21:01:27 #startmeeting nova 21:01:28 Meeting started Thu May 7 21:01:27 2015 UTC and is due to finish in 60 minutes. The chair is mikal. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:01:29 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:01:31 The meeting name has been set to 'nova' 21:01:33 Hi! 21:01:37 o/ 21:01:39 feels like deja vu 21:01:41 o/ 21:01:44 \o 21:01:45 o/ 21:01:49 o~ 21:01:49 o. 21:01:52 o/ 21:01:54 o/ 21:01:55 hi 21:01:57 well, ~o 21:02:04 o..../ 21:02:09 o/ 21:02:15 #topic Liberty release status 21:02:17 jaypipes: has long arms 21:02:31 #link https://blueprints.launchpad.net/nova/liberty 21:02:58 Do there are quite a few blueprints in "needs code review" targetted to liberty-1, so that's cool 21:03:08 Spec review is still ongoing as well 21:03:19 I'm not sure what else to say apart from that 21:03:24 We have a summit soon! 21:03:28 Anything else peoples? 21:03:39 non 21:03:44 I have a topic related to API 21:03:54 jaypipes: one sec 21:03:56 I just noticed 21:03:58 https://blueprints.launchpad.net/nova/+spec/devref-refresh-liberty 21:04:07 Which is a candidate for a non-spec fast pass 21:04:30 "The source tree of liberty must be refreshed from time to time with the blood of developers." 21:04:44 mikal: ++ 21:04:48 That's been filed by johnthetubaguy, but its not clear to me who is signed up to do that work? 21:04:56 he's doing it 21:04:58 mikal: that's johnthetubaguy 21:04:58 mikal: everyone is. :) 21:05:00 he has a bunch of patches up 21:05:03 Oh, ok 21:05:15 So, do we let that guy have an approved spec? 21:05:21 It seems like it meets the requirements to me 21:05:34 doesn't need a spec for it, right? 21:05:38 not worth a spec for sure, you mean "non-spec approved blueprint" right? 21:05:42 Sorry, BP I mean 21:05:43 I'm cool with that. 21:05:46 yes, +2 21:05:55 Ok, cool 21:06:01 jaypipes: ok, what was your API thing? 21:06:09 * alaski slinks in late 21:07:19 I think we should discuss the OverQuota 403 vs. 400 thing. is sdague around? 21:07:31 can that be open discussion? 21:07:48 Fair question, is it a release thing? 21:07:52 hey, sorry, waiting for my IRC client to catch up, but I am around 21:07:55 no, sorry, I will wait. 21:07:59 jaypipes: thanks 21:08:10 johnthetubaguy_w: we forgive you, you can tell me when I miss stuff 21:08:42 mikal: :) 21:09:20 cool, I think you approved my spec then? 21:09:20 Anything else for release? 21:09:25 johnthetubaguy_w: yes 21:09:28 cools 21:09:41 #topic Liberty design summit 21:09:51 Ok, so there is now a list of sessions for the summit online 21:09:59 #link http://libertydesignsummit.sched.org/type/design+summit/Nova#.VUvT8dqqqko 21:10:03 yeah, I have added some folks as leaders 21:10:18 Do you still need session leaders for anything? Or is that all done now? 21:10:20 I need to get in touch with you all though, and I haven't done that yet 21:10:28 so its not confirmed 21:10:45 I am trying to get the etherpads with the rough agenda, so its clear which session is witch 21:10:52 there are a few overlap kind of things 21:11:04 Are we going to try and set pre-reading again this time? 21:11:08 almost there, but I got dragged off into some fires this afteroon 21:11:10 mikal: +1 21:11:23 I have added some place holder stuff for some 21:11:30 like cells has a link to the cells devref 21:11:47 I am writing up some devref changes for my ones, where it makes snese 21:11:59 I'll have some specs for the cells stuff 21:12:01 I was going to send out that request when I contact each proposed leader or leader group 21:12:10 alaski: perfect 21:12:13 sounds fair 21:12:21 #link abhishekk mikal tjones cburgess jgrimm adrian_otto funzo mjturek jcookekhugen irina_pov krtaylor danpb alexpilotti flip214 jaypipes gilliard garyk edleafe