18:59:54 #startmeeting Cloud SIG 18:59:54 Meeting started Fri Feb 24 18:59:54 2012 UTC. The chair is rbergeron. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:59:54 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:00:00 #meetingname Cloud SIG 19:00:00 The meeting name has been set to 'cloud_sig' 19:00:04 #topic Roll Call 19:00:08 * gholms is in a staff meeting typing on a phone. Don't expect too much. 19:00:12 * mgoldmann says hi 19:00:17 * mdomsch 19:00:18 * rbergeron will chill for a bit while people drop in 19:00:27 gholms: you should make funny faces at mull and gregdek and obino 19:00:43 :) 19:01:28 I don't know your agenda, but you're probably concerned about https://fedorahosted.org/fedora-websites/ticket/83 19:01:30 ;) 19:01:52 * ke4qqq shows up 19:02:06 * shaiton is not sure to be there for the Open Floor 19:02:09 * jforbes is here 19:02:17 hello 19:02:41 shaiton: okay 19:02:55 shaiton: I usually do quite a bit of agenda roulette based on who is here :) 19:03:06 okay, i think we have a handful of people here, so, let's get started :) 19:03:13 #topic OpenNebula 19:03:19 spstarr: you get to go first today. 19:03:23 :) 19:03:46 So we've gone though some churn with Ruby 1.9 causing havoc, which is making upstream rethink dependencies going forward for new versions of OpenNebula 19:04:09 right 19:04:19 right now, im working with them on the opennebula.spec and going though the changes we need in Fedora packaging, other then that we have all the rubygem dependencies for Fedora 17, we will have to drop Ozones feature however 19:04:26 but this not criticial for using OpenNebula 19:04:43 We still need the comps.xml change to get opennebula into the Cloud group 19:04:50 notting: ping 19:04:53 and im working on the systemd/tmpfs/lock bits right now 19:05:02 the rest is just the package itself 19:05:05 oh, he's not here 19:05:10 spstarr: you're a packager- should be easy to make that change 19:05:23 I am :) but i cant review my own package and approve 19:05:34 spstarr: i meant comps.xml 19:05:51 oh just check it out w/ git make change, XMLlint then commit? 19:05:55 yep 19:06:01 dont need to notify anyone? 19:06:07 (well the commit would trigger emails anyway) 19:06:08 see, you should just ask ke4qqq things, because i have no idea :) 19:06:13 lol 19:06:20 for adding a package - no - massive changes - yes please 19:06:29 its, just one line :) 19:06:39 the rest gets pulled in via dependency walking in spec 19:07:10 this weekend i will be on irc with upstream hashing out the packaging bits 19:07:23 spstarr: rockin' 19:07:25 then i turn my home network into a cloud lab to begin testing 19:07:36 http://fedoraproject.org/wiki/How_to_use_and_edit_comps.xml_for_package_groups#How_to_edit_comps 19:07:42 * spstarr bookmarks 19:08:29 #info all rubygem deps are in F17; working on opennebula.spec and is also working through needed changes in fedora packaging 19:08:44 #action spstarr to add opennebula to comps.xml 19:08:48 :) 19:09:01 thats my update for now :) 19:09:17 awesome. thank you :) 19:09:33 #topic AWS and Mirrors and the like 19:09:38 mdomsch: you about? :) 19:09:42 yep 19:09:51 how are you coming along? 19:09:54 http://fedoraproject.org/wiki/Infrastructure/Mirroring/Amazon 19:10:05 I ran into some problems with s3cmd sync code 19:10:15 see bottom of that link 19:10:46 right now the showstopper is that at some point there's a null XML document running around, not sure where it came from or what it's used for, but it causes s3cmd to crap out 19:11:07 this is on syncs after the first one completed 19:11:26 I've gotten an initial sync done, ~77GB pushed over 48 hours 19:11:47 but this s3cmd bug has me stymied right now 19:11:57 I'll raise it with upstream and pray I guess 19:12:00 #info Amazon / MIrroring is coming along; initial sync done, ~77GB pushed over 48 hours; running into problems with s3cmd sync code 19:12:06 Any idea what request is getting the funny response? 