11:46:37 #startmeeting ptl_sync 11:46:38 Meeting started Tue Sep 9 11:46:37 2014 UTC and is due to finish in 60 minutes. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 11:46:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 11:46:41 The meeting name has been set to 'ptl_sync' 11:46:44 #topic Ceilometer 11:46:52 #link https://launchpad.net/ceilometer/+milestone/juno-rc1 11:47:09 Are thos 3 blueprints the only FFEs you consider ? 11:47:20 * ttx parses the ML for any request 11:47:22 yes 11:47:55 ok let's go through them, top to bottom 11:48:03 https://blueprints.launchpad.net/ceilometer/+spec/bigger-data-sql 11:48:21 how far are we on this one? 11:48:33 the main patch to remornalize the resource schema is approved and working it way thru' the gate as we speak 11:48:47 *remnormalize :) 11:48:54 ok, found it 11:49:08 is there any followup patch? 11:49:18 the second patch to switch from sql-a ORM to sql-a core is still under review 11:49:34 Mike Bayer (sql-a author) has helpfully chimed into the discussion 11:49:51 that would be https://review.openstack.org/#/c/113524/ ? 11:50:19 yes 11:50:24 #info pending on https://review.openstack.org/#/c/113524/ and https://review.openstack.org/#/c/111313/ 11:50:32 we need some further quantification of the performance gains, but it should be landable 11:50:33 https://blueprints.launchpad.net/ceilometer/+spec/paas-event-format-for-ceilometer 11:50:42 what about this one 11:50:42 approved, working its way thru the gain 11:50:52 *thru the gate 11:51:10 that's documentation only, non-controversial 11:51:17 ah ok 11:51:21 #info Doc-only 11:51:25 forgot about that 11:51:35 https://blueprints.launchpad.net/ceilometer/+spec/ipmi-support 11:51:51 impi-support is still the furthest away from landing 11:52:09 Edwin is engaging with the reviews, but the TZ latency adds to the delay 11:52:25 that's the two reviews at https://review.openstack.org/#/q/topic:bp/ipmi-support,n,z ? 11:52:27 last week I signalled that this one would be expected to land last of the 3 11:52:30 yes 11:52:46 #info https://review.openstack.org/#/q/topic:bp/ipmi-support,n,z -- a bit far away still 11:52:56 ok, that looks in good shape 11:53:04 Ideally the SQL stuff would merge this week 11:53:08 yep, agreed 11:53:12 the IMI stuff can wait until next Tuesday 11:53:16 IPMI* 11:53:22 yes, that's fair 11:53:32 then we'll rediscuss it 11:53:55 do you have a concrete date in mind for cutting RC1? 11:54:06 ... or driven by actual progress in landing stuff? 11:54:23 not really... it's normally when you empty your release-critical bug list 11:54:32 OK, got it 11:54:35 so you should target the "must fix before release" bugs to rc1 11:54:47 then depending on how fast you burn that list, we adapt 11:55:12 ideally the RC1 would appear before the end of the month 11:55:34 but if everything looks good and you have no bugs, then it could appear next week, I guess 11:55:40 eglynn: about the design summit, one question 11:55:42 right, I'll give the currently targetted bug list another trawl and punt anything non-critical that doesn't look like it's getting traction 11:55:44 shoot 11:55:58 If we do meetup-style on the Friday, would you rather have one full day, or only half a day? 11:56:23 i.e. would half-a-day be sufficient ? 11:56:32 I'd rather take the full day, even if folks are a bit zombie-ish by the afternoon 11:56:43 ok 11:56:47 (seeing as the pod-time will otherwise be limited) 11:56:54 trying to do a trade-off between comfort and time 11:57:06 is there an alternative event in mind for the Friday afternoon, or more an early finish? 11:57:23 you can spend time in other meetup rooms, or leave early 11:57:46 though some will get a Friday afternoon half-day, that's the nature of the game 11:57:46 yeah, making it kinda optional like that would be fine 11:58:20 ok. Anything you'd like to discuss at meeting today? 11:58:32 nope, I'm good 11:58:42 eglynn: cool, talk to you later! 