22:04:21 #startmeeting 22:04:22 Meeting started Tue Aug 23 22:04:21 2011 UTC. The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:04:23 Useful Commands: #action #agreed #help #info #idea #link #topic. 22:04:35 agenda: http://wiki.openstack.org/Network/Meetings 22:05:01 #topics nestack general topics 22:05:21 #info congrats to the whole team.... quantum was accepted as an official OpenStack incubated project today 22:05:29 congrats ! 22:05:36 allright :) 22:05:40 hearty congrats! 22:05:40 * markvoelker cheers and looks around for champagne 22:05:43 nice 22:05:51 congrats 22:06:01 with the significant participation from folks from multiple different companies... I think it was an easy call for them. Great work folks... take a minute to relax, then get back to working on D-4 reviews :) 22:06:02 Great News!! 22:06:31 mark: we'll have to bring some to the essex summit to toast :) 22:07:00 danwent: I dunno...last time I tried to bring drinks on a plane the TSA got frisky. =p 22:07:11 in other news, we'll be moving to the new github/gerrit setup post diablo 22:07:17 if you're curious, read up on: http://wiki.openstack.org/GerritWorkflow 22:07:36 best to identify any concerns you have early and we can work with the openstack team to figure out solutions 22:07:39 post diablo - cheers to that as well! :-) 22:07:39 danwent: do we have a leader for the migration effort? 22:07:46 i.e.: somebody to bugger with question? 22:07:50 salv: are you volunteering ;) 22:07:51 ? 22:08:03 jaypipes said he would help out 22:08:05 not really, I'd like to focus on blueprints for essex release :-) 22:08:12 mtaylor would be another guy to ping. 22:08:19 yes 22:08:20 what? 22:08:25 salv: no problem.... I was planning on driving this 22:08:29 you helping with gerrit and migration to git 22:08:38 sent an email to mtayler and heckj this morning. 22:08:46 yes. also jeblair will be involved 22:08:50 mtayler -> mtaylor 22:09:06 danwent: and - I haven't forgotten about your email about tarballs - I just had some world explosion earlier this morning I'm still cleaning up 22:09:19 danwent: I can help too if you want 22:09:20 #action danwent: send out email to folks helping with move to github 22:09:26 bhall: thx 22:09:42 mtaylor: no worries... tuesdays are always very busy days, I understand :) 22:09:47 <_0x44> jaypipes: ping 22:09:50 danwent: (sorry, I've got like, 5 todo list items for you guys) 22:10:01 mtaylor: we appreciate the help 22:10:19 one other general netstack point: essex summit signup is now open 22:10:31 #info sign up for essex summit if you can: http://summit.openstack.org/ 22:10:55 we'll definitely be having discussions about blueprints for quantum, and it is generally a great chance to meet team members face to face. 22:11:07 so you can put a face to the reviewers name :) 22:11:16 mine is particuarly ugly... 22:11:26 spots are limited and I think they ran out last time, so don't delay 22:11:54 salv: but your code is beautiful, so it makes up for it, right? :P 22:12:09 Ok, it troy around? 22:12:17 yes 22:12:21 About Essex Summit: Developers get a free pass to attend the conference I believe. 22:12:23 #topic melange 22:12:24 danwent: you're too kind. Anyway seats are going away fast. 1 per minut. 22:12:45 At this pace, the essex summit would be a sold out in 3 hours! 22:12:47 we did the initial merge prop for Melange and have gotten a great deal of feedback. 22:13:18 we're working on creating an OpenStack commons code area to address some of the concerns 22:13:30 we're also cleaning up the cli etc. 22:13:51 cool. over the weekend I did a first cut at melange support for the quantum manager. 22:13:51 it's too big of a change to get into Nova in Diablo but we should be setup for early Essex 22:14:24 danwent: we were just granted feature-freeze exception for quantum manager, is that correct? 22:14:25 plan is to focus on the merge prop and the quantum/nova integration 22:14:44 Ok. and if we get the quantum manager code in nova, it should still be possible to run nova with melange, i believe, you will just need to download melange separately... is that correct? 22:14:57 salv: correct 22:15:14 i believe so 22:15:23 I just created a new version of the qmanager branch that is not stacked on top of melange... still needs testing: 22:15:38 https://code.launchpad.net/~danwent/nova/qmanager-new 22:15:46 but you can look through the diff to get a sense of what it does 22:15:55 feedback appreciated. 22:16:16 ok troy, thanks for the update. I must say, the melange code has been rock solid in my testing against it. 22:16:25 any time there was a bug, it was in my code calling it :) 22:16:44 i hope we can start getting some scale testing done on the integration soon 22:17:00 we should be in good shape for that 22:17:03 agreed. 22:17:10 ok, anything else on melange? 22:17:41 #topic donabe 22:18:04 is debo or someone else working on donabe here? 22:18:22 I'm here, but don't have an update. 22:18:36 I'll see if I can get Rick or Debo to send an update to the alias. 22:18:42 #action #debo send update on donabe to email list 22:18:48 james: thx 22:18:51 #topic quantum 22:19:14 first off, congrats to everyone on incubation... 