08:58:47 #startmeeting ptl_sync 08:58:48 Meeting started Tue Nov 18 08:58:47 2014 UTC and is due to finish in 60 minutes. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:58:49 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 08:58:51 The meeting name has been set to 'ptl_sync' 08:58:52 #topic Heat 08:59:14 asalkeld: o/ 08:59:27 hey, first timer here 08:59:34 be gentle 08:59:38 :-) 08:59:45 My first question would be, do you plan to follow release management stuff directly, or would you prefer to delegate that to a release management liaison 08:59:46 ? 09:00:01 for now, me 09:00:09 cool. 09:00:10 we are having a weekly meeting 09:00:17 and i'll bring it up 09:00:28 (if anyone else wants to do it) 09:00:50 Second question -- about how we curate the blueprint list in LP 09:01:23 ok, I think i need to tweek some 09:01:25 As you know we use the "milestone target" field to curate per-milestone list of stuf being done 09:01:38 problem is, Launchpad lets anyone set that field 09:01:51 so we use a combination of priority *and* target milestone 09:02:00 only project drivers can set priority 09:02:23 so I have a script that will automatically remove from the milestone any unprioritized blueprint 09:02:30 i think i need to read this: https://wiki.openstack.org/wiki/PTL_Guide#Blueprints 09:02:43 For example, on https://launchpad.net/heat/+milestone/kilo-1 ... there are 3 undefined 09:02:59 (two of them implemented, so probably safe :) 09:03:30 I can run that script on a crontab, and it will automatically clean up the list 09:03:39 so that only the stuff you vet appears there 09:03:45 ttx, does anything automatically happen when a spec is approved? 09:04:15 or is there a way for me to figure out what is out of sync 09:04:18 asalkeld: nothing at thius point. I have another script that can set the blueprint fields for you on spec approval 09:04:29 nice 09:04:35 but it's not trigfgered automatically 09:04:41 and I need to rebase it 09:04:55 http://git.openstack.org/cgit/openstack-infra/release-tools/tree/spec2bp.py 09:04:58 checking all the bp's is going to get old fast 09:05:24 yes, and having two parallel paths is also a bit painful 09:05:38 which is why I wrote the autokick script to autoclean things up 09:05:45 ok, that should help 09:05:59 do I run that or you? 09:06:00 My question would be, do you want that script running for Heat (like it did in the past ?) 09:06:21 yes, if it can run automatically 09:06:26 It usually simplifies a lot (you choose what appears on the list, rather than have to curate it all the time) 09:06:40 yes I run it on a crontab from some server in the cloudTM 09:06:51 yes, please that amazing:-) 09:07:07 ok, let me tell you what happens if I run it now 09:07:38 so that you can check it's not crazy 09:08:10 i'd expect some need setting to approved 09:08:20 http://paste.openstack.org/show/134302/ 09:08:40 If I run the script, it would remove the milestone targets for both of those 09:08:51 that sounds right 09:08:59 so you may want to set a priority for them if you want them to stay 09:09:16 ok, makes sense 09:09:22 asalkeld: the general goal here is to use the milestone pages to communicate the main things being worked on 09:09:30 like a report 09:09:51 ok, I'll try keep on that 09:10:02 obviously once storyboard is everywhere it will be all integrated :) 09:10:30 ttx that spec2bp.py - I need to run that? 09:10:32 ok, script is running now 09:10:55 asalkeld: so spec2bp just helps if you want to set the fields when a spec is approved 09:11:07 I can send you an email about it 09:11:18 thanks, that would help 09:11:19 #action ttx to send aslakeld info about spec2bp 09:11:30 I have another version up for review that I need to rebase 09:11:36 and then I'll advertise it again 09:11:50 it would be nice if that ran on approval 09:12:27 asalkeld: how representative would you say https://launchpad.net/heat/+milestone/kilo-1 is? 09:12:38 is is missing a lot of stuff that is being worked on ? 