11:58:43 thank you for your time! 11:58:50 SergeyLukjanov: ready when you are 12:02:13 ttx, I'm here 12:02:16 ttx, morning 12:02:44 #topic Sahara 12:03:00 https://launchpad.net/sahara/+milestone/juno-rc1 12:03:01 #link https://launchpad.net/sahara/+milestone/juno-rc1 12:03:07 oops 12:03:14 https://blueprints.launchpad.net/sahara/+spec/edp-swift-trust-authentication 12:03:36 Is that all contained in https://review.openstack.org/#/q/topic:bp/edp-swift-trust-authentication,n,z ? 12:03:44 ttx, I think all CRs are on review for this blueprint 12:03:47 ttx, I think so 12:04:01 and how is that going so far ? 12:04:35 ttx, it's already working while manually testing 12:04:46 ttx, and all patches have at least one +2 I think 12:04:58 so just needs to adjust tests ? 12:05:03 ttx, yup 12:05:16 #info all in review @ https://review.openstack.org/#/q/topic:bp/edp-swift-trust-authentication,n,z -- needs some tests tweaks but in good progress 12:05:26 https://blueprints.launchpad.net/sahara/+spec/cluster-persist-sahara-configuration 12:05:49 two other blueprints are very small and easy and both are on review 12:05:57 the last one is re docs 12:06:09 is there more to it than https://review.openstack.org/#/q/topic:bp/cluster-persist-sahara-configuration,n,z ? 12:06:14 nope 12:06:23 #info https://review.openstack.org/#/q/topic:bp/cluster-persist-sahara-configuration,n,z on track 12:06:32 https://blueprints.launchpad.net/sahara/+spec/move-rest-samples-to-docs 12:06:35 #info doc only 12:06:36 we just need to persist some internal versions for validation, cluster-persist-sahara-configuration is about it 12:06:43 SergeyLukjanov: is that all your FFEs ? 12:06:47 yup 12:07:17 OK, so I think they should all land before next Tuesday 12:07:34 ack, I think there will be no issues with it 12:07:35 ideally the tust delegation part before end of week 12:07:46 because it's a bit more disruptive 12:08:07 do we need some additional actions to obtain FFE or I could just mail to the ML with the list of granted FFEs? 12:08:22 no, you can email the list 12:08:48 so next question would be about the Design Summit -- would Sahara like to have a meetup-like thing on the Friday ? 12:09:02 ttx, ack re FFEs 12:09:30 Do you think half-a-day would suffice ? Or would you rather have the full day ? Benefit to half-day is ability to participate to some others 12:09:31 ttx, in addition to the slots on other day, or only friday slots/ 12:09:44 in addition to slots on Wed/Thu 12:09:57 ttx, half of day is enough for sahara 12:10:20 ok, that way you can participate to the Infra/QA room 12:10:24 :) 12:10:46 SergeyLukjanov: Anything you want to add to meeting agenda for today? 12:11:01 just to ack - sahara will have some slots on wed/thu + we could have half day meetup like on friday? 12:11:44 yes. Current plan would be to have cross-project stuff on Tuesday, traditional slots scheduled on Wed/Thu and meetups on Friday 12:11:59 ttx, ok, thx 12:12:01 there just will be less "traditional" slots available overall 12:12:10 since we have two days for all rather than 3 12:12:29 ttx, so, I think it'll be cool to have something meetup-like for sahara at friday in addition to the traditional stuff 12:12:38 ttx, yeah, I understand 12:12:43 and a few new projects 12:12:51 yes, and limited space ! 12:13:03 my puzzle for the week, see if it can all fit 12:13:32 heh, I'm sure you'll make the best schedule as always :) 12:13:50 SergeyLukjanov: so basically when brainstorming sessions you need to see which can have a clear slot and schedule presence, and which could be solved in the Friday informal discussion 12:14:01 ttx, got it 12:14:06 usually the ones you want external feedback on are better as scheduled slots 12:14:22 yeah, to have them in publicly available schedule 12:14:30 and to have some fixed timing 12:14:32 project internal policies otoh are a good fit for Friday 12:14:54 yup, I think we have several good topics for friday 12:15:18 also release objectives setting isa good topic for the Friday 12:15:28 anyway, talk to you later! 12:15:36 thank you! 12:15:40 dhellmann: ready when you are 12:16:24 ttx, oh, I have one more q. 12:16:32 SergeyLukjanov: sure, shhot 12:16:35 shoot* 12:16:45 ttx, could we think oslo and migrate to oslo.XXX libs before the rc1? 