20:00:12 #startmeeting heat 20:00:13 Meeting started Wed Apr 24 20:00:12 2013 UTC. The chair is shardy. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:00:16 The meeting name has been set to 'heat' 20:00:23 #topic rollcall 20:00:34 here (technically on holiday) 20:00:34 shardy here :) 20:00:40 here 20:00:42 I'm here 20:00:48 me too 20:00:48 here 20:00:58 o/ 20:01:06 hi 20:01:14 Hi 20:01:19 Good to see you all again! 20:01:28 (read you all again?) 20:01:57 * mordred lurking 20:02:11 hi there 20:02:17 My first meeting… will be observing. 20:02:21 Hi all 20:02:24 howdy, just sent out a big thing about the future of heat, check openstack-dev :-p 20:02:31 zookeeper ftw, haha 20:02:45 cool, good to see you all :) 20:02:45 heh 20:02:54 #topic Review last week's actions 20:03:02 * shardy looks, I don't think there were any 20:03:18 *go to summit, behave at summit ? 20:03:20 definitely not from *last* week ;) 20:03:23 i mostly behaved 20:03:43 there was one. "Take a week to recover liver function" 20:03:44 zaneb: haha, yeah last *meeting* ;) 20:03:55 lots of actions from last week, which brings us to.. 20:04:00 #topic Summit review, action summary 20:04:22 So I've had a first pass over the BPs, trying to capture the Heat actions from all the discussion last week 20:04:42 https://etherpad.openstack.org/task-system is a big one, maybe more just discussion also 20:04:44 Can everyone please review, and check that I've not missed anything, that details and links are OK etc 20:05:06 harlowja: I raised a BP for that earlier but didn't have the link: 20:05:10 kk 20:05:12 thx 20:05:38 #link https://blueprints.launchpad.net/heat/+spec/workflow-library 20:05:56 good good 20:06:10 I've not yet targetted that to havana as I'm unsure who will do the work 20:06:27 ya, its a big cross-project topic 20:06:40 shardy: I will put some resources on task-library 20:06:42 *with discussion still ongoing 20:06:43 We have this so far (which already looks like a lot of work): 20:06:45 In starting to look at how rolling updates will be done, it may require a task library too 20:07:01 #link https://blueprints.launchpad.net/heat/havana 20:07:17 #action everyone to review havana BPs 20:07:30 kebray: great, lets talk after the meeting 20:07:43 Anyone else got anything summit-action related to mention? 20:08:02 shardy: also maybe we can talk about how said workflow-library could even be done (especially without affecting the other people working on heat/nova), that last part is the tricky one :-p 20:08:41 harlowja: that almost needs a separate meeting 20:08:42 harlowja: agree, I think it's a good idea, but there are still too many questions to commit to doing it for havana IMHO 20:08:55 harlowja: but I think more discussion on the ML is required first 20:09:08 zaneb: agree, should we pick that up in #heat and/or the ML? 20:09:11 one thing from the summit that bears repeating: Lots of working coming in will mean the heat core team will need help reviewing.. 20:09:16 zaneb shardy agreed, but u have to start with baby-steps, and baby-steps seem possible in havana 20:09:26 so, get out there and get reviewing. :) 20:09:57 using zookeeper seems to be polarizing, can someone summarize the argument against using it? 20:10:10 zaneb shardy i just don't want it to turn into, oh to big for havana, than oh to big for "I" series, then oh to big for "J" series :-p 20:10:27 stevebaker: starts with j, ends in ava .. 20:10:42 harlowja: agreed. baby steps. I haven't read your mailing list post yet, but I will look later 20:10:45 stevebaker: polarizing, i hope not, its a requriement in distributed systems imho, but maybe thats the polarizing aspect, haha 20:10:51 SpamapS: lol 20:10:59 zaneb: thx 20:11:02 harlowja: If someone steps up and presents a clear design/plan and some code, then we can target it to havana, not saying definitely no 20:11:25 zookeeper sounds like a big dependency to be taking on 20:11:29 Anyone want to take an action to formulate a plan in a wiki page? 20:11:29 shardy: of course, i'd like to have some code in both places, heat and nova starting to do the right thing with 'baby-steps' 20:11:36 I guess that is what makes it polarising 20:11:45 * shardy things harlowja is volunteering ;) 20:11:47 There are alternatives to zk 20:11:52 s/things/thinks 20:12:02 the *concept* is definitely one that should be investigated. 20:12:04 shardy harlowja When does a "clear" plan need to be presented by for Havana? 