21:05:22 <ttx> #startmeeting crossproject
21:05:23 <openstack> Meeting started Tue Jan 13 21:05:22 2015 UTC and is due to finish in 60 minutes.  The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:05:24 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:05:25 <ttx> Our agenda for today:
21:05:27 <openstack> The meeting name has been set to 'crossproject'
21:05:27 <ttx> #link http://wiki.openstack.org/Meetings/CrossProjectMeeting
21:05:30 <ttx> Should be a short one
21:05:37 <ttx> #topic CLI Sorting Argument Guidelines
21:05:38 <morganfainberg> yay short meetings
21:05:44 <ttx> #link https://review.openstack.org/#/c/145544/
21:05:55 <ttx> This is a new openstack-spec to standardize CLI sorting arguments across projects
21:05:55 <jungleboyj> morganfainberg: +2
21:06:40 <jungleboyj> ttx +2 from me.
21:06:42 <bknudson> this project has problems with consistency in general.
21:06:55 <bknudson> the most consistent thing to do here would be to have every project be different
21:07:21 <annegentle_> bknudson: noooooo.....
21:07:24 <ttx> Affects Cinder, Glance, Ironic and Neutron apparently, but also future projects that would make use of a sort parameter there
21:07:26 <jungleboyj> bknudson: -2 for synciality. ;-)
21:07:35 <ttx> Not sure if Steven Kaufer is on IRC
21:07:40 <ttx> But then that doesn't prevent us from discussing it here
21:07:40 <jokke_> o/
21:07:44 <jungleboyj> (sp) synicality
21:07:47 <thingee> sounds fine to me. we already approved the spec for cinder, but of course, curious to see what other projects think
21:08:54 <ttx> thingee: ideally we would have alignment from affected projects
21:09:09 <ttx> and then that would be a guideline for future such projects
21:09:20 <annegentle_> ttx: so what was the way TC members are supposed to vote in cross-project specs? +1?
21:09:28 <bknudson> one way to get alignment is to put the code in a library.
21:09:29 <ttx> annegentle_: yes
21:09:33 <annegentle_> ttx: k thanks
21:09:34 <ttx> same as anyone else
21:09:48 * devananda is back
21:09:50 <annegentle_> bknudson: I think this type of spec can then inform the openstack CLI of the best practice
21:10:09 <annegentle_> bknudson: so with dtroyer on board I think we can look for that sort of future
21:10:16 <annegentle_> bknudson: unless I misunderstand your point
21:10:21 <ttx> bknudson: not sure being in a library enforces alignment :)
21:10:49 <ttx> one can always ignore the lib. I'd prefer wide PTl approval on the openstack-spec, like for all openstack-specs
21:11:17 <ttx> mestery, nikhil_k: around ?
21:11:38 <ttx> if not, please review https://review.openstack.org/#/c/145544/ when you get a chance
21:11:55 <ttx> Other comments on that spec ?
21:12:39 <ttx> I guess not...
21:12:45 <morganfainberg> from a keystone perspective i don't think we have much of anything to say.
21:12:48 <asalkeld> looks reasonable
21:12:55 <morganfainberg> because we're relying on openstack client at this point
21:12:58 <bknudson> we've pretty much deprecated keystone cli.
21:13:05 <ttx> please cast your vote and i'll put it on the TC agenda for fainl tallying of the votes in a future meeting
21:13:09 <ttx> final*
21:13:21 <morganfainberg> our CLI is mostly deprecated / frozen and wont see support for v3/much future development
21:13:33 <devananda> looks reasonable to me, though I need to see what changes are needed in our current CLI/API
21:13:35 <jungleboyj> ttx +2
21:13:37 <thingee> bknudson: is openstack client using this format for sort/dir?
21:13:49 <ttx> devananda: cool
21:13:51 <jungleboyj> thingee: Assume you will put your vote out there?
21:14:04 <sdague> thingee: I believe currently it's just nova, because the conversation evolved out of novaclient patches
21:14:14 <morganfainberg> thingee, ttx, so i think the osc folks should be involved in this as well, specifically
21:14:25 <ttx> dtroyer: ^
21:14:30 <thingee> morganfainberg: +1 that's where I was going to go with that question :)
21:14:34 <sdague> morganfainberg: dtroyer has already commented on the review
21:14:39 <sdague> with a +1
21:14:39 <ttx> dtroyer: you might want to chime in on the review as well
21:14:41 <morganfainberg> sdague, ah cool.
21:14:41 <thingee> sdague: ah true
21:14:45 <dtroyer> thingee, morganfainberg: we are…and OSc will be very close to this if not exactly there
21:15:01 <ttx> hah, he did already
21:15:20 <ttx> OK, last comments before we move on ?
21:15:21 <morganfainberg> then i'm a +1 on the spec :) will vote accordingly w/ comment
21:15:52 <ttx> #topic Proposed Vancouver Design Summit format
21:15:57 <ttx> #link http://lists.openstack.org/pipermail/openstack-dev/2015-January/054122.html
21:16:09 <ttx> I posted proposed changes to the Design Summit format last week, if you have questions let me know
21:16:26 <ttx> So far the reception seems to have been good -- so unless someone objects I'll have the events team work on a room layout that would allow this format
21:16:31 <bknudson> I was wondering about cross-project sessions -- is that a fishbowl or a workroom?