12:17:03 ttx: good afternoon 12:17:37 SergeyLukjanov: depends how disruptive the change is... it's generally a good thing to adopt a library before release... but some of those adoptions may refactor a bit too much code 12:17:52 so it's a case-by-case thing 12:18:00 ttx, I think that's ok if it's mostly the import chaning CRs 12:18:05 dhellmann: opinion on that ^ ? 12:18:44 I would be conservative about that for now. Plan for K-1 adoptions of anything you aren't already using, unless you have a serious issue that means you need a fix in a lib. 12:19:02 and we do allow backporting fixes, too 12:19:09 to the incubator, that is 12:19:44 SergeyLukjanov: does that answer your question? 12:19:52 "only if you really need to" ? 12:19:58 ack, for sahara we'll check do we really need to switch or sync oslo 12:20:06 ttx, yup 12:20:10 ttx, dhellmann, thx 12:20:15 #topic Oslo 12:20:28 #link https://launchpad.net/oslo-incubator/+milestone/juno-rc1 12:20:35 SergeyLukjanov: some of the libs might be just as easy to adopt as a sync, if you have to do that anyway, so let's talk when you figure out how far out of date you are 12:21:09 dhellmann, ack 12:21:17 dhellmann: did we agree those graduate-* BPs should rather apear on the library project rather than on the -incubator? 12:21:41 If yes I can move them 12:21:45 or at least try 12:21:48 yes, that makes sense 12:21:54 I can move them after breakfast, too 12:23:43 I'm on it 12:23:53 That leaves us with... https://launchpad.net/oslo/+milestone/next-juno 12:24:26 Some of those should probably be kilo at this point, right? 12:24:28 ttx: I set up the pylockfile project yesterday, too, so it looks like all of these blueprints can move out of the incubator project to their library project 12:24:39 dhellmann: ok, will do it once we are done 12:25:32 I'll talk to josh about the taskflow thing. The amqp 1 driver is in oslo.messaging now but there's work left to be done. I suppose we should split that one. 12:25:47 dhellmann: yeah 12:26:02 I should create kilo series and next-kilo there 12:26:10 the mysql driver one is blocked on a review in sqlalchemy-migrate, let me find that link 12:26:11 so that stuff can be moved out 12:26:28 https://review.openstack.org/#/c/110611/ 12:27:28 I'll put the semver work for pbr in kilo at this point 12:27:37 I suppose that means moving the spec in oslo-specs, too 12:27:57 dhellmann: I'll create the kilo series so that you can do the cleanup 12:28:04 ok 12:28:41 any question on the juno release side ? 12:28:55 I think you have a good handle on all the lib releases now 12:29:13 did you clarify when we could cut proposed/juno? there were some fixes pending iirc 12:29:36 yeah, we have a list of things we'd like to merge, and I haven't checked them yet this morning 12:29:45 should be soon 12:30:04 dhellmann: ok, maybe track them using the -rc1 targeting 12:30:19 that will let you have a clean list of targets you have to complete before rc1 12:30:22 yeah, I'll make sure there are bugs for those 12:30:34 and remove what's not a "need to have" fro the list 12:30:49 dhellmann: ok, last question, about the design summit 12:31:31 I hadn't planned an Oslo meetup on the Friday, since I think you should be covered with the cross-project stuff and the scheduled sessions... but did you want one ? 12:32:06 it's tricky because you may want to visit the other rooms 12:32:28 yeah, we might use the pods that day 12:32:45 I think we'll have some things to decide that may take longer than the session slots 12:33:01 ok, so a smaller space could be enough 12:33:19 yeah, we probably only need room for a few of the cores to sit down together for those discussions 12:33:57 dhellmann: ok, that's good feedback. I originally planned to recover pod space to make an additional meetup room, but it might make sense to keep some around and pack more than one project in 12:34:30 oh, yeah, I assumed they would be available that day 12:34:44 dhellmann: ok, anything for the meeting today ? 