09:13:07 asalkeld: the main issue with spec2bp is that there is no way to CREATE the damn spec using launchpad api 09:13:21 so the spec has to be manually created 09:13:26 well there are people working on unapproved specs 09:13:26 err.. the bp 09:13:35 but that is ok 09:13:46 we should be approving some specs soon 09:13:51 and i'll clean up 09:14:17 (small'ish specs) 09:14:47 https://review.openstack.org/#/q/status:open+project:openstack/heat-specs,n,z 09:14:49 I think at some point we said we could track the "not approved yet" specs as "Blocked" 09:14:52 quite a heap 09:14:58 if they are part of the plan 09:15:31 not sure that is so nice 09:15:54 asalkeld: ok, I guess you should add the BP as the specs get approved. Don't forget to set a priority if you want them to survive the script. We'll review progress there next week 09:16:13 and I'll send you that email about the script 09:16:17 ok, thanks ttx 09:16:26 any topic for the cross-project meeting tomorrow morning ? 09:16:35 i don't think so 09:16:44 ok then, talk to you later! 09:16:49 later 09:16:53 mikal, johnthetubaguy: around? 09:17:04 I am! 09:17:22 yes, but kinda busy with a deploy right now :( 09:17:26 #topic Nova 09:17:44 kilo-1 is a mess that I hope to sort out last friday, but hopefully this afternoon now 09:17:46 ttx, hey, any chance to sync earlier today? (/me travelling and will be available for ~ hour today) 09:17:57 SergeyLukjanov: sure, just after Nova ? 09:18:22 ttx, awesome, thx! 09:18:50 mikal: as far as release management liaisons are concerned, I think we can safely say that John is the liaison for Nova, but you inteznd to still follow the release syncs and all ? 09:19:13 ttx: well, that's a little bit up to johnthetubaguy 09:19:24 ttx: he's our release guy, and I'll attend these if he finds value in me doing so 09:19:36 I guess I need to anyways as prep for tomorrow's release meeting? 09:19:38 mikal: trying to abuse him while he is busy with the deploy 09:20:15 I'm not sure johnthetubaguy attended release meetings in juno? Just the syncs? 09:20:21 mikal: the cross-project meeting might turn into a busy meeting for all PTLs and cross-project liaisons 09:20:31 Ahhh, ok. 09:20:34 That would make sense 09:20:42 So yeah, perhaps this is the john / ttx syncup then 09:20:44 I'd be fine with that 09:21:15 we can discuss that next time 09:21:38 Sure 09:21:39 #action johnthetubaguy/mikal to clarify formal release management liaison delegation 09:21:40 OK, its more a super bad time for me, that cross project meeting, but yeah 09:21:55 johnthetubaguy: I can cover for you in that one if I know what you need done there 09:21:56 I am happy to be release liaison, if that makes sense 09:22:06 johnthetubaguy: yeah, you're totally on the hook for that 09:22:09 johnthetubaguy: you don't *have to* attend the other one, the PTL can represent 09:22:12 mikal: cool, we need to sync anyways 09:22:19 johnthetubaguy: agreed 09:22:41 I'm just trying to empower the CPLs (cross project liaisons) to encourage people to step up 09:23:28 mikal, johnthetubaguy: that's what would get kicked off milestones if I enabled the adjust script right now: http://paste.openstack.org/show/134304/ 09:23:43 (remedmber: script kills all unprioritized stuff) 09:24:11 you may want to do a tour and ping me when I can enable the script 09:24:17 doesn't have to be just now 09:24:26 Huh, there's some weird stuff targetted and in that list 09:24:43 ttx: thats cool with me, I think, I haven't looked at that list since the summit, so it gonna be a bit odd right now 09:24:44 mikal: yeah, same issue as always, anyone can set the milestone is LP 09:24:47 Remind me what criteria the script uses to kick things? 