dims moshele anteaya Nisha sileht claudiub lxsli neiljerram markus_z swamireddy alevine tonyb andreykurilin ndipanov sc68cal akuriata 21:12:23 oops 21:12:26 johnthetubaguy: lol 21:12:30 #link https://wiki.openstack.org/wiki/Summit/Liberty/Etherpads#Nova 21:12:42 so I added etherpads in there 21:12:48 johnthetubaguy: so, I noticed there is a free session slot. are you considering any other session topics at this time? 21:12:59 jaypipes: yes, thats the next bit really 21:13:10 jaypipes: are we generally all happy with the list of slots 21:13:17 #link https://etherpad.openstack.org/p/liberty-nova-summit-ideas 21:13:29 jaypipes: that place holder, has a place holder... 21:13:33 johnthetubaguy: the only one I'm not sure about is the 9am Thurs one. 21:13:49 I think devananda would say he'd like to see discussion of futures for clustered hypervisors 21:13:54 so there is a NFV come resource traker one pencilled in 21:14:06 mikal: lol, that's exactly the topic I was going to propose. 21:14:09 mikal: yeah, thats a good point, its not in there 21:14:16 ugh 21:14:17 OK, so thats like consensus 21:14:24 mikal: that can be discussed re: RT 21:14:24 Sorry for a newbie question, but ... do I need to sign up in advance for the sessions I'm interested in? 21:14:27 dansmith: its not a happy topic 21:14:29 I feel a bit like cells is the future there, but I'd like to talk that through with people smarter than me 21:14:36 johnthetubaguy: no, it's not 21:14:40 neiljerram: no, just show up 21:14:42 bauzas: not really. it will take an entire session, IMO 21:14:45 johnthetubaguy: unless we decide "no future" then it's fine :) 21:14:46 neiljerram: Nope just rock up 21:14:49 i didn't realize we had managed to come up with a controversial NFV thing this time 21:14:58 so… 21:14:58 neiljerram: sit near the front if you think you might want to talk 21:15:07 mikal: Thanks; so the lists of attendees are just for info, then? 21:15:10 jaypipes: agreed that it's wide, but we can kick it off there 21:15:20 neiljerram: yep 21:15:22 mikal: cells could be the future, but that needs some discussion because it hasn't been thought through, that I'm aware of 21:15:23 jaypipes: and maybe leave details up to Friday ? 21:15:25 bauzas: are you talking about clustered hypervisors? 21:15:27 mikal: cool, thx 21:15:30 dansmith: yup 21:15:31 lets pencil that into the old resource tracker slot for now, I will try work out if that works (scheduling wise, like no clash with ironic, etc) 21:15:33 So, we don't have to talk about clustered hypervisors, but its a wound we need to heal sometime 21:15:38 bauzas: there is no question, that's at least an hour discussion 21:15:46 mikal: agreed. 21:15:49 dansmith: then we need a slot 21:15:53 do we think we're going to reach consensus?> 21:15:59 jaypipes: you had a question about a section 21:16:00 or even some actionable items? 21:16:03 dansmith: that's the quesiton 21:16:04 dansmith: if everyone agrees with me and you, yes. 21:16:09 jaypipes: hah 21:16:22 dansmith: we know that it's really wide, but I don't yet see what would be the outcome 21:16:25 dansmith: something about a big fire? 21:16:30 johnthetubaguy: yeah :) 21:16:34 dansmith: we had talked in the past about clustered hypervisors being presented as cells when the cells refactor is done. I would like to know if we still think that's an option at least. 21:16:37 johnthetubaguy: if you let me lead that, we can be done in ten minutes 21:16:38 johnthetubaguy: yeah, I was wondering if the Nova API 2.0, 2.1, 3rd party APIs thing was really necessary along with the Nova: API v2.1 next steps one. 21:16:49 mikal: I don't think it is really anymore 21:16:56 mikal: the way we're headed 21:17:00 jaypipes: so I should share what I am intending there 21:17:07 please do :) 21:17:14 we have moved slots that want operator input to the thrusday 21:17:19 as there are no ops tracks that day 21:17:32 basically I wanted to do v2.0 future with operators in the room 21:17:39 what's the goal of having a session discusing about distributed hyps ? to say it's barely supported and we need to fix stuff ? 