12:35:25 we should tell folks that we're trying to have final library releases on 18 sept to give time for testing before the rest of the release candidates 12:35:53 I only really expect a couple of updates at this point 12:36:01 #info we're trying to have final library releases on 18 sept to give time for testing before the rest of the release candidates 12:36:01 otherwise, that's it 12:36:15 dhellmann: ok, maybe just re-mention it when I link to the log 12:36:19 ok 12:36:26 dhellmann: ok, thanks and talk to you later! 12:36:32 have a good afternoon! 13:21:15 ttx: I think I have all of those bugs/bps updated 13:23:26 * ttx quicklooks 13:23:39 https://launchpad.net/oslo/+milestone/next-juno 13:25:39 looks good 13:39:02 ttx: I have our Neutron meeting from 1400-1500UTC today, can we move our 1:1 to immediately after that? 13:39:21 ttx: I can ping you in-channel if the neutron meeting ends early. 13:40:22 mestery: I'll be on a call then 13:40:38 mestery: we can do it now if that works better for you 13:40:42 ttx: OK, then lets just do it at our normal time and I'll multitask :) 13:40:47 ttx: I am about to get on a call with markmcclain now :) 13:40:53 hheh 13:41:10 ttx: I'll multitask, no issues, thanks! 13:42:34 mestery: my call may end early too. So maybe around 15:15 13:43:03 ttx: Sure! Anytime after 1500 works for me, so ping me when you can. I've cleaned up RC1 LP now, it should be a quick sync. 14:07:06 jgriffith: around? 14:13:57 dolphm: ready when you are 14:15:38 ttx: o/ 14:16:00 #topic Keystone 14:16:04 https://launchpad.net/keystone/+milestone/juno-rc1 14:16:09 All FFEs done I see 14:16:13 \o/ 14:16:20 you win! 14:16:37 (at the same time it's pretty sane that keystone is feature-complete first) 14:16:52 ha, i'm not sure that's a good thing for the overall project 14:17:13 no other FFE request pending? 14:17:22 we have two wishlist bugs in there, but that's all 14:17:44 we're going to discuss in today's meeting if we want to pursue them in juno, or delay 14:17:52 ok 14:18:26 so at this point it's pretty simple... compile a list of release-critical bugs, get them assigned and fixed.. and we can tag and open Kilo 14:19:04 https://launchpad.net/keystone/+milestone/juno-rc1 will be our target list. If you want to have "nice-to-haves" in the list, better use some tag 14:19:09 we're still seeing our bug count creep up with new bugs being opened though, so it's probably going to next week before we start counting in the right direction 14:19:16 like "juno-rc-potential" or whatever 14:19:32 alright, we have a lot of those :) 14:19:37 sure, the list should evolve over time anyway 14:19:59 you can also just get the nice-to-have bugfixes in without necessarily make a list of them 14:20:11 ok, questions on RC1? 14:20:29 what's the target date for RC's? 14:20:48 final RCs / release, i mean 14:21:09 "when the buglist is empty." 14:21:17 ideally before the end of the month obviously 14:21:29 so we'll aggressively refine the list as we go 14:21:37 ack 14:21:47 I'd say two weeks from now is a good time 14:21:59 For the design summit, as you know we are considering meetup-like space for the friday. Do you think you'll need half a day or would prefer to have a full day ? 14:22:28 would a half day be only afternoon? 14:22:45 yes, could be 14:23:04 i'd be fine with a half day - i imagine we'll be doing more reviews than anything else 14:23:23 ok 14:23:35 i'm not strongly opinionated either way though 14:23:47 sounds like a useful schedule change either way 14:23:49 I need to see how much spaces we need 14:23:54 so I can work on layout 14:23:58 cool 14:24:11 so more people asking for half-days means better spaces :) 14:24:42 obviously big projects like Nova and Neutron wil use the full day anyway 14:30:43 david-lyle: o. 14:31:25 ttx: o/ 14:31:34 #topic Horizon 14:31:48 https://launchpad.net/horizon/+milestone/juno-rc1 14:32:11 david-lyle: is that a complete list of the FFEs under consideration? 14:32:19 * ttx crosschecks with ML 14:32:29 one was proposed on mailing list 14:32:39 going to ask in an hour for another sponser 14:32:46 "Support of Cinder QOS Specs in Horizon" ? 14:32:51 1 owner, 1 sponsor so far 14:32:52 yes 14:33:10 let's see if we have room for it 14:33:21 let's go over those in your list 14:33:28 https://blueprints.launchpad.net/horizon/+spec/neutron-subnet-mode-support 14:33:28 sure 14:33:32 how is that one going? 