09:24:50 there are approved specs I need to deal with, etc 09:24:54 which is why we use the prio/milestone combo 09:25:05 Yeah, I agree we should wait for John 09:25:09 mikal: the script removes unprioritized stuff 09:25:15 But that script looks like a good thing to enable after that 09:25:24 I am good turning that on, it worked well last release 09:25:52 #action johnthetubaguy to do a final tour of milestones before pinging ttx to enable autokick 09:26:11 sure, will do, this afternoon hopefully 09:26:24 I'll rework spec2bp soon to help with setting fields 09:27:29 mikal, johnthetubaguy: is " I am happy to be release liaison, if that makes sense" a "yes" ? 09:27:38 * ttx can update the CPL page 09:27:57 ttx: yes 09:27:57 I am good with that being a yes 09:28:05 #info johnthetubaguy will be release liaison for Nova 09:28:08 johnthetubaguy is the CPL for Compute / Nova / Whatever we're called 09:28:23 the Nova thing 09:28:49 ok, any topic you want to raise for the cross-project meeting? 09:29:51 Not that I can think of, apart from that I'd like to see projects who intend to have mid-cycles announce dates ASAP 09:30:26 ok, we can talk a bit about those tonight 09:31:04 Ta 09:31:14 Anne and I would like to understand why people organize them, too, see if anything can be done to reduce the need for them 09:31:24 so that can be discussed tonight/tomorrow 09:31:39 mikal, johnthetubaguy: ok, talk to you later! 09:31:43 Yep, I am getting a lot of push back from a small number of companies about having one at all 09:31:55 ttx: cool, thanks 09:31:59 I do think its nessesary for Nova at the moment, but we're working to make that less true in the future 09:32:32 mikal: we could try a remote one, but yeah, its tricky, means someone is up late at somepoint 09:32:33 mikal: I think they fear that too many decisions would be made there, rather than just work getting done 09:32:52 hackathons/social gatherings are fine 09:33:11 "Core secret meeting", they kind of have to attend it 09:33:13 Hmmm 09:33:25 Its complicated 09:33:29 it is indeed 09:33:45 especially when you factor the delays for organizing travel in 09:33:46 Nova is under a lot of pressure to deliver a lot, and that's very hard to do without keeping people in sync 09:33:58 But some of the same people applying that pressure don't want to sync up 09:34:02 i.e. predicting 3 months in adfvance you'll need one 09:34:10 * ttx needs to run quickly for an errand 09:34:17 Fair enough 09:34:17 SergeyLukjanov: be right back 09:34:20 We can talk more later 09:34:25 yes, at that meeting :) 09:34:33 ttx, ok 09:45:16 SergeyLukjanov: ready now 09:45:24 ttx, me too 09:45:29 #topic Sahara 09:45:57 IIRC you said you would handle release liaison directly 09:46:08 ttx, yup 09:46:19 #info SergeyLukjanov will handle release liaison directly 09:47:05 Should I enable the script that clears out unprioritized targeted blueprints ? 09:47:15 at this point it appears it wouldn't kick anything out 09:47:27 ttx, yup 09:47:40 ttx, if we have something unprio than we need to pull it out 09:47:57 ok, the script will now do that automatically 09:48:31 ttx, ack 09:48:32 SergeyLukjanov: how complete is k1 ? 09:48:49 ttx, I think it's on a good shape 09:48:52 does it represent most stuff you plan to work on ? Or just a very partial view at this point ? 09:48:52 #link https://launchpad.net/sahara/+milestone/kilo-1 09:49:17 ttx, due to the summit and vacations after it - it's partial IMO 09:49:29 ttx, we probably have some other stuff that'll be included 09:49:31 ok, we'll refine it as we get closer 09:49:53 ttx, I'm on vacation this week too, so, will se next week I think 09:50:33 ok 09:50:47 anything you want to discuss at meeting today? 09:50:58 ttx, I think no, everythong is going ok 09:51:15 ttx, horizon stuff is in a good progress too 09:51:16 alright then, ttyl 09:51:21 ttx, thx! 13:00:04 ttx: knock, knock, ready when you are 13:02:16 yep ,just a sec 13:02:55 finishing an email, 1 min 13:03:19 cool, no rush 13:04:34 eglynn: o/ 13:04:37 #topic Ceilometer 13:04:44 #link https://launchpad.net/ceilometer/+milestone/kilo-1 13:04:45 sorry about that 13:04:49 np! 13:04:58 so we're begining to flesh out kilo-1 13:05:03 but early days yet 13:05:10 lots of specs reviews in flight 13:05:15 and more to come I expect 13:05:25 If I turn on the autokick script, it will kick out kafka-publisher (from kilo-1) 13:05:31 because it has no priority set 13:05:45 Should you fix it before I turn it on ? 