21:18:02 johnthetubaguy: I see. OK. 21:18:06 jaypipes: a valid question though, is do we need the previous one without the operators there 21:18:18 (and what would replace it) 21:18:36 johnthetubaguy: that's prolly a question for alex_xu and oomichi. 21:18:44 johnthetubaguy: I would tend to say "no". 21:19:22 so 21:19:30 johnthetubaguy: I have a suggestion for a session if it is replaced... 21:19:34 I thought that the first one was about our next steps for implementation, 21:19:46 and the second was about when we can deprecate v2.0, whether people have been poking it, etc 21:19:46 dansmith: isn't that pretty much settled, though? 21:19:55 jaypipes: which? 21:20:07 dansmith: what next steps are necessary at this point? 21:20:13 jaypipes: we still have work to do on the client, we have some things to work out around when and where we do microversions, etc, right? 21:20:15 maybe 21:20:26 I'm just saying, I thought that was the split 21:20:29 is it really likely ops have been heavily poking given kilo was only just released? 21:20:30 dansmith: I see what you're saying.. 21:20:36 yeah, that was my thinking too 21:20:41 dansmith: yeah, prolly good to keep it and discuss those things. 21:20:43 I mean its 1.5 sessions, probably 21:20:50 and we might be slow by the end of the day 21:21:05 one thing, jaypipes you had a replacement idea? 21:21:06 dansmith: it's a 5:20pm session, which means folks are gonna be antsy to cut it short anyway. 21:21:12 sgordon: maybe not, but we also said we'd circle with CD people to get them to push it out early and test it with some specific customers 21:21:17 jaypipes: yeah 21:21:59 johnthetubaguy: yes, I had an idea about a session for discussing higher-level grouping concepts like availability zones and regions and what AWS calls placement groups, and where to store that information properly in Nova (or outside of Nova) 21:22:18 johnthetubaguy: I just don't want that proposed session to be hijacked by the cascading openstack stuff. 21:22:24 jaypipes: I think we have time in the scheduler session (maybe) for that stuff 21:22:25 lol 21:22:34 so I've kinda kept it secretish :) 21:22:35 jaypipes: that a good point… 21:22:45 jaypipes: does your idea include fractals? 21:22:46 that's funny jaypipes 21:22:53 jaypipes: we should talk about that at some point, as it concerns cells I think 21:22:56 because if not.. snoozefest 21:23:15 alaski: tru nuf 21:23:20 jaypipes: I would suggest you to read https://etherpad.openstack.org/p/YVR-nova-scheduler-in-liberty about AZs 21:23:23 alaski: jaypipes: I will make a note to add that in one of the scheduler sessions, I think its sorto fo there already 21:23:27 ideas: https://blueprints.launchpad.net/nova/+spec/make-availability-zone-real-data-model 21:23:31 scheduler and cells are together, so we can just hit that in those 21:23:40 johnthetubaguy: sure, that works :) thx! 21:23:50 either one. cells or scheduler is fine. 21:24:18 So, we're done with that? 21:24:25 jaypipes: put it in the first one, it could slip if needed 21:24:30 mikal: I am good 21:24:35 thanks for the feedback folks 21:24:42 Ok, the last thing for design summit on the agenda is a stow away 21:24:44 jaypipes: I added an AZ section on the first session, see the etherpad 21:24:52 There is a proposal to run a mid-cycle meeting this release 21:24:59 jaypipes: that's tied to your NP 21:25:02 Do we think there is value in a mid-cycle meetup? 