21:16:35 <thingee> ttx: good job. Glad that were addressing the issue of ops feeling like second class.
21:16:35 <annegentle_> ttx: have you done a walkthrough with nova/neutron and cross-project?
21:16:43 <devananda> ttx: +1 overall
21:16:52 <ttx> bknudson: fishbowl, although technically we could also have smaller workgroups there...
21:17:07 <morganfainberg> ttx, i am a +1 on that - didn't reply to the list though.
21:17:10 <bknudson> ok, thanks.
21:17:13 * nikhil_k joins in
21:17:15 <ttx> annegentle_: walkthrough?
21:17:23 <ttx> One challenge we will probably have to solve is the scheduling.
21:17:25 <annegentle_> ttx: an imagining if you will
21:17:35 <annegentle_> ttx: yes, for scheduling and where the space runs out
21:17:50 <jungleboyj> Proposal looked good as long as the Work Group sessions are in individual rooms, not shared.
21:17:59 <ttx> annegentle_: I have a proposed room layout. We shouold have an awesome space in Vancouver
21:18:14 <ttx> About scheduling, not sure we'll be able to rely on pre-allocated slots and small tweaks (like we use to do) to avoid conflicts this time
21:18:19 <jokke_> ttx: by quick look your proposal is great ... by the Paris experience, that's exactly what's needed!
21:18:23 <devananda> ttx: +100 on clearly splitting fishbowl and workinggroup sessions
21:18:32 <ttx> We might have to tag sessions with the necessary audience and then use a constraint solver to produce the final schedule for us
21:18:35 <devananda> ttx: what will the timing / overlap be with the main conference?
21:18:45 <annegentle_> ttx: yeah I think the split is right but then the constraints haven't been revealed yet
21:18:47 <ttx> Main conf runs Mon-Thu, we run Tue-Fri
21:18:57 <annegentle_> ttx: so it helps a bit to see constraints is all
21:19:04 <ttx> friday will be all workinggroup
21:19:14 <morganfainberg> ttx, friday being working group will be great
21:19:22 <morganfainberg> it worked very well at the last summit
21:19:24 <ttx> and hopefully food will be worse so nobody will hang around just to eat.
21:19:29 <morganfainberg> hahah
21:19:40 <ttx> what do I say... Food can only be worse.
21:19:47 <jungleboyj> morganfainberg: +2
21:19:49 <devananda> ttx: ok. so ops overlap tue-thu concerns me a bit. what do you think?
21:19:52 <eglynn__> should help with the lunch queues
21:19:54 <asalkeld> ttx the changes seems ok to me
21:20:37 <devananda> ttx: not concerns-me in a way that we should change anything - i think this is a good structure
21:20:43 <ttx> devananda: you mean ops summit running Tue-Fri adds more conflicts ?
21:21:02 <morganfainberg> silly question, did we determine [a little off topic] if there was a 3rd level of participant that let people duck out after the main conf? i think we said a reason a lot of people hung around was because they felt like they needed to for the last day since it was expo or full-access
21:21:11 <ttx> There is the option os running some ops sessions on Monday... the trick being, it would still feel like a different event outside the design sumit then
21:21:25 <morganfainberg> and so they paid for the full access and stuck around + wifi was only in the design summit side the last day(s)
21:21:48 <morganfainberg> ^^ that question might be something for a different meeting
21:21:53 <morganfainberg> (mine that is)
21:21:59 <bknudson> is it all in the convention centre?
21:22:21 <devananda> ttx: it may split operators time more, without giving them a clear "here is the ops day" indication
21:22:24 <ttx> morganfainberg: not sure... I think Paris was special because people just stayed over the weekend
21:22:41 <ttx> so they were around on Friday when usually they fly back
21:22:41 <morganfainberg> ttx, fair enough that too. cause well Paris.
21:22:48 <ttx> bknudson: yes
21:23:11 <ttx> devananda: what we do with Monday is still up in the air
21:23:14 <devananda> ttx: ah. having it all in the same building will help a lot, and actually may obviate my concern about the overlap
21:23:18 <morganfainberg> devananda, ++
21:23:26 <ttx> depends a lot on how much room we'll actually have
21:23:34 <jokke_> devananda: at least I felt that the ops separation was awful ... we saw really small cross participation and it took at least me a full missed day to even find it
21:23:40 <ttx> fifieldt__ said he would rather have sessions on Monday too
21:23:45 <ttx> (ops sessions)
21:24:49 <ttx> On another note, I'll probably have to ask PTLs for their ideal allocation mix /before/ the new PTL election
21:25:05 <ttx> so the old PTl would pick the mix and the new PTL would use it
21:25:23 <ttx> I don't see much options to avoid that, we'll need to work on allocation way before the election
21:25:34 <eglynn> the old PTL is likely to get some team consensus on the preferred mix
21:25:44 <eglynn> ... i.e. not just pick it unilaterally I'd guess
21:25:44 <ttx> yes, and a lot of old = new
21:25:48 <devananda> ttx: seems fine to me
21:25:57 <ttx> just a detail I wanted to mention
21:26:09 <eglynn> yeap, doesn't seem like a huge issue
21:26:27 <ttx> OK, last questions / comments on that ,
21:26:28 <ttx> ?
