22:02:31 #startmeeting 22:02:32 Meeting started Tue Feb 21 22:02:31 2012 UTC. The chair is danwent. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:02:33 Useful Commands: #action #agreed #help #info #idea #link #topic. 22:02:48 hello netstackers 22:02:55 Hearty Congratulations! 22:03:02 did the delayed meeting start let salvatore get back in time for the meeting? :P 22:03:06 Hello 22:03:10 o/ 22:03:13 hi 22:03:15 o/ 22:03:21 ahoj! 22:03:24 yup, Quantum in going to be a core project for folsom! 22:03:24 Hola!! 22:03:32 :-) 22:03:36 Great job Dan!! 22:03:38 =) 22:03:40 hello and congrats team Quantum! 22:03:54 congrats all! 22:04:01 great news, indeed 22:04:04 its an exciting step, but it also means a lot of additional work for all of us to live up to that challenge :) 22:04:32 yes, we can! 22:04:38 I'll be sending out email with the some of the feedback from the PPB meeting (or you can probably look the minutes up yourself, as well) 22:04:52 edgarmagana: was that an obama reference? 22:04:59 :) 22:05:14 #link PPB minutes: http://eavesdrop.openstack.org/meetings/openstack-meeting/2012/openstack-meeting.2012-02-21-19.59.html 22:05:16 or bob the builder 22:05:27 thanks mark 22:05:34 Ok, so on to the meting 22:05:36 meeting 22:05:45 http://wiki.openstack.org/Network/Meetings 22:05:54 Some high-level feedback from the PPB: 22:06:13 1) need to get integrated into main CI infrastructure for system-test/integration asap. 22:06:36 we've been working on this for sometime, and have a lot of the peices, but need to push this work to conclusion. 22:06:56 2) work to get to full parity with existing nova capabilities. 22:07:33 3) API design work around integrating quantum + melange (to be discussed at summit, but we should really get started on this earlier). 22:08:12 4) documentation around how to transition from a non-quantum setup to a quantum-based setup. 22:08:38 all in all, this is a lot of work, but its important to making sure people can fully leverage quantum. 22:09:06 Will discuss this more via email, but wanted to give you a flavor 22:09:15 are there any questions/comments about the transition to core? 22:09:29 So melange is a part of the core as well (as a part of Quantum)? 22:10:11 What will the move to core do for the reviewers? Will all of the core devs be able to help? 22:10:12 mandeep: plan is that melange won't really exist anymore, though its IPAM capabilities will be folded into a new network API in the quantum project. 22:10:48 wwkeyboard: I think we'll really have to grow our team of core reviewers. 22:11:12 wwkeyboard: this is something I already chatted with the PPB about, as I think we'll need more support from the general openstack community. 22:12:18 wwkeyboard: the questions of whether plugins will be included in the core is a separate issue, but certainly will have an impact on core dev review load. 22:12:40 and CI 22:13:12 Running their unit tests is one thing, but we won't be able to do functional testing on most of the plugins. 22:13:13 wwkeyboard: from a CI perspective, it is very unlikely that we will do CI for all plugins. 22:13:47 though I think one goal of the CI team would be to let other run equivalent CI infrastructure themselves. 22:14:17 so if you have a plugin, and a lab with the necessary equiptment, you could run CI testing for that plugin. 22:14:24 Re: CI for plugins: Would that not depend on the plugin providing a test that can be run in the CI environemnt? 22:15:30 mandeep: I think CI is more about doing basic functional testing that works identically regardless of plugin. 22:15:43 Some plugins require specialized hardware 22:15:53 OK. 22:16:12 Salv-orlando: exactly, which is why were were saying that it would not be possible for openstack CI team to do CI across all plugins. 22:16:35 Agreed. 22:16:39 but the plugin creators themselves could choose to run the CI test infrastructure themselves with a lab that includes the specialized HW. 22:16:54 ok, any other questions? 22:17:21 Ok, next topic is Folsom design summit 22:17:29 invites 22:17:41 salv-orlando: I understand. My point was in the plug-in writers write a plugin mock that can run in CI environment, we could run that (and identify API breakage etc early). 