21:02:18 #startmeeting Cloud SIG 21:02:18 Meeting started Thu Jan 6 21:02:18 2011 UTC. The chair is rbergeron. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:18 Useful Commands: #action #agreed #halp #info #idea #link #topic. 21:02:18 and not have this problem? 21:02:18 -1 is empty. 21:02:22 OK meeting -3 21:02:31 #meetingname Fedora Cloud SIG 21:02:31 The meeting name has been set to 'fedora_cloud_sig' 21:03:09 * gholms runs off to the microwave 21:03:15 gholms :) 21:03:18 popcorn time? 21:03:24 meeting is continuing in -3 for infrastructure. 21:03:24 #topic Roll call! 21:03:26 Lunch time. 21:03:27 rbergeron, later 21:03:28 who's about today? 21:03:34 * jforbes is here 21:03:34 * clalance waves 21:03:36 * mmorsi says hi 21:04:19 well. alllllrighty then 21:04:30 #topic EC2 21:04:38 jforbes: what's the word? :) 21:05:02 Not much, I just got back yesterday... 21:05:25 I did manage to get the koji permissions and information to get things rolling, so working on that now 21:05:26 * rbergeron notes that we had a mail over the holiday break about large instances on EC2 being unresponsive 21:05:41 rbergeron: 21:06:00 dgilmore: yessir 21:06:16 21:06:21 jforbes: have we given more thought to doing a test day? 21:06:28 Dates are still available. 21:06:28 rbergeron: sorry was going to say i was here 21:06:34 If we know what we're testing. 21:06:40 my modem just rebooted 21:06:48 dgilmore: awesome 21:07:18 rbergeron: I can schedule the test day, but we need test plans lined up 21:07:41 jforbes: agreed. 21:07:56 rbergeron: so far, I think we need test plans for basic EC2 images, boxgrinder, openstack? 21:08:34 I think we need to circle up with silas on where openstack is at. I'll look at it a bit later in the meeting - but, yeah, I think that's a basic list. 21:08:47 jdarcy may also be interested in testing CloudFS. 21:08:53 Maybe we can do something with deltacloud core as well. 21:08:58 clalance: I don't know if you have any thoughts on deltacloud 21:09:03 aha, jinx. 21:09:10 rbergeron: The aggregator portion is just not going to be ready. 21:09:36 There is one additional component (oz/imagefactory) that I might get in before the deadline. 21:09:40 But I'm not certain about that. 21:09:43 clalance: yeah, I figured. 21:09:59 The core, though, is already in, so probably bears testing. 21:10:33 jforbes: how about we do this: I'll shoot a mail to the list about test day ideas. Maybe you can throw sometihng in about what you'd like to do with EC2 - do you think you'd have time to write up some test plans for that, and we can see what other folks want to contribute for their respective cloudy features? 21:10:33 rbergeron: Also, FYI; we are going to be renaming the aggregator in the very near future. 21:10:37 To reduce confusion with the API. 21:10:47 rbergeron: sure 21:11:10 clalance: :) any confusion-reducing is proably good there. 21:11:15 Seriously. 21:11:31 A website revamp will go along with that. 21:11:37 I'll announce this in Tempe as well. 21:11:46 (I hope to have the rename done by then) 21:12:05 clalance: excellent. I'm glad you're getting to come, btw. 21:12:21 #action rbergeron send an email to the list re: test days brainstorming 21:12:25 I tried getting more Eucalyptus people to go, but that's apparently not going to happen. :( 21:12:27 Yeah, my first business trip in a while :). 21:12:49 #info jforbes is working on getting koji permissions and information to get EC2 stuff rolling. 21:13:04 rbergeron: jforbes has the koji permissions 21:13:14 he just needs to get me a kickstart file 21:13:25 e can do his testing of the kickstart 21:13:34 yup, have the permissions, just need to finish testing and get a file out 21:13:41 so i can put it in spains-kickstarts with the other fedora kickstarts 21:13:52 spins-kickstarts even 21:13:53 #action jforbes to do some testing, get dgilmore a kickstart file. 21:14:09 Awesome. You guys rock. 21:14:19 Anything else ec2-ish? 21:14:40 none else here 21:15:24 sounds good. 21:16:24 mooooooving on.... 21:16:28 #topic CloudFS 21:16:36 jdarcy: how's feature-land going? 