21:02:01 #startmeeting 21:02:02 Meeting started Tue Oct 11 21:02:01 2011 UTC. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:03 Useful Commands: #action #agreed #help #info #idea #link #topic. 21:02:11 Welcome everyone... Today's agenda is at: 21:02:17 #link http://wiki.openstack.org/Meetings/TeamMeeting 21:02:29 Please all make liberal use of # for a richer meeting summary: 21:02:41 link (for links), info, idea (for notes), action (for TODOs), and help (for calls for help) 21:02:55 #agreed everyone should use # for a richer meeting summary 21:03:02 #info link (for links), info, idea (for notes), action (for TODOs), and help (for calls for help) 21:03:03 :) 21:03:14 mtaylor: actually, only the chair can use agreed :P 21:03:19 BAH 21:03:24 #topic Design Summit feedback 21:03:25 haha :) 21:03:36 So, last week we had the Design Summit 21:03:49 I hope mostly everyone made it back home now 21:04:00 home but still hung over. 21:04:03 We didn't have a closing feedback session, mostly due to the start of the Conference 21:04:14 While memories are still fresh, do you have any comment ? 21:04:16 ttx: I feel that sessions more on "bugs" / issues would be useful - in addition to forward looking blueprints (although that might be an issue of misuse of blueprints?) 21:04:17 * mtaylor thought that the Design Summit was rather productive this time 21:04:30 anotherjesse: actually, I think that's the RIGHT use of blueprints 21:04:35 ttx: the ec2 api session was an example of one that was very tactical 21:05:00 anotherjesse: some people propsoed that we have a pre-event that is more backward looking 21:05:03 we should create a bluerprint for bugs 21:05:07 but still technical 21:05:09 anotherjesse: really? It was basically a discussion around why we support EC2 at all. I don't think it hit it's original purpose of the feature gap 21:05:43 anotherjesse: something where we would present technical stuff so that most attendees are up to snuff when the design summit starts 21:05:45 I think a session on how to handle a room full of 80 people, where 20 of those people want to talk, may be useful. 21:05:56 hehe 21:06:12 Do you think we should run a SurveyMonkey thing to gather more precise feedback ? 21:06:15 ttx: yes, that was another comment - that while core developers know roughly how everything works, outsiders don't have context to join 21:06:35 I think we're getting to the point where more sessions, in more parallel that are smaller in scope would be helpful (getting closer to uds style) 21:06:48 ttx: survey monkey would be good 21:07:02 and if we're at the point where there are 8 folks who need to be in all of the sessions, perhaps that's a good incentive for more delegation 21:07:06 #idea - survey monkey...good one 21:07:10 ++ 21:07:18 bcwaldon: I didn't see it that way - I thought it was about "we use ec2 api, why does it have bugs" - beside the point though 21:07:46 #action ttx to set up some survey monkey for summit feedback 21:08:04 #idea more sessions, in more parallel that are smaller in scope would be helpful 21:08:20 #idea pre-event that is more backward looking 21:08:36 +1 on more smaller sessions 21:08:57 #idea some context setting session at the begining for newcomers 21:09:00 vishy: since you were the critical path, do you think you can handle more parallelization ? 21:09:40 we need to work on human cloning? 21:09:48 yes 21:09:57 i think breaking out the working groups will help 21:09:59 hahahaha like that one 21:10:10 like block storage, etc. 21:10:10 #action ttx to ask powers that be about next design summit location plans 21:10:24 #idea have a week long design summit without conference co-location 21:10:28 I like that y'all created an index of etherpad pages, but it would have been nice to have had that in advance so that people unable to attend could follow along. 21:10:35 mtaylor +1 21:10:39 glenc: +1 21:10:54 where is said index? 21:11:03 glenc: actually, the wifi was so bad etherpad was a bit difficult 21:11:06 #idea internet is important for a tech conference 21:11:08 #are these sessions recorded? 