22:18:45 invites have gone out to an auto-generated list of people identified as existing openstack contributors 22:18:59 based on git/gerrit history, I believe. 22:19:15 this has caused quite a stir, as some people didn't get invites. 22:19:50 PTLs have a small number of additional invites for people who have not yet contributed, but plan on contributing in Folsom 22:20:14 so email me if you have not gotten an invite, but think you need one. We're doing our best to work though these issues. 22:20:20 remember, this is for the developer summit. 22:20:29 openstack conference is open to a much larger number of people. 22:20:48 questions/concerns on developer summit invites? 22:21:18 Ok, Essex-4 22:21:35 Today is our target for having all features reviewed and in. 22:22:02 Looks like most things are in review: https://launchpad.net/quantum/+milestone/essex-4 22:22:25 though I had to do some clean-up earlier, as there were a lot of stale issues around. 22:22:46 So we have a lot of review work to do today. 22:23:10 Probably OK if it bleeds into tomorrow, but need to wrap it up very soon. 22:23:27 Please remember to keep an eye on the python-quantumclient repo as well: https://review.openstack.org/#q,status:open+project:openstack/python-quantumclient,n,z 22:23:57 Particularly, Salvatore has a branch on the 1.1 client changes which is very important: https://review.openstack.org/#change,4301 22:24:17 I have one review on there, but we'll need a second. 22:24:43 salv, any comments there, or is it good to go? 22:24:51 Yeah I'll be up for a few more hours tonight. Been at the stadium, too much adrebakine, can't sleep 22:25:12 I haven't yet loojed 22:25:23 Salv: I don't know what that is, but I'll assume its an illegal substance :P 22:25:23 Looked at the reviews 22:25:48 based on my last review, I think we're in good shape on that one. 22:25:55 Mobile phone swipe keyboard is set to italian ;-) 22:25:56 danwent: I think he meant adrenaline? :) 22:26:11 dan and salv: let me take a look today 22:26:11 ah… dissapointed :) 22:26:18 thanks edgar! 22:26:26 Thanks 22:26:30 f you have already +1 I can approve it 22:26:55 edgar: I still need to do another pass, as salv uploaded some changes since I last reviewed. 22:27:11 but I don't expect problems as my previous comments where fairly minor. 22:27:14 https://review.openstack.org/#change,3618 22:27:25 ryu plugin. salv and I have been doing this review 22:27:48 salv has a +1. I need to confirm that a few of my points where addressed, then I think its good to go. 22:28:10 nvp plugin review: https://review.openstack.org/#change,4358 22:28:41 would be good to get a non-nicira review on this. We've all reviewed it internally, and could reveiew publicly, if needed. 22:29:26 I can do a review in the morning (europe morning) 22:29:40 danwent: I will be reviewing it as well 22:29:47 great! 22:30:21 And there's a change to add a tunneling agent to OVS: https://review.openstack.org/#change,4367 22:30:47 ~250 lines. 22:31:02 I still need to review this, and bhall is already reviewing. 22:31:08 hmm, interesting, thanks for noting that 22:31:18 but if anyone else wants to review, please add yourself. 22:31:39 is bob here? https://review.openstack.org/#change,4286 22:32:01 danwent: bob is out this week 22:32:09 sumit and I have both +1'd his change, but I'm waiting on feedback regarding plugin binaries in the bin directory 22:32:12 cdub: ah... 22:32:49 Ok, so the question is whether plugin agent's should be in the main quantum/bin directory 22:33:08 my sense was that they should not be, but I wasn't sure if this was somehow needed for packaging. 22:33:19 danwent +1 22:33:52 cdub: is Bob completely offline? I'm wondering whether we can tweak this patch with him being out. 22:34:07 you mean in the tree? I think they should live with the plugin (maybe a bin directory in the plugin or something). If we split out the plugins to a repo at some point we'll have to move them out of the bin dir anyways 22:34:12 from a packaging view, plugins agent should be at plugins dir, this way it's easy to differenciate which binary belongs to which klg (althoug in this case is easier) 22:34:13 b/c this contains several things which are quite important. 