22:19:30 we've been making great progress on getting through the D-4 reviews 22:19:46 great work from many different folks on the team 22:19:53 https://code.launchpad.net/quantum/+activereviews 22:20:13 my understanding is that the cisco reviews are all stacked on each other, so they aren't quite as massive as they look. 22:20:18 is that correct? 22:20:30 yes 22:20:45 dont be too alarmed :-) 22:21:01 Sumit: great. As I mentioned in the email this morning, I'm confident that we can get these all into diablo, even if they aren't quite done by the D-4 deadline 22:21:06 https://code.launchpad.net/~salvatore-orlando/quantum/quantum-api-alignment/+merge/70788 22:21:10 I think the consolidation branch is stacked on top of the other three, which are independent 22:21:16 yes 22:21:34 there is one on the VIF-driver and Scheduler which is independent of the others 22:21:35 https://code.launchpad.net/~salvatore-orlando/quantum/quantum-api-alignment/+merge/70788 22:21:37 I had a good deal of feedback, especially from Simik 22:21:49 Somik, sorry. 22:22:13 Concerns have been addressed and updated code pushed to the branch. I.m pretty sure we can have it merged by thirsday. 22:22:18 Ok, anything we need to discuss here? 22:22:30 sounds like you're on top of it. 22:22:30 salv: I think both of your branches are pretty close, i'll do the final review soon. 22:22:51 somik: thanks. 22:22:53 salv: Great work on both those branches! 22:23:04 any significant concerns with any of the reviews that we need to discuss here, or can we just keep plugging away? 22:23:37 sounds like we're in good shape. 22:23:42 Not from the reviews I've done. Rohit updated the code after my review, and I think it should not be too much of an issue to merge that branch, 22:23:54 I have not received a reply on the vif-scheduler review, though. 22:24:17 thanks Salv, yes I am in the process of fixing that 22:24:25 good. 22:24:25 salv: Rohit and I made some changes based on your reviews.. Thanks a lot! 22:24:27 vif-driver/scheduler that is 22:24:51 salv: that particular branch is independent of other cisco branches 22:24:51 ok, great work folks. 22:24:53 We even increased our pylint score :-) 22:24:55 thanks salv again for the review...and good to hear that the branch is in good shape for merge 22:24:58 so it can go in even later 22:25:17 Ok, onto talking about Nova Integration with Quantum 22:25:18 SumitNaiksatam: very good 22:25:38 so Ryu's branch to add an extension to nova to expose vif uuids went in. 22:25:51 this should give us what we need for the GUI, is that correct arvind? 22:26:16 yup.. I integerated that into the dashboard waiting a merge on openstackx https://github.com/cloudbuilders/openstackx/pull/16 22:26:32 cool! 22:26:45 Does this extension also allow us to track to which instance a particular VIF belongs? 22:26:58 salv: I believe so 22:27:15 danwent: something worth looking at for our issue with authorization 22:27:55 yeah... we discussed that. Could definitely be used for that, at least in the short-term. 22:28:17 great 22:28:33 QuantumManager got an extension to try and get into diablo... so more coding for me :) 22:28:48 https://code.launchpad.net/~danwent/nova/qmanager-new 22:29:02 works with two modes for IPAM, melange, and the traditional nova DB IPAM. 22:29:24 danwent: great choice 22:29:27 it also supports a tushar's extension to allow specifying networks when creating a VM. 22:29:39 that could be something that we leverage in the GUI as well in the future. 22:30:09 the os-create-server-ext extension 22:30:31 other than that, wanted to touch on packaging, is tyler here? 22:30:59 once we get all of the existing reviews in for diablo, we'll need to figure out how to switch the directory structure around. 22:31:04 danwent: nope...this meeting time doesnt' work with his schedule unfortunately 22:31:40 ok, i encourage people to check out the branch: https://code.launchpad.net/~tylesmit/quantum/quantum-packages 22:31:59 I have some questions that i'll be sending to the list soon. 22:32:29 Only other agenda item is to start thinking about documentation.... we've done an awesome job writing code, now we need to make sure people other than us can use it :) 22:32:47 Both docs for administrators (i.e., quantum users) as well as for developers. 22:32:56 On this regard there was an interesting thread in the Openstack ML on documentation for openstack API 22:33:08 savl: indeed, thanks for bringing that up 22:33:19 vishy was proposing to embed the documentation in source code, and I agree with this approach. 22:33:32 Maintaning a separate wiki page for API spec has been quite cumbersome for me. 22:33:42 salv: +1 having an approval process to update an API spec seems very valuable 22:34:06 that way people are much more likely to see changes that are relevant to them. 22:34:20 +1 on the approval process. Second troubling issue was getting formal feedback and approval on the contents of the wiki page :) 22:34:35 I'd encourage folks to participate in that thread if they have opinions... =) 22:34:50 If I'm not wrong Keystone already has documentation in source tree, in rst format 22:35:10 salv: I think this is a great idea. Can you create a blueprint/bug for this? 22:35:16 plan to have the same thing for Quantum API. Will be a blueprint with integration freeze as target a good idea? 22:35:25 danwent: sorry did not see your last line 22:35:35 Okay, please put an action for that 22:35:35 salv: if you can do it quickly, I think that would make sense. 