21:11:23 also, uds carries an IRC room for each topic on the agenda and has that irc room on one of the projectors in the room so that people can attend rooms remotely as well 21:11:30 anotherjesse: ++ 21:11:47 anotherjesse: hopefully next time we can get rid of the conference pairing, which should allow to go to some lab or university, which have generally good coverage 21:11:52 worst conference wifi ever 21:12:07 glenc: we had serious connectivity issues for the conference 21:12:07 devcamcar: I had worse. 21:12:09 not to keep talking about UDS ... but the uds guys apparently ensure that all locations they use have at least a 45 meg connection that they can be handed 21:12:16 and then they run their own wifi on top of that connection 21:12:29 mtaylor: they arrive 10 days before the event to set it up 21:12:35 mtaylor: you propose to do that ? 21:12:38 it might not be a terrible idea to just use hotels that we know uds has been at before, so we can steal their research :) 21:12:39 mtaylor: if you like how they do it so much, why don't you go there ... 21:12:41 #idea copy uds wifi setup 21:12:41 as well ;) 21:13:01 ttx: sure. being in a place for a week without proper internet is kind of crappy 21:13:13 definitely 21:13:13 anotherjesse: I will! 21:13:17 UDS folks tell the hosting venue that each participant will consume 6 wifi connections and to plan accordingly. 21:13:33 ok, that's good feedback 21:13:41 (and then we bring our own wifi gear as well.) 21:13:42 I guess I shouldn't have torrented the whole time 21:13:45 will do surveymonkey for more 21:13:46 medberry: ++ 21:13:58 medberry: you wanna set up our venue for us? :) 21:14:15 ttx: I know we can't actually co-locate with uds because of timing 21:14:22 mtaylor, my mifi isn't going to handle that load either. :^) 21:14:30 ready for next topic ? 21:14:31 * medberry was tethered all week though 21:14:46 ttx: but what if we went to the same venue they were about to go to just a few weeks earlier ... and then figure out if we can share the setup costs for the infrastructure with canonical 21:15:04 that way the venue would only have to outfitted once 21:15:16 mtaylor: they basically set up their own, separate from the hotel. 21:15:17 it would be like, I don't know - collaboration! 21:15:21 ttx: I know 21:15:29 ttx: we could rent it from them or something 21:15:37 mtaylor: don't want to be the one that asks elmo to come 3 weeks early. 21:15:45 ttx: I'll do it :) 21:15:45 ok, next topic 21:15:47 ok 21:15:47 #topic EssexReleaseSchedule final discussion 21:15:57 Based on the session we had at the Design Summit, I updated the page at: 21:16:02 #link http://wiki.openstack.org/EssexReleaseSchedule 21:16:16 The main change from Diablo is that we have a "RC at will" period starting after the last milestone... 21:16:32 ...where each project produces release candidates until one is deemed good enough for release (and open the next cycle) 21:17:00 Compared to the summit chart I adjusted things a bit (only 4 milestones) 21:17:13 Please doublecheck and let me know if you're OK with it 21:17:29 bcwaldon: I know jaypipes is fine with it, but your input is welcome 21:17:41 ttx: thanks :) 21:17:47 It's 4 balanced milestones, E4 on March 1st, with 5 weeks at the end for stabilization work and production of Release Candidates 21:18:26 that's good...to see stablization work is included now 21:18:29 ttx: lgtm 21:18:37 I like it 21:18:42 +1 21:18:48 ttx: yep, works for me. I know glance is planning to release our RC 1 week ahead of time 21:18:49 We expect keystone and glance to use the very beginning of that last period, and nova / dashboard to use the late part 21:18:59 ttx: not sure if we want to document that somewhere 21:19:14 devcamcar, zns, vishy: ack ? 21:19:21 vishy said ack 21:19:24 ttx: maybe just the milestones listed in launchpad are good enough 21:19:54 bcwaldon: one week ahead of nova, actually. 