22:34:21 danwent: pretty much offline, yeah. i can bug him, but i already did once...hate to interrupt his time off 22:34:23 bhall: +1 22:34:34 all else fails we can push his patch and generate a patch to it 22:34:48 anyway, at packaging time we will move everything around 22:34:49 bhall: you thinking something like quantum/plugins//bin ? 22:34:57 yeah, something like that 22:35:02 +1 22:35:06 +1 22:35:25 that would make more sense to me, but had wanted to hear from bob as to whether he had a reason for his approach. 22:35:26 +1 22:35:42 yeah, the package process places things whereever, just makes sense to have the basic pre-package layout be sane to being with 22:36:00 begin with... 22:36:19 Ok, cdub, should we abandon his patchset and have someone else tweak it, or do you think he would want to tweak it himself soon (given our time constraints). 22:36:59 danwent: can i let you know later today? 22:37:17 cdub: sure. 22:37:25 ok, cool 22:37:45 cdub: we just definitely want this patch in soon, as it has several things important to making sure the sdist tarball is not busted, which we need to be able to test the E-4 release. 22:38:00 *nod* 22:38:07 i'll chase down right after this 22:38:11 thx 22:38:24 Ok, there are a few other small commits that we should also review today/tomorrow. 22:38:45 but unless someone wants to call them out explicitly, I'll skip them here so we can get back to work :) 22:39:25 Ok, last thing on the agenda was just to comment on what people should expect in terms of the master branch in the next few weeks. 22:39:49 next week, only thing going into quantum master should be bug fixes or packaging work. 22:40:17 focus should be on testing 22:40:36 And docs :-) 22:40:54 Salv_: indeed… I always welcome help :) 22:41:34 Salv_: and yes, you'll be working on 1.1 API docs, I suppose, right? 22:42:25 Correct 22:42:39 After E-4 is cut, we'll go into a release candidate phase. 22:43:33 we will aim to cut our quantum essex release well in advance of the actual release date, and then re-open master for commits. 22:44:03 (tehnically, master will be open before then, but team resources should primarily be focused on getting the release candidate polished, and docs in shape) 22:44:12 I co 22:44:16 Any questions/comments on that? 22:44:32 Sorry did not want to enter texy 22:44:38 Text ... 22:44:42 shoot, looks like i skipped to agenda items. are debo-os or mjfork around? 22:45:04 yes\ 22:45:06 don't see debo 22:45:18 mjfork, anything to add on the horizon front? 22:45:29 shuold be commiting in about 5 mins 22:45:45 great! 22:45:57 can you send a link out to the netstack list so people can help review? 22:46:12 Gabriel Hurley gave me a few pointers, finally have tests passing and was able to use the network name 22:46:26 mjfork, cool. when is the horizon freeze? 22:46:45 some rough edges to flush out in creviews (e.g. hwo to detect if Quantum is present) 22:46:47 i don't know. 22:47:13 danwent: where is no blueprint for this right? 22:47:25 ? 22:47:34 there is no blueprint? 22:47:44 none that I am aware of in horizon. 22:47:44 whoops, meant there is no 22:48:36 might just create a bug there. That's why I was asking about when they were freezing. 22:48:51 but we can figure this out offline. Let us know if we can help review. 22:49:31 mjfork, anything else? 22:49:46 yes, i will catch you offline. specific question around default networks. 22:50:12 Only other things we need to figure out is the status of debo's devstack stuff. Dave, I believe you said you have some feedback around using the script with devstack? 22:50:21 let's just take that offline 22:50:31 danwent: yes... 22:50:42 i've been working on multi-node devstack + quantum 22:50:52 will send out scrpits to the ML... 22:51:21 ok, great. getting integrated into the CI infrastructure is really important, as its part of what the PPB really wanted to see. 22:51:38 Ok, anything else to discuss? Other than that, we just really need review cycles today. 22:51:46 (and likely tomorrow…) 22:53:10 Ok, thanks folks, and congrats again to the team. This is an exciting step toward Quantum being used widely throughout the OpenStack community. 22:53:25 #endmeeting