13:05:46 yeah the spec hasn't landed yet 13:05:51 ah, hm 13:06:02 I'll manually kick, not sure how got targetted 13:06:10 I'll manually kick *it, not sure how *it got targetted 13:06:19 ok, cal I enable the script crontab for ceilometer now ? 13:06:35 can* 13:06:47 yep 13:06:51 please do 13:06:58 ok done 13:06:59 so on the specs process for kilo 13:07:13 I missed this session at summit https://etherpad.openstack.org/p/kilo-crossproject-specs 13:07:41 we mostly said that would be used only for truly cross-project specs (i.e. likely to affect everyone) 13:07:41 but I take there was some unease at how spec across the board worked in juno? 13:08:14 so the practice within individual projects will remain as-was? 13:08:19 It felt like people were happy with it and wanted to standardize 13:09:27 The TODO in that etherpad translate what we'd like to try pretty well 13:09:49 I guess the key question in my mind is "not all blueprints need a spec?" 13:10:03 from the etherpad it seems that wiggle room was closed off somewhat? 13:10:15 I would say, not all features need a spec, and not all features need a blueprint 13:10:33 no, projects can still very much set the bar where they want 13:10:53 as far as "integrated release" is concerned, it's the blueprint view (i.e. the subset of features that we track for release purpose) that matters 13:11:30 cool, it's the set different of those two categories above that I was wondering about 13:11:55 i.e. features that do need a BP (for tracking) but not necessarily a spec (as the way forward is self-evident) 13:12:19 set *difference 13:12:27 my view on it is... use spec for stuff which requires design/choice, and/or use blueprint for stuff that makes sesne to track as part of the release cycle 13:12:36 cool, got it :) 13:12:47 there may be smallish stuf that falls between the cracks 13:12:59 in which case the commit message should be self-explaning 13:13:33 I assume you will handle release management directly ? No release management liaison ? 13:13:55 yep 13:14:17 I'm also down as stable-maint liaison for now 13:14:29 no one jumped at the call for volunteers 13:14:57 a lot of the more active cores are already liaising on something 13:15:10 (oslo, docs, qa etc.) 13:15:14 Note that if they are part of your roadmap, you can also track "features that are waiting for spec approval" in blueprints using "Blocked" status 13:15:31 cool, I'll adopt that practice 13:15:33 i.e. if those are a priority in your release plans 13:15:39 * dhellmann makes a note of that, too 13:15:46 Working on https://review.openstack.org/#/c/108041/ 13:16:14 cool 13:16:29 that way blueprints are truly a reporting tool, disconnected from spec status 13:16:36 reporting/planning 13:16:40 nice :) 13:16:54 would be nice if only Launchpad could create a blueprint from the API 13:17:15 ok, I think I've got the urgent points covered 13:17:28 yep, thanks for your time :) 13:17:36 #info Eoghan will do release management duties directly 13:17:47 eglynn: anything to add to agenda for meeting tonight ? 13:17:52 nope 13:18:12 ok then 13:18:26 adding you to https://wiki.openstack.org/wiki/CrossProjectLiaisons 13:18:56 dhellmann: ready? 13:19:01 ttx: here 13:19:05 #topic Oslo 13:19:27 I suspect you'll stay the release management liaison for Oslo, whatever that means ? 13:19:37 i.e. you'll be the one attending this sync ? 