14:00:59 #startmeeting 14:00:59 Meeting started Wed Oct 10 14:00:59 2012 UTC. The chair is mattdm. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:59 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:01:14 #meetingname cloud-sig 14:01:14 The meeting name has been set to 'cloud-sig' 14:01:22 #meetingtopic Fedora Cloud SIG 14:01:31 anyone around? 14:01:39 * mmorsi says hi 14:02:52 hmmm we'll wait a little bit for other people to trickle in.. 14:03:02 sure 14:03:14 you work on aeolus? 14:03:20 didn't know what chatroom it was gonna be in till you said it on fedora-cloud 14:03:37 yessir 14:03:40 yeah it was on the mailing list but i figured it wouldn't hurt to reinforce. 14:03:59 ah yeh thats what i figured, prolly missed it in the flood 0_o 14:04:09 ya have an update for the cloud-sig list 14:04:42 who, me? 14:04:49 on the meeting or in general? :) 14:04:54 no i do 14:04:58 had hoped to have it out b4 the meeting, unfortunately was blocked on a couple things, but those should be taken care by end of day 14:05:04 ohhhh. well coool. 14:05:20 ya can briefly talk about it when we get things underway 14:05:29 oh and this is just my 2cents 14:05:37 i will take 2 cents. 14:05:43 but am not a big fan of having the unofficial meeting 14:05:49 feel like it causes confusion 14:06:08 i know, but the morning didn't work for some people we need to hear from. 14:06:14 but doesn't really matter to me either way 14:06:29 but I see that no one is here at all so, hmmm. 14:06:36 ya perhaps encouraging people to be more vocal on the ML 14:06:54 will need to try and post advance agendas and make it more useful... but more vocal in the mailing list is even better really. 14:07:03 ya thats tru 14:07:04 let's go ahead with yuor thing 14:07:14 topic is aeolus? 14:07:22 #topic aeolus update 14:07:25 sure 14:07:32 go! 14:07:35 * tdawson notes that he is here. 14:07:43 well as i mentioned b4 i was working on a few templates for fedora infra 14:07:53 eg so you can deploy fedora infra to multiple cloud providers 14:08:06 the barebone / preliminary versions of those are about done 14:08:14 should be added to our templates repo soon 14:08:21 https://github.com/aeolus-incubator/templates 14:08:22 in this context 'fedora infra' means? 14:08:30 mock, koji, bodhi 14:08:45 starting to do things like orchestrating configs 14:08:57 so a component on one cloud provider can talk to a component on another 14:09:01 that's cool. 14:09:01 and so on and so forth 14:09:22 eventually doing things like dynamic pricing adjustments (we're far from that point tho) 14:09:33 and of course a cloud provider could be a local openstack instance 14:09:54 so ya the basic templates to just deploy the services to any provider are just about done 14:10:09 i'm gonna blog (and hopefully screencast it) in the in near future 14:10:13 * mattdm nods 14:10:14 so it should show up on fedora planet 14:11:04 hoping to drive some interest on the fedora side in aeolus and externally in the fedora / cloud efforts and technologies 14:11:21 of course, obligatory any help w/ this effort would be appreciated 14:11:27 cool. have you been working with seth on this? 14:11:42 should all be very easily to follow along one the blog post is out 14:11:46 not atm 14:11:58 want to show it working before i formally approach fedora infra 14:12:15 kinda giving ya'll the preview / plan here, but still in works 14:12:19 probably good to approach them informally too :) 14:12:32 yeh, i mentioned it at hte last fudcon 14:12:36 well the last one i went to 14:12:40 blacksburg 14:13:11 * mattdm nods 14:13:16 ah hey 14:13:29 anything in particular you need? 14:13:53 not atm, again just keep a lookout for specific instructions / steps via the planet blogroll 14:14:04 ok i will. 14:14:07 #topic openshift 14:14:19 tdawson openshift? 14:14:42 We're still doing pretty good. 14:14:58 i looked at the feature page -- looks like there's basically one outstanding review request? 14:15:01 WE've progressed on testing, but nobody has finished the full test. 14:15:24 mattdm: Yes, and that review isn't totally necessary unless we need to update to the newer code. 14:15:45 The extra week for Beta has helped us. 14:15:58 * mattdm nods 14:16:10 The problem is that our setup page is so long that nobody has made it all the way through it ... at least nobody that's reported back. 14:16:45 https://openshift.redhat.com/community/wiki/build-your-own 14:17:04 I admit that I saw that and then thought "maybe later". 14:17:29 We are adjusting it to work with Fedora 18 stuff ... but as I said, nobody has finished, so we don't have that up. 14:17:41 The good news is that upstream has automated all of that 14:17:53 what's the automation tool of choice? 14:18:14 The bad news is they don't consider it stable yet, so we aren't updating to that code unless we're broken. 14:18:24 mattdm: majority of it is just scripts. 14:18:54 the openstack people are packaging up a puppet module that does a simple local deploy 14:19:05 If you go through the "build-your-own" you will see that majority of the stuff can be just cut and pasted ... which makes for good scripts. 14:19:16 something like that would be awesome for openshift too 14:19:24 not necessarily for F18 release :) 14:20:16 mattdm: Yep, I know that somebody has mentioned doing that, and it's on a roadmap ... but for the initial F18 stuff, it's a bit of overkill. 14:20:40 In the meantime, "find some volunteers for testing" is the main thing, then? 14:20:45 tdawson: how many instances do you need? 14:20:58 mattdm: Well, we have 3 testers volunteered ... none have finished. 14:21:01 tdawson: and how tightly coupled are they for bringing up 'openshift' in a private cloud? 14:21:26 skvidal: For our initial "it works" test, we just need one. 14:21:28 tdawson: one might consider that "test failed".... :) 14:21:52 tdawson: b/c my first thought is to take the set of instructions you have and make it into an ansible playbook 14:22:03 since that requires no infrastructure to setup and use, unlike puppet 14:22:22 (puppet doesn't either; masterless puppet ftw.) 14:22:25 skvidal: Although some people have noted that having just one node do everything doesn't show up several bugs ... but ... we're starting with just one. 14:22:39 mattdm: *cough* ruby crap *cough* 14:22:51 tdawson: okay 14:22:58 * mattdm stands back, looks sideways shiftily 14:23:55 so, do we need someone to make the docs less scary? 14:24:11 mattdm: If there is someone who can do that ... that would be good. 14:24:20 I think that first diagram is probably the scariest bit 14:24:30 I've gotten halfway through, and I've taken notes on each step I take. 14:24:43 skvidal: *laughs* Ya ... maybe put that in the bottom. 14:25:38 I'm in another meeting and need to talk ... so I'm going to be quiet for a bit. 14:25:59 But, I think we are on track to be at 100% by friday unless we actually do find a big bug. 14:26:15 cool. I'll look at testing today and tomorrow. 14:27:01 #info cloud images 14:27:07 #undo 14:27:07 Removing item from minutes: 14:27:11 #topic cloud images 14:27:18 Just some quick notes here: 14:27:55 tdawson: if I wanted to tinker with that process 14:27:56 1. I have a meeting coming up with dgilmore from releng about making this happen as part of the release process 14:28:00 tdawson: and I need help, can I ping you? 14:28:14 skvidal: Sure 14:28:23 tdawson: thx 14:28:25 2. mirror admins are generally happy with the idea of cloud images, or neutral 14:29:02 3. gholms is working on getting cloud-init working so that the raw image size will be small 14:29:09 more on that later. 14:30:02 for now.... 14:30:07 #endmeeting