21:16:39 Hm, let's see. 21:16:58 Little progress on writing code since the last meeting. 21:17:11 I did get some EC2 performance results showing that the new functionality doesn't cost much. 21:17:18 Generated lots of content for CloudFS.org 21:17:45 We had meetings with Gluster today, some partner-relationship kindsa stuff I can't talk about but a few technical tidbits. 21:18:14 They've announced their own fairly limited form of (active/passive, file-granularity) replication, which is good for DR but not much else. 21:18:37 We talked about how to migrate/converge toward the kinds of replication CloudFS is targeting. 21:18:52 Cool! 21:18:56 Some talk about multi-tenancy and authentication (they're getting a lot stronger there, driven mostly by CIFS/AD). 21:19:15 They're actively reviewing some of the CloudFS bits now. 21:19:23 I think that's about it. 21:20:01 Do you think you're still on-track as far as feature completion goes? 21:20:22 :D 21:20:28 I think so. This week is being chewed up by other stuff, but after that I have a pretty clear shot to feature freeze. More eyeballs would be nice, though. 21:20:45 * rbergeron nods 21:20:47 And if anyone wants to volunteer to help with building/packaging kindsa stuff (about which I know zip) that would be a huge help. 21:21:09 we were talking before you got here also about doing test days - I just sent a mail to the list about it. 21:21:17 Excellent. 21:21:27 But that's some food for thought for you as well, if you want to do some test plans and whatnot. 21:21:33 * jdarcy nods. 21:21:47 I would imagine if you shot a mail to the cloud SIG list looking for some packaging help, you might find some. :) 21:21:54 Will do. 21:21:58 at least, it's worth a shot. 21:22:12 Anything else? :) 21:22:26 Not that I can think of. 21:22:33 coolio. 21:22:41 #topic Deltacloud 21:22:51 clalance, mmorsi: anything fun going on? 21:23:13 rbergeron: Nothing hugely new since we last spoke. 21:23:26 yea alot of dev work really 21:23:27 rbergeron: We have someone packaging up more of the dependencies to get into Fedora. 21:23:47 Yeah, other than that, more development, but nothing hugely new. 21:23:58 also a better web interface is in teh works, also a better installation and configuration tool 21:24:01 The biggest externally visible thing will be the name change. 21:24:06 Oh yeah, that stuff too. 21:24:42 we're tightening up alot of hanging ends and what not too, tackling security over deltacloud as we speak 21:24:53 all of this should be in the next release we hope 21:25:01 Are you guys still planning on doing part of deltacloud as a feature in F15, and if so, is someone going to be making that feature page before the submission deadline on 1/25 :) 21:25:31 rbergeron: I think not. 21:25:44 clalance: oh? we're not adding the core at least? 21:25:49 rbergeron: As we discussed before the break, we have a lot of dependencies. 21:25:51 mmorsi: That is in already. 21:25:57 ah aight 21:26:03 Not sure if it needs a feature page, though maybe that will get it some additional exposure. 21:26:13 clalance: that's the idea. 21:26:16 rbergeron: What do you think about that. 21:26:19 sure thing, fine w/ me either way 21:26:32 rbergeron: OK. Let me ping a couple of people to make the feature page for the core. 21:26:46 rbergeron: As far as the aggregator, I think we are just going to continue to get dependencies in. 21:26:55 clalance: I think it would be fabulous, both from the getting you guys some extra PR perspective, as well as getting the Cloud SIG some extra marketing love. 21:26:59 With any luck, condor 7.6 will be ready for F-16. And then we can do a feature for the aggregator. 21:27:03 Yeah. 21:27:05 Plus testing and so forth. 21:27:29 You can put an action item for me for that. 21:27:35 #action clalance to investigate finding someone to do a feature page for deltacloud core 21:27:43 * rbergeron was just in the process of doing so :) 21:27:44 Oh, I guess I could just do it ;). 