13:19:39 yes, I think that's going to be simplest 13:19:57 I'll put myself on the list 13:20:02 we don't exactly discuss the same things as other projects, but I think we still benefit from a sync 13:20:06 I'm on it 13:20:08 ok 13:20:50 Looking to the autokick script, not sure it makes sense in Oslo world 13:20:51 I've asked for volunteers for stable maint and documentation liaisons, too, but no one raised their hand 13:21:13 that's the thing that removes blueprints if they don't have a priority set? 13:21:13 that's the script which will remove unprioritized blueprints from kilo milestones 13:21:29 I'd be happy to have that running to help keep us honest 13:21:43 we'd have to list all the launchpad projects 13:21:53 oh, so it wouldn't run against the project group? 13:22:04 hmmm, worth a try 13:22:11 if not, it's not a big deal 13:22:22 https://api.launchpad.net/devel/oslo object has no attribute 'getSeries' 13:22:41 I would have to hack the script to dereference the projectgroup 13:22:43 I've explained to everyone about using the new release tool, so we shouldn't get too cluttered 13:22:47 I guess it's doable 13:22:56 if you find time, fine, otherwise don't worry about it 13:23:17 #action ttx to look into supporting projectgroups in autokick.py, so that we can apply it to Olso 13:23:28 in the mean time it won't run 13:23:32 ok 13:23:48 I wanted to read https://wiki.openstack.org/wiki/Oslo/ReleaseProcess and comment 13:23:57 maybe I can do that now 13:23:58 please do! 13:24:13 I'll ping you in 5 min once I read it ? 13:24:30 I'm going to the bakery for breakfast right after we're done, but I'll be back in 20 min 13:24:52 OK, let's close the 1:1 first 13:24:56 sure 13:25:01 anything to add for todays meeting agenda ? 13:25:18 I don't think so 13:25:28 is it worth mentioning the alpha version decision? 13:25:42 that's the only thing that really comes to mind for me 13:26:14 dhellmann: I think the only person who I can remember opposing a part of it (stable dep capping) would be markmc 13:26:25 ok 13:26:35 so not sure disucssing it in meeting would help/change 13:26:47 I have a note to myself to propose the cap change in the stable global requirements list, so I'll try to do that today 13:26:48 although pinging him could be seen as a courtesy 13:26:55 yeah, I'll do that, too 13:27:08 ok then, go to bakery, buy croissants, ping me when back. I'll read that page in the next 5 min 13:27:31 and we can quickly discuss it when you're back 13:27:33 ok, bbiab 13:33:21 * ttx braindumps random remarks 13:34:20 - the main goal of the "fix leftover bugs" stop is to let people try to manually fix the bugs Launchpad is timeouting on, could be good to mention that 13:35:38 - the email example is using an alpha pre-release, would be good to switch to a "normal" release template 13:36:30 - Would be cool to add the link to the Launchpad milestone page, since we build one 13:36:47 Otherwise looks good. 13:44:36 ttx: back 13:44:48 points 1 and 2 ack, I'll do that today 13:44:54 not sure what you mean for point 3? 13:46:31 the announcement email doesn't include a link to the Launchpad milestone page that oslo_release.sh helps build 13:46:41 ah, ok, that makes sense 13:46:58 I need to automate generating that email body more anyway 13:48:40 ack 15:03:39 ttx: Here when you're ready! 15:04:19 mestery: hi 15:04:24 #topic Neutron 15:04:51 mestery: you fine with being the release management liaison for kilo, or would prefer to delegate that ? 15:05:07 ttx: Ack, I'm good with that. 15:05:25 Been looking into enabling autockick for Neutron/kilo 15:05:35 http://paste.openstack.org/show/134396/ shows what would happen 15:05:39 Please do! I thought that was very useful during Juno. 15:05:42 * mestery looks 15:05:54 you might want to fix those by hand before I enable the crontab 15:05:54 Looks good to me 15:06:07 ok, enabling 15:06:10 Thanks! 