21:20:09 which still falls after E4 in the "RC at will" period 21:20:10 ttx: yes, that's what we're planning on as well 21:20:10 aye 21:20:11 ack 21:20:37 ttx, RBP isn't listed on the schedule? I can't remember whether there was a conclusion about branching vs freezing trunk 21:20:37 ok, will make it "official" then 21:21:00 markmc: RBP will actually happen when the RC is deemed "good enough" 21:21:09 that's when F will open 21:21:14 ttx, k, thanks 21:21:25 since the PTLs want people to focus on RCs until one is "good enough" 21:21:42 OK, let's go to project status, in roughly the dependency order... 21:22:08 is zns around ? or anyone else from keystone ? 21:22:20 ttx: I will agree for ziad 21:22:38 anotherjesse: but can you replace him for Keystone status ? 21:23:07 Let's start with Glance and give him a chance to join 21:23:14 #topic Glance status 21:23:26 bcwaldon: So is a plan coming together for Essex in Glance ? 21:23:53 or is it for Glance in Essex ? 21:24:08 ttx: definitely 21:24:11 (No Swift this week since nobody replaces John, who is in vacation) 21:24:14 * Vek glances at essex 21:24:18 Be sure to check out https://bugs.launchpad.net/glance/+milestone/essex-1 for everything we have targeted at the first Essex milestone. Currently we're looking at one blueprint and 23 bugs. 21:24:26 Jay also asked me to mention that he sent a summit summary email to the list highlighting what we are going to focus on in Glance during the Essex release. It outlines some major changes planned for the HTTP APIs and how we are going to make Glance more robust. Any questions I cannot answer today regarding anything in that email I will make sure to relay to him. 21:25:27 Looks good. Did you start filing blueprints for the remaining stuff mentioned in that email ? 21:25:53 ttx: I haven't but I am assuming Jay will 21:26:00 ok. Questions on Glance ? 21:26:01 ttx: I think some already exist that aren't yet targeted 21:26:23 bcwaldon: ideally you set the "Series goal" to Essex 21:26:37 and then they all appear in https://blueprints.launchpad.net/glance/essex 21:26:42 ttx: ok, I'll leave that up to the *offical* ptl 21:26:48 ready to be milestoned to a specific one 21:27:07 No other Glance question ? 21:27:22 #topic Nova status 21:27:28 vishy: o/ 21:28:11 vishy: How is the Essex plan coming up for Nova ? 21:28:38 vishy: you'd mentioned pulling Core together to try to figure out priorities? 21:29:04 +1 21:29:25 hey 21:29:26 sorry 21:29:30 bunch of people just showed up in the office 21:29:49 so I haven't made a huge amount of progress since friday since I took yesterday off 21:30:58 i'm trying to collect all of the plans here: http::etherpad.openstack.org/essex 21:31:06 and convert relevent ones into blueprints 21:31:25 vishy: fwiw I filed the rootwrapper and targeted to Essex already. 21:31:26 so if anyone has items to add, I could use some help 21:31:41 feel free to unset the series goal if you don't want it 21:31:50 @Vishy - is there a plan to open up a branch for fixes to D4? 21:31:53 ttx: yes I was just looking over the existing blueprints filed and trying to map them 21:32:10 patelna: no, there is a separate group working on that 21:32:26 Daviey: any news on the sru group setup ? 21:32:27 pointers to that team??? 21:32:31 I added my two etherpads to the wiki , and was planning to supply blueprints to go with them (VM State Managment and Nova Image Cache Management) 21:32:33 i'm not sure where it is going to live yet 21:32:50 patelna: they should send an email to ask people to join the group. Probably this week 21:33:04 Got it...thanks 21:33:12 patelna: group does not exist yet afaik 21:33:37 patelna: ping Daviey on IRC during European office hours 21:33:40 @ttx who is PTL for D4 maintenance group? 21:33:47 k... 21:34:00 I'd like to organize a meeting for nova-core to prioritize and target blueprints 21:34:04 meng-hp: no PTL, a workgroup will equally handle it 21:34:16 pvo: any preference for timing for that? 21:34:40 not tuesdays. : ) 21:34:43 other than that, no 21:35:08 soonish, obviously 21:35:18 Thursday maybe? 