20:12:09 shardy: haha, i'm gonna make it happen, i'm just trying to figure out how to get there also ;) 20:12:10 agreed. zk is a possible solution to distributed state management 20:12:15 SpamapS: excellent. let's hear 'em ;) 20:12:39 zaneb: https://github.com/ha/doozerd 20:13:08 written in go.. MIT license 20:13:10 i just wonder about the maturity of all of those other ones, pretty much every big distributed system out there is using ZK :-p 20:13:16 so its not a slam dunk for OpenStack ;) 20:13:19 kebray: I'm not proposing a hard deadline, we can decide when we have a plan and resources to implement 20:13:30 harlowja: and yet, ZK still doesn't have SSL... :-/ 20:13:49 #action someone to create workflow-library wiki page 20:13:53 i can push for that hard @ yahoo (where ZK came from) if thats what we want 20:13:55 SpamapS: Go? Oh well, nice try ;P 20:14:26 Ok, anything else from summit or can we move on? 20:14:55 #topic Bug triage/milestone assignment 20:15:35 Ok, so I'd like everyone to change the bug states from "New" to something else (Triaged/Confirmed) when they assign a bug to themselves or someone else 20:15:51 ok 20:16:06 #link https://github.com/knipknap/SpiffWorkflow 20:16:15 Also please target the bug to the next release, so we can track all the assigned bugs for each milestone 20:16:38 I can do this, but it will make it much easier (for me ;) if everybody just does it when they change the assignee 20:16:46 it that's cool with everyone that is ;) 20:16:56 shardy: should we target bugs we don't actually know we'll fix? 20:16:58 sandywalsh: all for libraries, but we need to nail down something simple first imho, haha 20:17:00 s/know/think/ 20:17:16 I'd think that we only want bugs we're willing to commit to at least trying to fix 20:17:21 shardy: do you know about ttx.py? 20:17:24 SpamapS: If you're assigned a bug then I'm assuming you will either fix it or close it 20:17:52 Then we'll bump any bugs which aren't fixed before we tag a release (to the next milestone) 20:17:52 shardy: some bugs have medium/low priority and get done more when the time is right, not "now or never" 20:18:08 shardy: launchpad sanity checking script https://github.com/ttx/bp-issues 20:18:25 SpamapS: Ok, well that's pefectly fine, just change the status to something other than New 20:18:35 Right, Triaged seems appropriate for that :) 20:18:56 like "I have looked, I ack its a bug and is workable by a dev.. but Medium/Low probably means it isn't getting done soon. 20:19:07 I'm just trying to keep some sort of bug pipeline organised without doing pointless-clicking myself all day :) 20:19:16 stevebaker: thanks, I'll check it out 20:20:11 Ok, anything else on bug workflow? 20:20:41 #topic Blueprint review for Havana 20:21:07 #link https://blueprints.launchpad.net/heat/havana 20:21:38 So is everyone happy with BPs they are assigned atm? 20:22:04 I probably have too many so shout if anyone sees something I've got that they'd like to take ;) 20:22:23 it will change over time, no doubt 20:22:45 stevebaker: Yep, just want to make sure we're all vaguely happy for now :) 20:22:49 hm 20:22:51 shardy: i'd keep the workflow-lib open, i think the picture will get clearer soon i hope 20:22:56 which one of those is "run n+1 engines" ? 20:22:59 hoping for baby-steps in hava :) 20:23:17 to me that is the only Critical BP we discussed at the summit 20:23:20 shardy: I've been mulling over trusts, so I may be asking where it is in your queue at some point 20:23:34 as in, Heat is pretty craptastic for any sort of large scale usage without it. 20:23:46 https://blueprints.launchpad.net/heat/+spec/multiple-engines 20:23:58 SpamapS: I've not targetted that one correctly, thanks 20:24:17 fixed 20:24:21 shardy: np, just wanted to make sure it wasn't lost (in fact I keep losing it and forgetting the name.. ;) 20:24:44 who wants to be assigned to it for now? 20:25:06 seems like asalkeld was most actively pursuing it 20:26:03 I'll speak to asalkeld and see if he's happy to take it 20:26:32 If anyone wants to take any of the remaining unassigned ones, please do, then we can take another look next week 20:26:48 too late, I assigned him ;) 20:26:54 If anyone spots any other errors please let me know 20:27:00 stevebaker: haha :D 20:27:09 shardy: sorry for my ignornace but is it just core devs who can sign up, or can us new guys take one 20:27:18 stevebaker: that will