21:27:57 actually, i think you have to be chaired, but i'm not a meetbot expert, per se. 21:28:06 Ah. 21:28:06 I just type things, and usually it all comes out in the wash. 21:28:10 LOL 21:28:11 Love it. 21:28:15 shout out if you need help w/ that, owned the F14 feature for Ruby 1.8.7 21:28:31 also the F15 feature for Rails 3 is still in the work 21:28:32 works 21:28:35 mmorsi: I'm going to make lutter or mfojtik do it ;). 21:28:46 mmorsi: yes please, send that in ;) 21:28:46 which we could also prolly tied into deltacloud, eg that can be one of the first major use cases 21:28:59 like u said, the more exposure the better :-) 21:29:03 sure thing 21:29:09 mmorsi: Yeah, but it doesn't need to be tied. 21:29:20 agreed, just a mention 21:29:23 mmorsi: If we get rails 3 into F-15, we should mark that as its own feature. 21:29:25 Yeah, I think rails 3 stands on its own :) 21:29:28 agreed 21:30:20 But - like I said - we're getting there pretty quickly - just 2.5 more weeks. 21:30:35 #info Feature Submission deadline is 1/25 21:30:39 ya hear the clock ticking :-p 21:30:53 mmorsi: I do now. >:( 21:31:02 :P 21:31:05 lol 21:31:08 haha. 21:31:15 * rbergeron hears it quite a bit :) 21:31:32 Anything else from deltacloud land? 21:31:41 we need an irc plug that just spews out "tick-tock", keeps us on our toes :-p 21:31:53 nothing from me 21:32:20 Yeah, that's about all I have. 21:32:26 Thanks for the updates, folks. 21:32:33 #topic OpenStack 21:32:42 * rbergeron is going to peek in on that bz and see where it's at. 21:33:15 #link http://fedoraproject.org/wiki/Cloud_SIG/OpenStack 21:33:41 #link https://bugzilla.redhat.com/show_bug.cgi?id=649495 21:33:51 Okay - looks like silas is still waiting on python-carrot. 21:35:24 #info if someone is interested in helping out with OpenStack, there are some reviews still outstanding (python-carrot) 21:36:03 #link https://bugzilla.redhat.com/show_bug.cgi?id=611054 21:36:26 * rbergeron would really like to see us be able to get this in - I'll send an email to silas and see if there is anything else he's hung up on. 21:37:15 Okay. Moving along 21:37:19 #topic Open Floor 21:37:26 Anyone? :) 21:37:47 rbergeron: What do we have planned for Tempe? 21:37:52 (cloud-wise, that is) 21:37:54 * rbergeron notes that FUDCon is coming up, of course, lest you think I would let y'all go for a week without a reminder - if you'd like to do Cloud SIG stuff, i'm taking ideas. 21:38:02 Heh ;). 21:38:03 clalance: you have an uncanny way of being in my head today. 21:38:06 Seriously. 21:38:10 * gholms steals rbergeron's '/' key 21:38:13 great minds think alike :-) 21:38:29 Soooo - there are a few cloud-related proposals already on the wiki. 21:38:32 * rbergeron pulls it up real quicklike. 21:38:36 Hm. Need slides. 21:38:39 ... 21:38:47 jdarcy: Yeah, me too. 21:39:03 rbergeron: How does that work, by the way? How is it chosen which sessions are given? 21:40:01 zaitcev has someting on Fedora Cloud storage.... sdake has something on sheepdog.... jdarcy has something on cloudFS...clalance has something on cloud management. 21:40:06 * rbergeron looks around 21:40:11 dgilmore: you still here? 21:40:26 rbergeron: si 21:40:27 or someone wanting to give a better explanation than I can of how the sessions get chosen / organized 21:40:54 rbergeron: for fudcon? 21:41:04 dgilmore: Yes. 21:41:11 rbergeron: everyone pitcjes their session idea 21:41:13 it's a little known secret that I have never been to a FUDCon. So this part is still somewhat mysterious to me, but from what I understand, it's very barcamp style - people will be pitching ideas in front of others in the morning, and we'll be doing the session grid onsite. 21:41:20 then the attendees vote on them 21:41:23 Ah, OK. 21:41:30 dgilmore: OK, thanks. Easy enough. 21:41:38 they then get ordered so that the most people can get to what they want 21:41:41 we then do them 21:41:46 dgilmore: So basically the first session is used to choose the rest? 21:41:56 jdarcy: yeah 21:42:03 its done bar camp style 21:42:04 * jdarcy thinks about scheduling algorithms. 