21:35:36 yes, would be good. 21:35:38 definitely maybe 21:35:45 <-- not nova-core :) 21:35:48 afternoon is clear 21:35:52 I will try for thursday at this time. I will send out a message to the ml today 21:35:58 great 21:36:07 vishy: anything else ? 21:36:11 <-- not nova-core :) 21:36:13 * Vek suggests #action'ing that 21:36:21 #action vishy to invite nova-core for a planning discussion 21:36:29 Vek: on my way 21:36:37 ttx, I'll have a first cut proposal on the stable branch tomorrow 21:36:45 meanwhile if anyone has anything to add regarding blueprints add them to the blueprints page above 21:37:03 markmc: cool ! I expect Dave to be a bit busy this week. 21:37:13 Questions on Nova ? 21:37:54 #topic Dashboard status 21:38:00 devcamcar: o/ 21:38:05 Did you change the name, btw ? 21:38:23 or rather, did you select a cool code name ? 21:38:27 ttx: we didn't get to do the lightning talk for the naming ceremony 21:38:33 we'll be taking it to the list for a final decision this week 21:38:57 ack 21:39:04 we now have a diablo friendly branch available on git 21:39:04 #info We are after the Design Summit, so now the first step is to get your feature plan filed as Launchpad blueprints, and set the "Series goal" to Essex 21:39:10 #info collecting plans/blueprints for nova on http://etherpad.openstack.org/essex 21:39:10 #action vishy to schedule nova blueprint prioritizing meeting for Thursday and send an email to the list 21:39:38 i'll be adding blueprints from the essex roadmap discussion this week as well 21:39:38 devcamcar: In the end that should give you a nice subset of blueprints at: 21:39:45 #link https://blueprints.launchpad.net/openstack-dashboard/essex 21:39:55 ttx: yep 21:40:02 (All members of ~dashboard-drivers are able to confirm the series goal) 21:40:15 devcamcar: I'll also need you to set the "Maintainer" at https://launchpad.net/openstack-dashboard to ~openstack-admins 21:40:25 should avoid having to add extra people to ~dashboard-drivers for admin reasons 21:40:35 will do 21:40:45 #action devcamcar to set "Maintainer" at https://launchpad.net/openstack-dashboard to ~openstack-admins 21:41:26 that's all I had. Is it reasonable to try to get an Essex plan to present by next week meeting ? 21:41:36 yes i'll have it 21:41:51 devcamcar: anything else on your mind ? 21:41:57 all: Questions on Dashboard ? 21:42:02 not today! mostly just admin stuff 21:42:38 zns: around now ? 21:43:01 #topic Incubated project news 21:43:03 ttx: I can give a broad summary of the summit stuff if not 21:43:20 anotherjesse: ok, will do after that topic. Thanks 21:43:23 danwent: news ? 21:43:25 hi 21:43:40 slides, etherpads, notes from netstack sessions at the summit are at: http://wiki.openstack.org/NetstackEssexSummit 21:43:48 danwent: when do you push other projects to join you in incubation ? You're a bit lonely :P 21:44:15 #link http://wiki.openstack.org/NetstackEssexSummit 21:44:19 #info based on discussions at summit, donabe will no longer be handled within netstack. debo sent email to list about scheduling separate donabe meetings 21:44:32 ttx: i'll just have to talk with myself :) 21:44:38 for some reason the network guys had a working network and could use etherpad. 21:44:51 other than that, we had an unconference session to identify priorities, will be writing up blueprints this week. 21:44:51 haha :) 21:45:01 that's lucky 21:45:02 there's no nestack meeting today…. just write blueprints 21:45:11 * medberry guesses network guys had a 3G laptop 21:45:12 haha, we bring our networks with us :) 21:45:20 or 4g 21:45:21 that is all 21:45:34 any questions? 21:45:49 danwent: nice summary page 21:45:55 ttx: thanks 21:46:14 danwent: would be great if all track leads had such a page, but that sounds a bit unfair to vishy. 21:46:37 yes, i had it pretty easy 21:46:37 ok, back to keystone 21:46:43 #topic Keystone status 21:46:53 anotherjesse: quick design summit summary ? 