14:33:53 all patches I can find seem merged 14:34:08 I need a status update, in an hour as to where support in the client is at 14:34:27 ok, might just actually be finished? 14:34:29 and determine if we need a client release or not 14:34:56 https://blueprints.launchpad.net/horizon/+spec/tagging 14:35:28 there are 3 related bps 14:35:33 that is one of them 14:35:49 around the new meta-data catalog feature in glance 14:35:51 + nova-metadata-flavors + nova-metadata-aggregates ? 14:35:56 yes 14:36:09 these are in good shape and have been previously reviewed 14:36:12 how are they collectively coming up ? They sound a bit far away still 14:36:29 How much time do you see them collectively take? 14:36:30 we were just waiting on the client to release and a requirements bump 14:36:36 that happened yesterday 14:36:49 I hope to start merging them later today 14:37:07 #info this one is related to nova-metadata-flavors and nova-metadata-aggregates 14:37:18 this one has been on our radar for a while and new we were going to be time constrained due to the larger process 14:37:22 david-lyle: ok, so ideally they would be all in by the end of the week? 14:37:34 I feel confident about them landing by sometime Fri 14:37:42 (all related 3) 14:37:43 ok 14:37:45 yes 14:38:12 That means you still have "room" for another one if you have core support for it 14:38:16 the QOS proposal contains 3 patches that haven't landed 14:38:43 what is their current state? 14:38:46 yeah, I need more core support 14:39:11 they have had a few reviews, but not the same level of scrutiny as the meta-data patches 14:39:36 how about we add it to the RC1 list, as "Deferred", so that we keep track of it ? When you finally decide, you can either put status back to "Needs code review", or remove the milestone target 14:39:49 ok 14:40:11 ok, that looks good. 14:40:27 we'll make more drastic cuts next week if any are still open 14:40:33 sure 14:40:50 david-lyle: other questions on RC1? 14:41:13 estimate for a target date? 14:41:35 as soon as your targeted bug list is empty 14:41:44 ideally in two weeks, in all cases before the end of the month 14:41:58 ok, just wanted the window 14:42:15 we have a lot of it would be nice bugs, but less that we can't live without 14:42:18 the window is already opened, but I don't see the first ones happening before mid-Sept 14:42:50 sounds reasonable 14:43:06 david-lyle: about design summit -- we may be able to set up meetup-like space for the Friday. Would Horizon need a full day of informal discussions, or would half a day be enough ? 14:43:43 (so that you can actually participate a bit to others meetups if you want to see what is coming your way) 14:44:01 that would be in addition to how many more formalized slots? 14:44:27 that would be in addition to "less" scheduled slots. 14:44:40 since we would have two days instead of 3 for those 14:44:59 * ttx checks numbers 14:45:14 I imagine a half-day would by enough, especially if we have the pods or other informal meeting areas 14:45:39 in ATL you had 7 sessions, so in Paris you would have 4-5 scheduled sessions and at least half a day of unscheduled time 14:45:53 to address "all other business" 14:46:29 ttx: that sounds good 14:46:45 cool. All depends how much time you need to reach alignment 14:46:55 I think Horizon team is pretty well aligned generally 14:47:03 typically 14:47:24 david-lyle: ok, that's all I had -- anything you want to discuss at meeting tonight ? Any red flag ? 14:47:39 I think we're set now that the glanceclient has released 14:47:48 no blockers 14:48:14 thanks! 14:51:31 david-lyle: cool; thx 14:51:38 jgriffith: still not around? 15:05:57 ttx: Ready and waiting, ping when your call is complete. 15:07:57 mestery: ack 15:08:03 ttx: \o/ 15:09:15 mestery: probably in 10min 15:09:51 ttx: OK, ping me here, I may step out for a bit while waiting. We'll connect soon though :) 15:28:33 ttx: still here 15:28:33 mestery: I'll ping you when I have a hole in my schedule 15:28:50 jgriffith: ok, let me cover swift first 15:28:55 ttx: Ack 15:28:57 notmyname: around? 