's not blocking release, don't count on us testing it
17:34:32 <dustymabe> tagging for the docker images
17:34:33 <roshi> we're spread pretty thin
17:34:40 <maxamillion> mattdm: the update to docker requires a manual pull request
17:34:45 <dustymabe> I feel like all of Fedora is spread really thin
17:34:52 <maxamillion> mattdm: we could automate everything leading up to that though
17:34:57 <roshi> we all do work though :p
17:35:06 <mattdm> maxamillion yeah. we could at least produce the images and push 'em to our own Future THeoretical Registry
17:35:14 <roshi> fedora peeps++
17:35:16 <maxamillion> roshi: oh yeah, y'all are spread super thin
17:35:44 <dustymabe> give me job descriptions!!
17:35:51 <dustymabe> if we have any
17:35:51 <maxamillion> mattdm: right, that we could do ... I actually have a card in Taiga to spec that out, scope it, and draw up a design doc
17:35:53 <dustymabe> :)
17:35:54 <roshi> if someone has a spare req, we'd take it
17:35:59 <mattdm> maxamillion++
17:36:00 <zodbot> mattdm: Karma for maxamillion changed to 1 (for the f23 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:36:01 * roshi has job descriptions :p
17:36:09 <dustymabe> roshi: PM me
17:36:09 <maxamillion> \o/
17:36:10 * roshi will sync with you later dustymabe
17:36:21 <maxamillion> mattdm: http://taiga.cloud.fedoraproject.org/project/acarter-fedora-docker-atomic-tooling/us/282
17:36:24 <dustymabe> ok back to meeting
17:36:47 <dustymabe> #topic Fedora Vagrant Boxes in Atlas
17:36:54 <dustymabe> #link https://fedorahosted.org/cloud/ticket/116
17:37:05 <mattdm> wait, so, the plan for previous: ship as is, document in common bugs, and do a respin at some future point?
17:37:11 <dustymabe> so me and lalatend1M are going to be releasing f23 vagrant boxes in Atlas
17:37:14 <roshi> that's what I read
17:37:21 <dustymabe> we have access and permission from legal
17:37:29 <mattdm> dustymabe++
17:37:30 <zodbot> mattdm: Karma for dustymabe changed to 1 (for the f23 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:37:33 <mattdm> lalatend1M++
17:37:43 <dustymabe> we will be publishing a few boxes soon and send to list to have people test
17:38:07 <dustymabe> #topic Producing Updated Cloud/Atomic Images
17:38:14 <dustymabe> #link https://fedorahosted.org/cloud/ticket/94
17:38:25 <mattdm> checkit out, everyone:
17:38:27 <mattdm> https://www.youtube.com/watch?v=r1jrqClho3U
17:38:42 <mattdm> #info two week atomic release tooling demo
17:38:42 <dustymabe> ok now that we are getting close to f23 release that means 2 week atomic is around the corner
17:39:13 <dustymabe> mattdm: nice.. were we invited to that?
17:39:22 <mattdm> dustymabe: uh, maybe? :)
17:39:36 <dustymabe> ok. so my question is how close are we to something that we can use?
17:39:38 <mattdm> shoulda been if not.... sorry
17:39:46 <dustymabe> how much work needs to be done to pull everything off?
17:40:01 <mattdm> very. team is confident that it'll be ready with what was previously "plan b"
17:40:19 <mattdm> that is, ship ga image at ga, but then switch to two week process for release two weeks after
17:40:26 <mattdm> currently targetted at Nov 10
17:40:32 <dustymabe> mattdm: ok
17:40:39 <mattdm> and I'm not going to jinx anything by guessing
17:40:52 <dustymabe> so "we" (Fedora cloud WG) probably need to start being much more involved in that process
17:41:28 <dustymabe> I know maxamillion and walters are here
17:41:29 <mattdm> dustymabe++
17:41:29 <dustymabe> but..
17:41:37 <dustymabe> we need more people who know more
17:41:39 <dustymabe> I think
17:41:48 <mattdm> yeah. I just pinged acarter, who is doing project management, about that
17:42:36 <mattdm> we *definitely* need more if jzb and I can convince y'all to remain convinced about atomic-as-primary
17:42:37 <dustymabe> #action dmabe,mattdm we need to integrate the Cloud WG and the 2 week atomic efforts
17:42:41 <maxamillion> yeah, the "pull that off" bit is basically done ... at least for the first run at the deliverable, folks have been working hard on that for a while now and a lot of it recently came to fruition
17:43:10 <dustymabe> ok for the rest of the tickets I am going to open it up and let people comment if they have status
17:43:15 <dustymabe> #topic open floor
17:43:17 <roshi> me could use more knowledge on the subject - that's for sure
17:43:34 <mattdm> ftr amanda says she sent an invite to the list
17:43:41 <mattdm> i'm checking for whether it got caught in moderation
17:43:50 <mattdm> or, I will, after this meeting :)
17:44:27 <dustymabe> mattdm: I don't recall seeing one, but that doesn't mean it didn't come through
17:44:40 <dustymabe> anyone have anything for open floor?
17:44:50 <mattdm> release annoucement
17:44:55 <mattdm> see mailing list thread I started
17:45:08 <dustymabe> mattdm: should we talk about ansible in there?
17:45:47 <mattdm> only if it is "ansible is awesome and we have it now"
17:45:59 <mattdm> we should talk about it in the release notes or common bugs, though
17:46:07 <dustymabe> haha. yeah definitely
17:46:22 <mattdm> for the annoucement, set the level at "could run in Infoworld"
17:46:31 <maxamillion> mattdm: we've always had it, it's open source :)
17:46:33 <mattdm> if infoworld is even a thing anymore.
17:46:38 <mattdm> pc magazine
17:46:51 <dustymabe> ok i'll leave meeting open for another meeting to see if anything else comes up
17:47:56 <dustymabe> another minute
17:48:03 * dustymabe sets fuse
17:48:09 <dustymabe> 3..
17:48:18 <dustymabe> 2...
17:48:23 <dustymabe> 1....
17:48:27 <dustymabe> #endmeeting