21:25:02 BP even 21:25:06 cheers 21:25:15 The current proposal is Tuesday 21st July till Thursday 23rd July in Rochester, MN 21:25:24 jaypipes: bauzas: can we allow hosts to be part of multiple orthogonal availability zones? 21:25:26 I do 21:25:31 they're always useful, IMHO 21:25:32 I've always found value 21:25:37 cfriesen: let's discuss that off-meeting :) 21:25:41 cfriesen: leave it for the discussion in YVR! ;) 21:25:53 I guess the question could be rephrasesd as "will enough people come if we run one to make it worth while?" 21:25:59 I keep wondering because of what ttx said about the summit, but they are useful 21:26:11 I find midcycle valuable too 21:26:13 mikal: that's a fair question, I would love remote possibilities 21:26:35 we'd have a few cross-project people here too, and could maybe rope mestery in if he's around (lives 90 miles away) 21:26:39 bauzas: I think we can do someting with google hangouts 21:26:41 if we wanted to talk neutron things 21:26:42 mikal: dims and I will be there pretty much regardless of when and where... unless it's in like Antarctica. 21:26:45 bauzas, as in remote-in to a "real life" meeting? 21:26:47 bauzas: like we did in Jan only better 21:26:52 bauzas: yeah, the fancy mike worked ok last time 21:26:52 21-23 Jul is also conflicting with EuroPython IIRC 21:27:04 ++ 21:27:22 mikal: sure, just wanted to say 'please, continue' :) 21:27:33 So, my concern is if we debate the details too much we'll never get around to actually doing the thing 21:27:45 So should we just open up a signup and see how many people say yes? 21:27:49 yes 21:27:52 And cancel if we get say less than ten? 21:27:53 mikal: +1 21:27:53 mikal: +1 21:27:59 ++ 21:28:05 given how many we had last time, is there a concern? :) 21:28:18 dansmith: I don't know how much of that was the CA bump though 21:28:20 johnthetubaguy: you 21:28:20 mriedem: I'd come down especially if you're providing John Hardy's BBQ as a dinner option :P 21:28:21 we had 55 register and 30ish show up 21:28:27 johnthetubaguy: you're ok with me doing that thing? 21:28:37 mikal: please do, thank you! 21:28:52 Ok, I will create and eventbrite thing today sometime 21:28:52 mikal: thats a good Segway to the next topic 21:29:04 Ahhh, nice call 21:29:12 #topic Czar and CPL Refresh 21:29:25 so I am going to keep this really short… 21:29:48 basically, if you want to help out with the running of Nova, taking on a particular aspect 21:29:51 then do reach out 21:30:10 mikal: is working with mriedem on sorting out the midcycle, and thats awesome 21:30:18 johnthetubaguy: sign me up for gate i guess 21:30:24 I think mikal is going to become our Security Czar 21:30:27 mriedem: awesome thank you 21:30:29 johnthetubaguy: I'd be happy to do the security coordination as long as alaski and tonyb aren't offended by it 21:30:40 mikal: I am the opposite of offended 21:30:43 johnthetubaguy: alaski especially has more tenure than me there 21:30:48 alaski: awww, thanks man 21:30:56 sweet 21:31:02 so no need to all step up now 21:31:04 * jaypipes self-nominates as Czar of Awesomeness 21:31:09 but if you fancy something, let me know 21:31:24 jaypipes: sure, probably needs some kind of jingle theme music 21:31:29 jaypipes: I'm running for that too, so they'll have to vote 21:31:29 I really like the idea of a mentoring czar by the way 21:31:29 :) 21:31:36 But I don't think that's a job I'd personally be good at 21:31:37 * artom fancies ice cream. Is that in Nova yet? 21:31:57 anyways drop me an main or IRC ping if you interested in something 21:32:01 johnthetubaguy: is now a good time to discuss the need for a counterpart to sc68cal as a Neutron liaison from Nova? 21:32:05 mentoring czar ? 