15:29:13 ttx: hi! 15:29:20 #topic Swift 15:29:42 notmyname: how is your next release coming up? 15:30:01 ttx: for project scheduling, we're still on for oct 6 for the RC, with the final on oct 16 for the integrated release 15:30:07 https://launchpad.net/swift/+milestone/next-juno 15:30:13 #link https://wiki.openstack.org/wiki/Swift/PriorityReviews 15:30:25 that has a section for current open patches that I'm looking at to land for juno 15:31:27 I'm a bit worried on the timing 15:31:36 oh yeah? 15:31:44 maybe have a small safety buffer and plan RC earlier? 15:32:06 because of the length of the RC or because of getting stuff in for the RC? 15:32:15 getting stuff in on the RC 15:32:42 I guess we could shorten the RC baking period 15:33:00 i.e. if last patch lands on Oct 8 it's not really a showstopper 15:33:31 we've got four weeks until the RC. I'm not too worried abotu getting those listed patches in master in that timeframe. they've all been in progress for a while 15:33:49 notmyname: ok, if it's not too busy, Iguess it works 15:33:57 and, like you, I'm also ok with targeting the 6th and slipping a couple of days if necessary 15:34:16 I just want to avoid an unpleasant plan B where we'd ship the previous release in Juno final :) 15:34:39 of course. I find that highly unlikely :-) 15:34:49 ok! 15:34:52 I'd prefer to cut a release sooner with some of those missing before doing that :-) 15:35:01 anything else you wanted to mention? 15:35:25 any other thoughts on doing any gap analysis follow-up before paris? 15:36:04 notmyname: maybe we could try to squeeze it in one of the last TC meetings before reelection 15:36:09 also, I'm not too happy that the defcore group asked for designated sections feedback, and I gave it, and they didn't update their proposed havana sections 15:36:26 notmyname: you probably didn't give the right feedback! 15:36:41 heh. I said their list was good! 15:37:00 ttx: ok, no real worries on the TC front. if there's time, I'd love to get a feel for things before paris, just to make sure the in-person things are time-efficient 15:37:06 there are new meetings this week, if you want to get that remark on the record 15:37:15 notmyname: I understand that 15:37:38 notmyname: about design summit and the Friday meetup-like setup; would you rather have a full day or would a half-day be enough ? 15:38:00 It's a bit of a two-edged sword 15:38:37 that meetup-like setup would be self-scheduled by who's there? 15:38:42 yes 15:38:49 ie kinda like a mid-cycle meetup 15:38:57 yes 15:39:12 is that the only time we'd have to meet as swift contributors? 15:39:32 formally 15:39:38 there would be pods, but I doubt we ca, make space for one pod per project 15:39:46 so the pods would be reusable 15:39:50 ok 15:40:01 i.e. you would claim a table as your pod by putting your sign on it 15:40:11 not sure that will scale perfectly 15:40:33 in general, full day probably means less scheduled time, more unscheduled time. 15:40:41 and the other scheduled sessions on wed and thurs are for cross-project stuff? 15:40:45 I still have to build a layout 15:41:11 no. Tuesday for cross project, wed.thu our traditional scheduled sessions (just less of them) and Friday unscheduled time 15:41:23 (will appear on schedule as "swift contributors meetup" 15:41:29 ah ok 15:41:44 so does each program get some scheduled time on wed.thurs? 15:41:48 yes 15:41:58 ah ok 15:42:12 just expect ~30% less, maybe more if you pick the full-day option 15:42:28 obviously the less time at the summit makes mid-cycle meetups more important 15:42:51 agre 15:42:52 e 15:43:09 but it seems that if there are already some scheduled sessions, plus the ad-hoc pods, a half-day on friday would probably be a good place to start 15:43:28 you don't have to decide now, I just want to build a proposal based on this informal poll 15:43:39 (proposal = room layout) 15:44:17 ok, anything you want to discuss at meeting later today? 15:44:28 my cross-project questions are around defcore and gap analysis and paris scheduling. so no :-) 15:44:40