17:11:51 #startmeeting releng-demo 17:11:51 Meeting started Tue Dec 15 17:11:51 2015 UTC. The chair is dgilmore. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:11:51 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:11:51 The meeting name has been set to 'releng-demo' 17:11:59 http://taiga.fedorainfracloud.org/project/acarter-fedora-docker-atomic-tooling/wiki/demo-videos 17:12:14 17:02 * rdieter lurking watching @ https://www.youtube.com/watch?v=UO-_WEVYXvo 17:13:44 rtnpro: for later, I think a better name than "family" would be "variant"? 17:14:33 what is the strategy for installing openshift origin in fedora? ansible? 17:14:47 rtnpro: looks slick :) 17:15:01 threebean, agreed! 17:15:22 what is the role of autocloud in the larger plan being discussed here? just curious, and I'm new at this ☺ 17:15:23 threebean, creating an issue for it 17:15:36 dustymabe: the official OpenShift Ansible installer 17:16:24 maxamillion: is it going to run openshift in containers? or natively on the machine? 17:16:54 dustymabe: natively on the machine, I plan to also enable openshift in containers but that's still unstable upstream to the best of my knowledge 17:16:57 dustymabe: https://github.com/openshift/openshift-ansible/blob/master/README_origin.md 17:17:14 bowlofeggs: in this diagram -> https://fedoraproject.org/wiki/Changes/Two_Week_Atomic#Schematic 17:17:20 bowlofeggs: it is the orange box. 17:17:41 rtnpro: any chance we could have a field that links to the image that was tested (the link from koji) 17:17:44 oh cool, thanks! 17:17:55 maxamillion: ok 17:18:16 will fedora also be signing docker manifests with F24? 17:18:17 once we get it running in containers I think it would be interesting to use atomicapp to run openshift 17:18:20 dustymabe: the links shoudl be to the compose not koji 17:18:45 dustymabe: we need to think about thinsg not as koji tasks but as composes 17:18:49 bowlofeggs, autocloud is an application to test Fedora cloud images, using tunir, and a dashboard to visualize the test results 17:19:09 rtnpro: yeah it looks cool - thanks for sharing! 17:19:24 dgilmore: but there could be multiple images output of a compose right? 17:19:33 bowlofeggs: afaik docker signing is all done by docker 17:19:41 dustymabe: yes 17:19:57 but the link should be the compose location not teh koji location 17:20:01 dustymabe, isn't link to the koji build enough (the 1st column)? 17:20:09 dustymabe: that's on the roadmap but I'm unsure if it will land before Fedora 24 so I don't want to include it in the scope of the current changes 17:20:17 rtnpro: we should not provide koji links 17:20:33 dgilmore: then how do we say exactly what image we tested 17:20:47 dgilmore: do you mean docker the company, dockerhub, or the docker registry instance that fedora will be using with pulp? 17:21:06 dgilmore: that's the base images on docker hub 17:21:12 bowlofeggs: I am unaware of any signing that we can do 17:21:15 bowlofeggs: does pulp support it? 17:21:17 :) 17:21:19 dgilmore, ok, could you file an issue for it at https://github.com/kushaldas/autocloud? 17:21:28 bowlofeggs: afaik all teh signing happens by docker in the dockerhub 17:21:29 bowlofeggs: because if pulp supports it, I'd love to enable it 17:21:39 jgreguske: where can i get a copy of this "kojay" that you are using? is it superior to koji? ☺ 17:21:51 :D 17:21:57 bowlofeggs: its koji 17:22:23 dgilmore: the Docker manifest has a "signatures" section - i believe fedora *could* insert their own signatures there 17:22:37 maxamillion: yes, pulp will "support" it, but only in that it won't interfere 17:22:38 bowlofeggs: if we can it is news to me and not scoped 17:23:12 maxamillion: so if you give pulp a manifest that is signed, it will be validly signed when docker clients pull it from pulp/crane 17:23:55 dgilmore, maxamillion: perhaps we could schedule a time to talk about this in more detail 17:23:58 bowlofeggs: :) 17:24:00 bowlofeggs: oh ok, so it's not something pulp provides? 