21:42:15 yeh, most often you pitch the idea, and write it on a eice of paper, all of which are collected at the end of the first session and put up on a wall 21:42:20 we have the welcome/pitching sessions session 21:42:32 then its a big jungle of voting and sorting 21:42:36 I was basically do a 2-parter about Hail and Swift, how Fedora can be used for it now that we have packages. 21:42:50 followed by an easy to follow schedule where you are free to choose your own path 21:42:51 then everyone is given a chance to vote on whichever ones by putting a tally mark on the papers they are interested in 21:43:04 mmorsi: correct 21:43:07 Does bribery work? 21:43:11 I think we have a big enough cloud contingent that we'll all be going to each other's proposed sessions. 21:43:14 * mmorsi gave the deltacloud talk at toronto (combined w/ jforbes ec2 talk) 21:43:15 jdarcy: it does not 21:43:20 you can propose multiple talks too 21:43:24 jdarcy: ponies have been known to work though 21:43:25 dgilmore: what? bribery works for me all the time. :) 21:43:32 * rbergeron has an arsenal of cookies. 21:43:33 rbergeron: your special ;) 21:43:38 dgilmore: ;) 21:43:51 I guess I'll just have to focus on content. Darn. 21:44:52 jdarcy: Well I have your comparison with Ceph, maybe you just read that aloud ^_^ 21:45:06 jdarcy: lol 21:45:13 I'll re-do it as an epic poem. 21:45:20 YES 21:46:18 the pitching is honestly pretty fun 21:47:02 Anyway, to circle back a bit, I think we have some good opportunities to talk about the individual features people are working on, but I htink we could also think about organizing something around "how this all works together" 21:47:14 since, I think for most poeple, that's the big mystery of cloud things. 21:47:27 How does it work together, if it even does, and if it doesn't, what can we do to get it there, etc. 21:47:44 rbergeron: Yeah. Some of the additional pieces we are building for aggregator fit well in there. 21:47:51 * rbergeron nods 21:47:55 (there is a project called audrey to do configuration-like things that I would like to talk about) 21:48:07 I think Chris's talk will shed some good light on that side of things. 21:48:14 And I think it's probably also a good place to run it by people who are actual users of the stuff, who may have good feedback as well. 21:48:20 Yeah, definitely. 21:48:21 Or potential users. Etc. 21:48:29 We are missing some of that. 21:48:56 So if someone wants to organize or propose a session around "How does all this cloud stuff work together," Or ... well, anything else, really, I think that would be a great idea. 21:49:27 rbergeron: probably a good hackfest thing also 21:50:09 dgilmore: yeah. Do you think that would be best in tha tformat? 21:50:37 rbergeron: could be good for both 21:50:58 or maybe have a question and answer session 21:51:06 where we get all kinds of dev folks 21:51:19 and the viewers can ask about how things work 21:51:28 you know put the devs on the spot and all 21:51:39 OH. i love tha tkind of thing 21:51:46 mostly because i'm not a developer ;) lol 21:52:25 Maybe it's a good question to take to the mailing list. 21:52:35 Heh, nice. 21:52:36 * clalance runs 21:52:52 There is also the option of getting the cloud folks together and maybe going out for some f00d together one night. 21:52:54 ;) 21:53:09 That works too 21:53:18 * gholms is all for an all-hands meeting 21:53:27 * rbergeron knows the secrets of getting people to show up. Food. 21:54:03 #action rbergeron to send a FUDCon prodding message about "what to do!" to the cloud sig list. 21:54:11 Anyone else have anything? :) 21:54:21 * dgilmore would like to sit down with jforbes and other to talk about floud and releng 21:54:26 but at fudcon 21:54:33 dgilmore: hackfest? :) 21:54:56 rbergeron: probably more a intimate discussion 21:55:02 but could be a hackfest 21:55:10 hallway track? :) 21:55:18 We'll figure it out. 21:55:21 probaly hallway track 21:55:23 or hackfest 21:55:43 yup. we will certainly do something there 21:55:51 okay. 21:56:02 * rbergeron looks around for any other comments before closing things out 21:56:56 Thanks for coming, all! 21:57:03 #endmeeting