teach him not to show up to the meeting ;) 20:27:26 randallburt: absolutely anyone can take one 20:27:42 * shardy would be very very happy to see some non-core devs on that list ;) 20:27:42 cool. I guess I'll take the dsl one unless there are objections 20:27:59 randallburt: Ok, thanks 20:28:29 I guess that one will be a bit of a team-effort, perhaps we'll need to create some child-BPs when we've fully defined what needs to be done 20:28:38 randallburt: an approved blueprint is just a vague indication that you're not wasting your time developing something 20:28:49 shardy: agreed 20:29:17 Yep, and the message I have ringing in my ears from summit is people want dsl/hot, so we definitely approve of that BP ;) 20:29:23 I can ride herd and expect the work to be broken up quite a bit 20:29:39 randallburt: Ok, cool, well thanks for offering to drive it 20:29:43 Thanks randallburt. 20:29:44 my pleasure 20:29:52 randallburt: right, the BP assignee is really just the one who answers the "how is it going" question at the meetings. You can break it up into a bunch of bugs or even sub-blueprints. 20:29:55 shardy: the message I heard was that people want features that cloudformation does not support 20:30:10 cool, thanks! 20:30:25 zaneb: Ok, true, that is a more accurate statement 20:30:50 people certainly talked about alternate template language a lot too tho ;) 20:30:56 anyway.. 20:30:59 a *lot* 20:31:12 developers talked about that a lot 20:31:29 Anything else on BP's before we move on to open discussion? 20:31:57 #topic Open discussion 20:32:11 I'm on leave next week 20:32:39 So I'm in the process of getting openstack/heat-templates repo created, ref ML discussion 20:32:51 shardy: sweet. :) 20:32:57 shardy: start by doing it in a personal github repo 20:33:06 lol 20:33:09 haha 20:33:34 then import happens magically when you do the gerrit review to openstack-ci-puppet 20:33:47 oh 20:33:54 do we still want to move to testr? 20:33:59 YES 20:34:02 I'm happy to take that on 20:34:07 stevebaker: I assumed we just wanted a new empty repo? 20:34:08 cool 20:34:14 is there already a bug? 20:34:23 since I guess we don't care about template revision history that much? 20:34:27 shardy: you could have a crack at an initial structure 20:35:03 stevebaker: Yep, I'm planning to create a tentative dir structure, copy in the CFN templates and heat-jeos TDLs 20:35:17 then other stuff can be added later 20:35:47 SpamapS: I though there was one, maybe there is a project-wide bug we can attach to 20:36:26 https://bugs.launchpad.net/heat/+bug/1172468 20:36:27 Launchpad bug 1172468 in heat "Replace nose with testr" [Medium,Triaged] 20:37:23 SpamapS: just remove the integration tests, I'll be looking at tempest soon 20:37:39 any other topics or did we all do too much talking last week :) 20:37:58 * SpamapS has been feeling fairly antosocial since Friday 20:38:03 anti even 20:38:55 randallburt: did you have any dsl interpreted versions of the example templates to share, or is that still in-progress? 20:39:19 I saw a few 20:39:26 * shardy must've missed the link 20:39:37 still in progress 20:39:48 randallburt: ok, cool 20:39:56 I thought Adrian added some to his articles, no? 20:39:56 we have a couple of the simple ones but are still refining 20:39:58 one thing I saw that was missing that I want is a separation from describing "an app" and "a vm" 20:40:31 but I think thats just because I don't know how to read them yet :) 20:40:32 we should probably get together soon and get everyones input since this will be a bit of a bear 20:40:43 final input that is. 20:40:53 kebray: they were just example snippets, I'm talking about fully converted versions of those in https://gist.github.com/hardys/5408410 20:41:04 (not necessarily all of them ;) 20:41:11 shardy: Andrew and I are working on that this week 20:41:13 shardy got it. Ok, definitely something randall can help ride herd on. 20:41:29 Ok, cool, not chasing just wondered if there was anything to look at yet :) 20:41:35 k 20:41:46 If there's nothing else we can end the meeting early? 20:41:46 we'll ping the list when we get something to show 20:41:55 early +1 20:42:00 randallburt: Ok, great, thanks 20:42:19 randallburt: if you want to have some offline discussion on the samples, I'd be happy to discuss 20:42:31 tspatzier: very much. thanks! 20:42:45 Ok, we can follow up re templates etc on ML and #heat, thanks 20:42:49 #endmeeting