21:26:49 <asalkeld> all good
21:26:52 <ttx> #topic Open discussion & announcements
21:26:57 <thingee> ttx: o/
21:27:06 <ttx> thingee: go for it
21:27:34 <thingee> For the juno release, there have been deprecation warning going out from Cinder on v1 being removed by Kilo.
21:27:44 <thingee> Some have expressed to me one release is not enough notice.
21:28:06 <thingee> regardless of v2 being around since G. Comments?
21:28:12 <bknudson> is it causing problems to keep it there?
21:28:42 <ttx> agree with bknudson, it's more a tradeoff than a hard rule. One release is the minimum
21:28:43 <eglynn> isn't two full cycles the norm for the deprecation path on major API versions?
21:28:53 <morganfainberg> eglynn, typically
21:29:14 <thingee> bknudson: that's a good point. It has been little technical debt the past two releases.
21:29:14 <eglynn> but yeah, not an iron law or anything
21:29:47 <ttx> thingee: if you can afford to keep it for two in deprecation, by all means, please do :)
21:29:47 <jungleboyj> thingee: Just the occasional reminder to not add function there.
21:30:04 <thingee> mostly there is an issue that exists with two service types being around for volume and volumev2.
21:30:12 <morganfainberg> thingee, i recommend making people happy and keep it for 2
21:30:24 <thingee> Ideally I would like volume to point to x.x.x.x:8776 instead of a particular version endpoint, and just do discovery.
21:30:47 <jungleboyj> thingee: Can't we do discovery and leave v1 around still?
21:30:52 <thingee> jungleboyj: yes
21:31:13 <thingee> ok that's all for me. I guess expect me going around to everyone using cinder client to rid volumev2.
21:31:15 <thingee> ttx: thanks
21:31:19 <morganfainberg> jungleboy, but at a certain point you want to drop that stuff from the codebase is all
21:31:28 <asalkeld> i don't think heat is on to cinder v2 yeat
21:31:29 <thingee> getting rid in L-1 that is
21:31:35 <thingee> asalkeld: it is
21:31:40 <ttx> We had 1:1 syncs today, where I tried to collect all project-specific deadlines...
21:31:44 <ttx> Logs at:
21:31:47 <ttx> #link http://eavesdrop.openstack.org/meetings/ptl_sync/2015/ptl_sync.2015-01-13-08.59.html
21:31:52 <asalkeld> ok, thought there was a bp for it
21:31:53 <ttx> We'll skip 1:1s next week.
21:31:57 <jungleboyj> morganfainberg: +2 ... Was just making sure we could hold on to it a little longer and still move forward.
21:32:03 <morganfainberg> ttx, confirmed 2-wk prior to k3 code-feature-freeze for keystone.
21:32:52 <ttx> wow, that's what rehubbing means
21:32:53 <morganfainberg> ttx, erm code-feature-proposal
21:32:53 <morganfainberg> wow.
21:32:53 <morganfainberg> nice netsplit
21:32:53 <ttx> I guess that means the meeting is over
21:32:53 <devananda> heh
21:32:53 <morganfainberg> hey and we didn't lose the openstack bot
21:32:53 <ttx> I think we still have the meetbot on our side
21:32:53 <morganfainberg> ttx, ++
21:32:53 <ttx> BTW, if anyone is interested in chairing the cross-project meeting, let me know. I'm fine sharing chairing :)
21:33:18 <morganfainberg> ttx, ^^ if you didn't see the comment about code freeze for keystone
21:33:29 <ttx> I did see it fly
21:33:29 <morganfainberg> ttx, 2wks prior to k3 as we discussed
21:33:32 <morganfainberg> :)
21:33:35 <morganfainberg> cool
21:33:44 <ttx> Anything else, anyone ?
21:34:00 <morganfainberg> (anyone that is left)
21:34:00 <morganfainberg> aaaaannnd round 2
21:34:00 <ttx> we still win
21:34:03 <morganfainberg> hah!
21:34:05 <dhellmann> woo
21:34:12 <devananda> \o/
21:34:16 <ttx> tempted to cotinue to play
21:34:24 <asalkeld> thingee, gone
21:34:33 <ttx> but then if I can't close the meeting we may have an issue
21:34:35 <morganfainberg> ttx, let it ride! one more time.... it'll be all good.
21:34:57 <ttx> woosh
21:35:08 <jokke_> yeah ... I think we split out as that's around 300 on the other side
21:35:09 <samueldmq> lol
21:35:09 <asalkeld> thingee: yeah heat is ok: https://github.com/openstack/heat/blob/master/heat/engine/clients/os/cinder.py#L38-L46
21:35:33 <ttx> any last comment before we close the meeting ?
21:35:39 <ttx> or before we get rehubbed
21:36:07 <ttx> I'll take that as a "no"
21:36:10 <ttx> #endmeeting