19:12:43 gholms, not really - I added some prints to look for it, but it hasn't been helpful yet 19:13:09 the rest is done - DNS CNAME, mirrormanager entry (disabled for now) 19:13:41 dgilmore: are we having alpha up on aws on tuesday morning? 19:13:52 and I still need Max or someone to tell me how to restrict access to a bucket by region 19:13:54 Did you manage to find a way to track ec2's ip ranges? 19:13:59 gholms, yes 19:14:17 #info still need to find out how to restrict access to a bucket by region 19:14:45 spevack ^ 19:14:45 On Tuesday? Crap. I'm still working on a cloud-init fix. 19:15:17 gholms: sorry, we delivered this hot dog on time for a change 19:15:41 #info *we all look at spevack* 19:15:47 No worries. It shouldn't be any worse than it is now. 19:15:59 gholms: is it a srs broken thing? 19:16:46 The bare minimum works. Everything else does not. But that is what the f16 image does, fwiw. 19:17:13 I wouldn't hold anything up for it. 19:17:14 ... okay 19:17:19 gholms: gotcha 19:17:30 mdomsch: are you aiming for any particular point in time? 19:17:36 beta, final, F22, ? :) 19:17:49 rbergeron, not specifically 19:18:08 it's independent of any release schedule 19:18:09 That doesn't have to coincide with a release. 19:18:13 and is a "nice to have" 19:18:40 mdomsch: okay :) 19:19:13 mdomsch: anything we can do here other than tempt spevack's response in the meeting logs? 19:19:32 fond a replacement for s3cmd that works? 19:19:37 s/fond/find/ 19:19:39 Heh 19:19:42 i am fond of that idea 19:19:45 * rbergeron snickers 19:19:53 gotcha. well, thanks for plugging away :) 19:20:12 * gholms will write one eventually for $dayjob, but not yet 19:20:44 #topic Get Fedora webpage updated to include cloud 19:20:53 #link http://fedoraproject.org/get-fedora 19:21:13 #link https://fedorahosted.org/fedora-websites/ticket/83 19:21:16 shaiton: still around? 19:21:29 So we were looking at this for around beta time, AIUI 19:21:32 yep 19:21:50 shaiton: I guess the question is: what information do you need from us to get moving on this? 19:22:05 ok, I just wanted to keep that not forgotten 19:22:12 probably a nice picture 19:22:16 from design… 19:22:19 short text 19:22:25 of a fire engine? or maybe a cloud 19:22:28 probably a cloud 19:22:31 LOL 19:22:35 :) 19:22:46 but it's not urgent, we still have time 19:23:00 http://fedoraproject.org/en/get-fedora-options 19:23:07 actually, I think it is supposed to go in there 19:23:12 * rbergeron apologizes for being confuzzling 19:23:44 yes rbergeron that's what I think too 19:23:47 I guess maybe a cloud tab on that page? 19:24:10 rbergeron, nice :) exactly 19:24:22 (but we need more content then ) 19:24:27 okay 19:24:33 well, i can take a first crack at it i guess 19:25:04 don't bother with html, just give us few text and a picture 19:25:06 i konw there was also some talk of making a more information page similar to ooboontoo's page on ec2 options 19:25:28 #action rbergeron to take a first crack at the get-fedora-options tab for ec2 stuff 19:25:43 more informational, rather 19:25:49 * rbergeron is tired, heh 19:26:07 now this task is in few mind, great :) 19:26:09 shaiton: thanks for the ping on this, it had long since been forgotten by my brain :) 19:26:40 no probs 19:26:47 shaiton: can you add me to the cc or make me hthe owner of that ticket? 19:26:51 if you still have it open? 19:26:59 i will 19:27:07 thank you :) 19:27:17 #action shaiton to add robyn to the websites ticket 19:27:41 #topic Cloudstack/Euca/AS7 dependency fun 19:27:56 ke4qqq, mgoldmann, mull who is not here becuase he's in a staff meeting: how is this stuff going? 