21:32:08 jaypipes: beat me to it 21:32:11 heh 21:32:12 jaypipes: thank you for the reminder, yes 21:32:14 nice idea, in particular for Nova 21:32:15 mikal: I like that as well 21:32:18 bauzas: its in the wiki list of possible hats 21:32:24 https://wiki.openstack.org/wiki/CrossProjectLiaisons 21:32:32 we have lots of blanks in the above 21:32:35 * sc68cal needs to find weights to stop jaypipes from typing faster than he does 21:32:37 mikal, possible hats? Link? 21:32:42 and we need someone to go to the neutron meating 21:33:00 mikal: I couldn't find that dammed page earlier 21:33:00 https://wiki.openstack.org/wiki/CrossProjectLiaisons 21:33:04 oops 21:33:09 https://wiki.openstack.org/wiki/Nova#People 21:33:10 ah, there is another one somehwere 21:33:24 mriedem: to be clear, I am happy to add more :) 21:33:33 with funny names if that makes you feel good 21:33:35 anyways 21:33:41 anyone up for the neutron thing? 21:33:55 beagles we keep winking at you, is that working at all? 21:33:56 I guess we need to add a table in cross project wiki for cross project 21:34:00 last week beagles was asking his mgmt for permission 21:34:05 sc68cal: that would be awesome, thanks 21:34:07 johnthetubaguy: question... if nobody self-nominates, will you appoint someone? 21:34:13 I would like to be tangentially involved, for cells stuff, but can't commit to it full time 21:34:27 jaypipes: I might start asking people more forcefully, yes 21:34:32 that's all about committing ourselves :) 21:34:41 how very British of you, johnthetubaguy! 21:34:56 jaypipes: if everyone says no, I believe it defaults to me, which is bad news, as I am likely to make a hash of it 21:35:03 jaypipes: :) 21:35:13 so if anyone wants a nice cup of tea… its yours 21:35:20 Last time the British did something forcefully an empire happened... 21:35:38 OK, well action item for sc68cal to send another email out, cc'ing johnthetubaguy to the openstack-dev ML asking for a Nova community member to be the Neutron liaison. 21:35:50 artom: :) 21:35:50 yeah, its worth a nudge 21:36:01 Next topic? 21:36:09 * mikal looks at the clock concernedly 21:36:13 yeah 21:36:14 wasnt there a neutrob liason proposed by mestery? 21:36:16 whoops I am here 21:36:23 claudiub: that was me 21:36:31 #topic Bugs 21:36:45 (interrupted) ... I am up for it btw 21:36:50 so I made a little triage 21:37:01 or recomanded? 21:37:07 si, low hanging fruit bugs 21:37:19 we need those soon, for the upstream training 21:37:22 And I pinged the ironic people cause they have more than their fair share 21:37:30 so if some one fancies that, its cool 21:37:38 so a Bug Czar would be awesome 21:37:39 basically 21:37:41 johnthetubaguy: I tried to find some 21:37:44 to help push this 21:37:47 johnthetubaguy: do you know how many low hanging fruit bugs the training people need as a minimum? 21:37:52 johnthetubaguy: for the low-hanging-fruits 21:37:57 johnthetubaguy: or just as many as we can find? 21:38:02 mikal: as many as you can 21:38:08 Ok 21:38:15 reed: ^ 21:38:18 mikal: that's depending on the students who will pick what htey want 21:38:36 I guess its time to discuss the gate 21:38:38 sc68cal: i see, thanks for the answer. :) 21:38:40 mikal, many :) 21:38:44 Hey, I can write some bugs if you need 'em 21:38:52 johnthetubaguy: mikal: another big bunch we can offload are the ec2 related bugs. right? 21:38:55 yeah, I was going to say, I can make some new ones 21:38:56 johnthetubaguy: well, can we quickly do coresec first? 