17:24:03 bowlofeggs: we should file an issue in taiga to have it be visable for us to look at 17:24:27 maxamillion: no pulp doesn't provide it, but it certainly will work with manifests that are already signed 17:24:30 bowlofeggs: yeah, I think we should ... it's definitely something I would like to have but I thought it was something pulp needed to provide and that it just didn't yet 17:25:14 hm, the youtube is showing jay as opposed to ralph's screen right now 17:25:27 another thing worth noting is that docker is changing the manifest format (again!) and this may impact how the signatures will work. they have not decided on the format yet to my knowledge 17:25:42 acarter1: can we control who's in focus on the youtube stream? it's showing Jay fullscreen 17:25:47 stupid youtube 17:25:51 bowlofeggs: of course they are 17:25:55 hahaha 17:26:00 it's my world of pain 17:26:08 bowlofeggs: https://github.com/Zyzyx/koji/tree/signed 17:26:15 bowlofeggs: the work so far 17:26:18 please fix 17:26:23 thanks 17:26:41 he's sharing it, but jay is zoomed in 17:26:54 much zoom 17:26:58 wow 17:27:10 cant see it 17:27:17 acarter1: maybe try clicking on threebean's thumbnail? 17:27:18 https://fedoraproject.org/wiki/Changes/ProductDefinitionCenter 17:27:25 https://pdc.fedorainfracloud.org/ 17:27:43 acarter1: there you go 17:27:46 jay, youtube has chosen YOU 17:27:51 threebean: yep 17:27:52 ys good to go 17:28:06 yeah i can see 17:30:52 we've got the time 17:31:02 audit tools! \o/ 17:33:37 THIS IS AWESOME 17:33:58 https://github.com/fedora-infra/pdc-updater/ 17:33:58 I *might* be the only one *this* excited about PDC and the audit/retry tooling 17:34:04 :D 17:34:17 ralph: that beard is legit 17:34:21 threebean: you're basically my new hero 17:34:48 oh man.. I forgot to show off the test suite. (there is a test suite...) 17:35:07 let me rejoin 17:35:12 threebean: your beard has a test suite‽ 17:35:34 dgilmore, adam is going to do his last demo while you rejoin 17:35:36 acarter1: you should arbitrarily share peoples' screen during the meeting and creep on them 17:35:40 bowlofeggs: o-o ... not yet. 17:35:45 hehe 17:37:44 jgreguske, I'm sure everyone is just working on their Taiga cards 17:38:18 FYI, as a probable new contributor, the guide looks great 17:38:20 did maxamillion lock up? 17:38:24 or me? 17:38:26 dgilmore: nope, you 17:38:57 dgilmore, :< 17:39:06 * lmacken hopes for chipmunk mode 17:39:18 lmacken: YESSSSS 17:39:23 dgilmore, you joining back? 17:39:32 lmacken: live, on youtube, recorded for the world to preserve 17:39:37 Google Hangouts had better have a chipmunk voice plugin. 17:39:40 LET'S DO THIS 17:39:41 maxamillion: priceless 17:39:59 awwww 17:40:02 womp womp 17:40:11 err... I mean, yey! :D 17:40:32 lol 17:40:49 I got hidden :( 17:40:58 dgilmore: think we lost you 17:41:27 maxamillion: I think so 17:41:42 firefox is not responsive 17:42:37 dgilmore, has that happened to you before? 17:42:41 Good stuff 17:42:51 thanks mclong! 17:43:11 dgilmore, maybe it's because you disabled the chipmunk plugin? 17:43:22 dgilmore, your public is demanding it 17:43:34 acarter1: no I think it is because my desktop is dual screen 17:44:09 I tried rejoining from my laptop but it says noone is in the call 17:44:14 dgilmore, I might grab some time with you tomorrow or Thursday to try to record a quick run through? 17:44:16 dgilmore, we gave up 17:44:21 :( 17:44:26 3 strikes and you're out 17:44:49 acarter1: that was two strikes 17:45:05 dgilmore, 3 -- adam stepped in after the first one 17:45:22 acarter1: that was the first, then I had one more after that 17:45:26 then you gave up 17:47:09 #endmeeting