19:28:11 soooo :) 19:28:18 we're making progress - iirc we have 2-3 4-th level java deps that still aren't approved 19:28:27 My current work is to make AS7 usable (still) 19:28:44 part of it compiles, and boots, which is a good sign 19:28:56 fortunately jboss/ovirt/euca/cloudstack all share sooo many deps. 19:28:57 with every day/week I'm adding new modules 19:29:07 mgoldmann: i think you should work on getting other folks from jboss-land to also know how to do things in fedora so you're not forever stuck as "that guy who knows about fedora" :) 19:29:12 mgoldmann: Keep up the good work, please! And let us know if we need to publicly shame anyone :-p 19:29:13 for example this week I added security and transactions 19:29:22 which both are pretty important 19:29:32 * jsmith solved a jboss problem this past week, so they owe us one :-p 19:29:33 I think I'm close of making oVirt running on what we have 19:30:00 rbergeron: it's not so easy :) 19:30:02 mgoldmann: i am sure that will please them as well ;) 19:30:06 mgoldmann: i know it :) 19:30:27 jsmith: I'm going to write a blog post about it and maybe do some presentations on fudcon(s)/judcon(s) 19:30:40 mgoldmann: That would be fantastic 19:30:52 I'm keeping notes :) 19:30:59 * gholms had a couple euca deps done this week 19:31:01 ke4qqq: i assume all those java deps are actually assigned for review and so forth? 19:31:05 gholms: ;) 19:31:10 I woudl like to extend spot's weighting on bad projects 19:31:19 mgoldmann: HAHA 19:31:23 * rbergeron snickers 19:31:23 and do something for java :) 19:31:27 Heh 19:31:29 like this* 19:31:34 mgoldmann: i am sure he would welcome those patches 19:31:41 heh :) 19:32:00 mgoldmann: /me would love to see such a thing - and use it. :) 19:32:00 so, In general - I'm busy as hell on packaging 19:32:27 and so everyone hears - thanks everyone for being supportive! 19:32:54 btw, the plan is to package web profile for F17 19:33:02 mgoldmann: and thank you for plugging away, i know it is probably not the most thrilling thing on earth :) 19:33:02 I'm short on time, but I hope I'll make it 19:33:14 indeed :) 19:34:07 well, rockin. i'm glad you guys (in a twisted weird way) are at least able to share the pain 19:34:12 and willing to do so :) 19:34:22 so props to all of you :) 19:34:33 we need to have a party afterwards :) 19:34:48 mgoldmann: a FAD you mean :) 19:34:50 #info still a few dependencies left shared by euca/cloudstack/as7, but making progress 19:35:01 #info as7 is compiling and booting partially, which is awesome 19:35:12 but really, considering the amount of new features in Fedora over the last year - maaan, it's huge :) 19:35:25 indeed 19:35:33 #info plan is to package web profile for F17 19:35:38 FROM ZERO TO HERO 19:35:39 or something 19:35:40 ;) 19:35:52 nah, we had kernel 19:36:17 from .03 to hero! 19:36:23 okay, well. moving on a bit 19:36:29 (thanks, guys!) 19:36:32 #topic Open Floor 19:36:46 * rbergeron is willing to take any other topics at this point from people who are here 19:36:57 hello, I've been lurking 19:36:58 updates welcome 19:37:01 nilsson: hi! 19:37:09 Kaleb's at Connectathon, improving GlusterFS/NFS compatibility. 19:37:11 I'm looking for any junior jobs that might be available 19:37:49 jdarcy: ooh, that's right 19:38:24 iirc, there was mention on the wiki of needing to package noVNC 19:38:24 nilsson: as far as cloud sig work? we usually have a ton of people looking for packaging help - if that's your cup of tea, i would certainly advertise your interest on the cloud sig mailing list. 19:38:42 https://admin.fedoraproject.org/mailman/listinfo/cloud 19:39:01 nilsson: you should look at mull's eucalyptus page in the wiki 19:39:13 he is going to need tons of help from a review perspective 19:39:15 we also need some documentation help, if that's more interesting 19:39:20 spstarr needs reviews 19:39:24 etc. 19:39:53 well the doc work is something I'll be able to do 19:39:58 yes I will need :) 19:40:00 I don't have any packaging experience 19:40:11 nilsson: if you're cool wit reviewing java packages, join #fedora-java and we'll help you with everything (eucalyptus, as7, ovirt deps) 19:40:22 nilsson: well if you decide you want to build the experience we can help there too. 19:40:28 +1 19:40:38 * rbergeron laughs at everyone being like OMG