15:06:17 I had been manually kicking a bit the past few weeks already ;) 15:06:27 updating wikipages 15:07:09 ok done 15:07:12 Thank you! 15:07:21 #info Kyle will handle release management directly 15:07:35 https://launchpad.net/neutron/+milestone/kilo-1 15:07:54 feels like this is incomplete at this point 15:07:57 This will be changing over the next week as we start approving specs 15:08:10 We're in the process now, people are spending lots of time reviewing specs 15:08:17 Note that if you want to track specs-still-needing-approval as part of your milestone roadmap, you can use "Blocked" 15:08:28 Ah, right, I'll do that! 15:08:34 useful for tracking essential stuff 15:08:40 that you know you need to do anyway 15:08:51 Perfect 15:08:57 We'll discuss meetups in general at the cross-project meeting today 15:09:06 anything else you'd like to add ? 15:09:47 That's about it! markmcclain and I are in the process of proposing the advanced services spin-out from neutron now 15:09:54 ack 15:09:56 We'll be emailing hte list and gettingthat on a TC agenda soon 15:10:11 We have our mid-cycle Dec 8-10 as well, may be worth noting that in the logs here. 15:10:25 #info Neutron mid-cycle Dec 8-10 15:10:39 Otherwise, that's it for now! Hopefully next week I'll have more interesting BPs in kilo-1 ;) 15:10:51 ok then, have a good day! 15:10:55 you too, thanks ttx! 15:16:13 nikhil_k: ready when you are 15:16:23 ttx: ready 15:16:26 #topic Glance 15:16:41 have one item from my side 15:16:49 nikhil_k: sure, shoot 15:17:07 ttx: the author of this commit https://review.openstack.org/#/c/133858/ 15:17:15 requested it to be in juno 15:17:41 ttx: wanted to see if we can get that one in and what was "general" process of allowing such changes in stable/* 15:18:01 do we delegate to the stable maintainers? (now per project) 15:18:45 nikhil_k: feels like it should be acceptable per stable rules 15:19:06 if stuck, you can raise a thread on -dev with [stable] prefix 15:19:22 you can also reach out on #openstack-stable 15:19:33 sure, I'll remove the -2 and make it +2 then 15:19:36 ah nice, sure 15:19:49 nikhil_k: I assume you'll be handling release management liaison duties for Glance ? 15:20:12 ttx: yes 15:20:27 #info Nikhil will do relmgt liaison for Glance 15:20:56 The other thing I wanted to discuss are autocleaning of milestone plans 15:21:17 We are using blueprints and target milestone to communicate which features are likely to land in which milestone 15:21:31 it's like a roadmap that is communicated outside the project 15:21:49 so we maintain pages like https://launchpad.net/glance/+milestone/kilo-1 15:22:14 BUT launchpad unfortunately doesn't restrict who can set "starget milestone" on a blueprint 15:22:29 sure, I'm aware (for the most part at least) :) 15:22:31 so we use a combination of priority being set AND milestone 15:22:45 ah ok 15:22:50 and I have a script that will remove any unprioritized stuff from milestone 15:23:03 so that nobody can sneak stuff in without the project drivers approval 15:23:19 That script is not enabled yet for glance in kilo 15:23:24 but I can enable it 15:23:26 sure, the thing is in Glance we use specs 15:23:33 (it wouldn't have any effect right now) 15:23:37 so just wanted to get some of you thoughts on it 15:23:45 So, you can use spec to vet for design 15:23:53 and blueprints to communicate a roadmap out 15:24:11 I have another script called spefc2bp that can be used to sync status between spec and bp 15:24:24 this would be an issue after a spec is approved right? 15:24:38 but unfortunately that script can't create a blueprint, so you still have to create it byu hand 15:24:42 that script sounds nice 15:24:57 you can track a feature in blueprints even befgore the spec is approved 15:25:07 sure, that's ok. BP are only for bookkeeping anyways with most design details in the spec 15:25:12 if you know a specific spec is really on the roadmap 15:25:25 ohk, that issues sounds bad 15:26:01 Imagine you have a key feature you want delivered in Kilo release, you can add a blueprint to track it even before the spec is approved 15:26:09 just set status to "Blocked" 15:26:31 I'll rebase my spec2bp script and land it, then send an email about it 15:26:33 gotcha 15:26:35