22:35:54 #action #salv will create a blueprint to put API spec into the source tree 22:36:00 Sure, I think it can be achieved by next tuesday 22:36:13 salv: thanks for bringing that up 22:36:56 ok, and just on general docs, I think we have some work helping the rest of the nova team understand how quantum fits in... some of this material has already been created, we just need to present it in an organized form. 22:37:34 now that we have a mechanism for nova/quantum integration (actually... two mechanisms), we can create some tutorials which will help a lot. 22:37:43 ok, anything else on quantum? 22:37:47 quick update on the dashboard: it's complete with vif plugging .. just waiting on a couple of merges 1) Merge to the main dashboard trunk https://github.com/4P/openstack-dashboard/pull/92 and 2) The openstackx class to grab VIFS https://github.com/cloudbuilders/openstackx/pull/16 22:38:20 that was quick :) nice work asomya 22:38:36 #topic open discussion 22:38:48 There's a openstack meetup in the bay area tonight 22:39:14 I'd love to attend, but my teleport is broken :) 22:39:17 http://www.meetup.com/SVForum-CloudSIG/events/18687151/?a=md1.2p_grp&rv=md1.2p 22:39:37 salv: you'll be in boston though, right? 22:39:43 You west coasters have all the fun...we east coasters just get earthquakes and hurricanes this week. =p 22:39:54 danwent: yes, my passport has been renewed! 22:39:58 yeah, that earthquake was crazy 22:40:02 salv: good to hear :) 22:40:05 ok, anything else? 22:40:08 Tyler and Arvind, which are not part of the core team, are giving a great contribution in terms of reviews (and code) 22:40:26 would they like to be part of the core team? 22:40:28 shall we consider an extension of the core team? 22:40:31 (wink, wink?) 22:40:41 I'm definitely in favor. 22:40:42 +1 22:40:45 if they agree, it's +1 from me 22:40:46 danwent: do we have to bribe someone for that :) 22:40:46 +1 22:40:56 +1 22:41:06 +1 22:41:11 +1 22:41:14 +1 22:41:15 Ok, seems like approval to me. 22:41:28 if nobody complains in 10 seconds, it's approved 22:41:43 this is great point. please chime up if you are not a core but want to nominate yourself... or if you're shy, just talk to someone who will then nominate you 22:41:43 approved! 22:42:20 sometimes people act like they are already in core, and we forget that they are not :) 22:42:23 ok, anything else? 22:42:23 Another little thing on documentation: it would be great if we can provide use cases and reference architectures for quantum. 22:42:42 For instance, how to build the network topology for a three-tier app 22:43:00 salv: once the quantum manager gets merged in, that would be the next step. 22:43:09 but you are correct. 22:43:20 somik: good, we had a blueprint for doing something similar 22:43:30 https://blueprints.launchpad.net/quantum/+spec/use-case-milestones 22:43:36 salv: agreed. That demo link I sent out showed what is probably one of the most common use cases. 22:43:59 danwent: demo link? 22:44:15 yeah, I must have missed it as well... 22:44:16 salv: I actually cleared that out once we started using launchpad milestones 22:44:26 it was buried in my update email to the openstack list... 22:44:56 oh ok will search 22:44:58 found it: http://wiki.openstack.org/QuantumOVSDemo 22:44:58 have been meaning to do a screencast, but until then the text will have to do: http://wiki.openstack.org/QuantumOVSDemo 22:45:13 thanks guys! 22:45:45 Showing how we can use quantum to do cool things that aren't possible with existing nova will be key to helping people understand and get excited about quantum. 22:46:02 danwent: that's very useful 22:46:17 for example, it would be really cool to do a demo that shows the cisco plugin, with extensions, doing something not exposed in nova 22:47:00 that code is a bit outdated now... i will update it to use real melange, but you should get the idea. 22:47:20 highlighting different use cases in this way will definitely help. 22:47:24 anything else? 22:47:57 Just one little thing: who's doing the milestone release on thursday EOD? 22:48:07 I was planning on doing it 22:48:12 but its not a big deal 22:48:44 danwent: great. I just wanted to make sure that it was EOD PST! That would be Friday Morning for me :) 22:48:48 the diable-rbp release is coming up pretty soon after it, so I'm not too worried about what does / does not make it into D-4 22:48:54 salv: haha 22:49:21 if there's anything that is looking really tight, email me. we're small enough that we can fudge the times by a bit if there's a good reason 22:49:35 ok, keep up the great work on D-4 reviews folks, and congrats again on incubation :) 22:49:47 #endmeeting