21:38:58 it should be a habit of leaving the small stuff available for people to get trained :) 21:39:07 mikal: sorry yes 21:39:17 just mentioning that the trivial bug list is back up 21:39:27 So, we're doing a bit of a reboot of the group that does embargoed security vulnerabilities 21:39:28 bauzas: link? 21:39:28 edleafe, I knew I could count on you :) 21:39:40 reed: it's what I do best! :) 21:39:41 Mostly because of some burnout from very hard working people 21:39:50 There are a few bugs in flight I am getting my head around 21:39:52 mriedem: https://etherpad.openstack.org/p/liberty-nova-priorities-tracking 21:40:06 mikal: I left them new 21:40:07 But if you're a subject matter expert in an area we have a bug for, expect a tap on the shoulder 21:40:22 mikal: since we don't yet know if it's a config problem or more than that 21:41:08 Also, if you're interested on helping out on the security bug thing more directly, come talk to me 21:41:13 Especially if you don't work for Rackspace 21:41:20 Cause we're kind of hogging it all at the moment 21:41:22 And we love sharing 21:41:26 +1 21:41:33 mikal: +1 21:41:49 +1 21:41:51 Oh, good point. We should also say who is working on this at the moment... 21:41:53 mikal: i can help with security stuff if you need a hand 21:41:57 Which is alaski, tonyb and I 21:42:10 dims: we were worried you were too big a deal in oslo now for us little people? 21:42:32 says the person who is core on many things :) 21:42:35 * johnthetubaguy hopes distro folks are able to help out with this 21:42:50 dims: we'd love your help, let's talk more after this? 21:42:51 are we including -drivers like we said? 21:42:55 johnthetubaguy: distros help on the reporting end :) 21:42:55 mikal: ack 21:43:02 I was assuming I was already on the hook for this because of that 21:43:06 dansmith: I think that's fallen out of my brain 21:43:20 The VMT likes the group being very small, but I am not opposed 21:43:23 johnthetubaguy: opinion? 21:43:25 well, 21:43:35 that's cool with me of course 21:43:47 can also just reach out to drivers first when you need some generalist feedback or something 21:43:49 mikal: well dims isn't a particularly tall person. 21:43:50 so, I flip flopped on this 21:43:51 they like a small group for initial triage at least 21:43:54 mikal: so you're cool. 21:44:00 dansmith: definitely on the reach out thing 21:44:02 :) 21:44:04 basically we are going to use that group to do initial traige 21:44:06 : 21:44:08 its normally one person 21:44:08 :) 21:44:15 but we are too big 21:44:38 I think we need to do better at roping people in quickly, then making sure stuff gets complete and pre-reviewed 21:44:50 the mode of interaction just sucks 21:44:55 And also at pushing back to stupid bugs 21:44:57 patches in bugs, review comments in comments 21:45:00 sucks real bad 21:45:02 There's a least one we should just close as being silly 21:45:13 dansmith: yeah, it sucks, to try keep it private 21:45:19 yes, some seem to be blown out of proportion 21:45:20 dansmith: I'm going to talk to VT and infar about a private gerrit 21:45:23 dansmith: that's true. I thought infra has said we could have a private gerrit by now, but I don't have recent status on that. 21:45:26 johnthetubaguy: I know 21:45:28 tonyb: that would rock 21:45:33 +1 21:45:41 so for those not sure what we are on about 21:45:41 tonyb: +1 21:45:42 http://security.openstack.org/vmt-process.html 21:45:46 anyway, I don't do enough around here, 21:45:55 so if you need another nova person, that's cool 21:46:06 and if not, because you're too large, then that's also fine 21:46:28 While I love this topic, I think we need to move on to end on time. dansmith, I'll chat to you about this after the meeting? 21:46:32 dansmith: don't take this the wrong way, but I would like you on the list so its not all rackspace 21:46:34 sure 21:46:49 heh 21:47:05 So, did we cover gate status? 21:47:17 nope 21:47:24 gate is fine 21:47:28 Yay! 21:47:30 done 21:47:32 Heh 21:47:33 moving on 21:47:34 :) 21:47:35