PACKAGING PLZ PLZ YES 19:40:44 heh :) 19:40:48 mgoldmann and I are sponsors and happy to sponsor folks who want to help 19:41:05 * mgoldmann is laughing at being "sponsor" 19:41:29 nilsson: if doc work is interesting, maybe we can fix you up with sparks and he can point you in the right direction for getting started 19:41:38 sure 19:42:01 Just a quick update on openshift. We are still targeting F18 even though we're still a little quiet about it. 19:42:18 I have to head out now, I'll join the mailing list and see you guys there 19:42:20 tdawson_: hey, that's nice! 19:42:43 nilsson: maybe a first good step would be to send a mail to the mailing list (link i posted above) advertising your interest and introducing yourself :) 19:42:52 tdawson_, is openshift/cloudforms open source? I couldn't find anything that would indicate yes 19:42:55 nilsson: thanks for coming! :) 19:42:59 rbergeron, will do 19:43:05 I don't know about cloudforms 19:43:06 tdawson_: i bet you never get that question :) 19:43:13 haha 19:43:31 But openshift stems from a company we bought, so we've been working on open sourcing all that. 19:43:40 ok, thanks 19:43:40 nilsson: cloudforms is based in part off of aeolus, which is in fedora and is open source 19:43:43 tdawson_: timeline yet? 19:43:50 aeolusproject.org would be a good place to read aobut it 19:44:01 and also based on pulp, kateloo and candelpin ( and that's also open ) 19:44:01 No ... they haven't given me a definate time ... hopefully in this first quarter. 19:44:03 ok, see you on the mailing list 19:44:16 tdawson_: who's going to be packaging? 19:44:31 But it should be open sourced in time for us to get everything ready for F18. 19:44:55 tdawson_: do you have any rough list of deps? 19:44:58 tdawson_: it's all in java, right? you should prolly start buying mgoldmann beers now. :) 19:45:06 I know I'll be part of that team, I believe russell harrison. 19:45:13 ke4qqq: you mean cases of beers, right? 19:45:19 Actually ... the openshift code is written in ruby. 19:45:26 mgoldmann: yes. 19:45:34 hmm 19:45:35 But we have cartridges, and one of the most popular is java/jboss 19:45:39 ahhh - heard theywere using jclouds at one point, so assumed from there 19:45:40 tdawson_: might be a good time to start thinking about learning how to package so you can be a packager when it's time - there are a few people here who could sponsor you 19:45:54 if you're willing ot pick up a package or two and do some reviewing :) 19:46:02 * rbergeron cna't remember if russell is already one or not, /me looks 19:46:13 tdawson_: the biggest issue will be in making as7 mulit-tenant in a good way 19:46:31 russell already has been sponsored, and I think his first package might have been approved by now. 19:46:33 oh, he is, that's good 19:46:51 * rbergeron sees packager in his fasinfo 19:47:16 I'm working on rubygem-mongo, but I keep getting swamped with other things ... 19:47:51 tdawson_: you might take a look at some of the wiki pages others have made (for cloudstack, euca, as7, etc) to see how they started their prep-for-packaging-dependency-fun 19:48:07 just ... priming you for being able to get going :) 19:48:15 i know there are plenty of people who will be excited to see openshift 19:48:30 that might even be willing to help out ;) 19:48:44 Will do. 19:49:16 tdawson_: thanks for coming :) looking forward to seeing you and rharrison here more! 19:49:42 * rbergeron feels like she is 12, she found a blow pop and had forgotten how tasty they are 19:50:02 (though i probably shouldn't raid candy my kids got at school for valentine's day) 19:50:19 okeedokee, anyone else? 19:51:16 [Time passes] 19:51:45 [gholms makes awesome funnies] 19:51:46 ;) 19:51:56 :3 19:51:59 *hug* 19:52:05 okay, anyone, anyone, bueller 19:52:17 Thanks, everyone! 19:52:18 * rbergeron will strap on the jetpack and start counting backwards from 87 to end meeting 19:52:25 42 19:52:26 21 19:52:27 18 19:52:28 16 19:52:40 #endmeeting