21:47:35 ttx: we had several sessions on the auth protocol - the major open question is whether we do: 21:47:43 scoping by more than just a tenant 21:48:05 more capabilities into the identity service or not 21:48:45 we had good discussions from folks at hp and other locations - we will be talking about scope of keystone in essex on the mailing list 21:49:07 but try to nail down the api very early in the cycle 21:49:27 that's good Jesse... 21:49:28 anotherjesse: so the essex plans are not settled yet, right, still needs a bit more discussion ? 21:49:38 ttx: that is correct 21:49:50 anotherjesse: yes, I'd like to freeze the APi by E3 21:49:57 to avoid any bad surprise 21:50:15 even if that means having a keystone RC early and open F for keystone earlyt 21:50:32 compared to everyone else 21:51:02 anotherjesse: did you have anything else ? 21:51:17 ttx: good for now - if there are any questions feel free to ask here or in dev 21:51:55 ok then 21:51:58 #topic Open discussion 21:52:05 question! 21:52:09 I mean comment! 21:52:11 I wanted to bring to everyone's attention the email I sent to the list proposing the removal of OSAPI v1.0 from Nova. I think it's a really important decision we need to make and I want as many people involved as possible. In favor of not wasting anybody's time here, if anybody does have any comments, please share them on the mailing list. 21:52:44 I might as well ask what would be a great starting place for a new contributor, myself. 21:53:11 bcwaldon: do you think you can reach consensus, or does it look like the sort of discussion we should have had last week ? 21:53:28 * ttx admits not having read the whole thread 21:53:33 ziyadb: I think getting everything running and working on low hanging fruit is good for getting started … we hope to be finished with a script that helps with that (devstack) and email the list tomorrow 21:53:52 ttx: I'm referring to a second email. The conventions email is a separate discussion 21:53:58 ziyadb: and there is a "low hanging fruit" tag in the bug list 21:54:00 ttx: we can definitely reach consensus around removal ov v1.0 21:54:32 anotherjesse, is devstack.org broken atm? 21:54:48 medberry: it's up for me 21:54:49 medberry: ask in #dev 21:54:51 it is github 21:55:04 anotherjesse: excellent, I'll start there. Although I must comment, a "getting started" page for new devs, aside from that which includes information regarding joining projects and signing the CLA, would be great. 21:55:33 ttx: is that something you or stef can tackle? 21:55:35 ziyadb: ahh, that is on the wiki I think - one sec 21:56:06 anotherjesse: #dev? did you mean a diff channel? 21:56:08 ziyadb: the top right corner of http://wiki.openstack.org/ talks about contributing, ... 21:56:11 and I'm not sure how the dev ecosystem is, would I for instance be able to work with other individual contributors to contribute a particular feature in an ad-hoc fashion? 21:56:14 westmaas: sorry, meant #openstack-dev 21:56:17 anotherjesse: thanks, I'll have a look. 21:57:00 bcwaldon: not sure how we can help, the discussion seems to stay on track. Do you mean it needs outsiders to sum it up and help it making progress ? 21:57:20 #action someone needs to update the "How to Contribute" page; it's a bit out of date 21:57:21 ttx: sorry, I was referring to ziyadb's request for a jumping off point on the wiki 21:57:26 ttx: ignore me ;) 21:57:44 glenc: I'll take it 21:58:02 #action ttx to update the "How to Contribute" page 21:58:37 ok then, I think I filled my bag of TODOs for the week. 21:58:49 ah, I suppose I'll start with the low hanging fruit and work my way up to triaged bugs. 21:58:51 thanks for the information. 21:59:09 ziyadb: I'll try to update the page -- if you have any doubt, ping me 21:59:10 #action add dashboard & keystone to wiki.openstack.org overview 21:59:26 the wiki definitely needs some love 21:59:27 ttx: as I will, thanks. 21:59:31 lots of change in essex - probably a lot of things need updated :) 21:59:35 err - diablo 21:59:51 ok, let's close it and leave the room for the netstack folks 21:59:53 well, same will be true of essex, I'm sure. 22:00:01 #endmeeting