22:01:50 #startmeeting Horizon 22:01:50 Meeting started Tue Dec 10 22:01:50 2013 UTC and is due to finish in 60 minutes. The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:01:51 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 22:01:54 The meeting name has been set to 'horizon' 22:02:00 o/ 22:02:00 Hello Horizon folks! 22:02:02 hello 22:02:05 hi all! 22:02:05 o/ 22:02:07 hi 22:02:07 hello! 22:02:25 hi 22:02:50 howdy 22:03:00 https://wiki.openstack.org/wiki/Meetings/Horizon 22:03:27 hi! 22:03:46 No general announcements today 22:03:52 so let's jump in 22:04:12 #topic Discuss integration of TripleO UI 22:04:41 So this item was added to the agenda based on a mailing list thread 22:04:54 david-lyle, so what are the conditions for the merge? 22:05:33 david-lyle, given that make sense, cause all UIs belongs in Horizon 22:05:39 The general question posted was would it make more sense for the UI component of Tuskar to merge into the Horizon program rather than the TripleO program 22:06:21 I think from a code make-up perspective and an eventual home perspective this makes perfect sense, however.. 22:06:28 a note: on the TripleO meeting we agreed that yes. 22:06:34 So far we've carried UIs in our own tree when a project gets integrated 22:06:58 I'm not sure what the organization structure would be 22:07:21 jpich, to be fair the aims of tuskar UI is much larger than say Trove as far as Horizon is concerned 22:07:22 david-lyle, not sure either 22:07:48 russellb: o/ 22:07:59 I am torn to pieces as well 22:08:08 If I understood correctly a big part of Tuskar was a new "Infrastructure" dashboard, at the same level than project and admin? Maybe my information is outdated 22:08:10 I think it will fit there 22:08:22 yes, It's a well populated dashboard 22:08:23 david-lyle, well I do see it as separate panel, that just shows when Tuskar is present 22:08:26 jpich: that's my understanding 22:08:26 at one side it all make sense to be part of UI project 22:08:42 jpich: correct 22:09:01 we have had several requests about adding features, which belong to infrastructure 22:09:02 lblanchard, jtomasek: Thanks! 22:09:11 I do think it makes sense, but I think it would still have to be a separate code base for now 22:09:18 so it would make sense for horizon to have it 22:09:38 and there will be a need for core-reviewer overlap 22:09:48 david-lyle: +1 22:09:49 david-lyle, so do you mean separate codebase placed under Horizon program? 22:09:49 david-lyle: +1 if that is not an issue from Horizon point 22:10:13 david-lyle: Until when? 22:10:15 lsmola_: yes, until TripleO/Tuskar is an intergrated part of the release 22:10:21 I think this sounds very reasonable 22:10:22 That makes sense to me 22:10:39 david-lyle, I am not really sure how that can be done, but yes that seems to be reasonable 22:10:48 The code base will be what I image it is today, an extension of the current Horizon repo 22:10:58 we even could add it now or very soon 22:11:06 yes please 22:11:08 and make tuskar a config option 22:11:14 the sooner the better 22:11:23 david-lyle, ok 22:11:30 rdopieralski had the plugin concept merged today 22:11:46 mrunge, well yeah, we could add it now, and it would stay hidden unless user will have Tuskar 22:11:47 mrunge, that's right 22:11:53 and a demo option is about configuring tuskar 22:11:55 mrunge, which they wont 22:12:04 switch on/off 22:12:09 less overhead for the tuskar folks then 22:12:18 exactly 22:12:48 that sounds good 22:12:54 So, this jumps topics a bit, but review backlog worries me a lot 22:13:00 even with the current scope 22:13:03 and we don't get any evil surprises, when tuskar becomes official 22:13:33 adding more core-reviewers from tuskar should help, but I think patches are languishing much too long now 22:13:35 mrunge, I have to check that plugin patch 22:13:44 so, david-lyle should we move to topic #2? 22:13:52 (reviews) 22:14:07 well, I think it effects #1 22:14:17 yes, I agree 22:14:26 So let's circle back to #1 in a sec 22:14:27 david-lyle: I thing good amount of tripleo-core reviewers working on tuskar-ui will jump to Horizon reviews right away (me included) 22:14:38 #topic Reviews 22:14:53 that was brought by me 22:15:03 I would appreciate a general rule from when we accept new components' UI into our tree. We've turned away code before because the project was incubated but not integrated, a clear guideline would help being fair and treating everyone the same way 22:15:21 and I'm currently concerned about a long review queue 22:15:28 jpich +1 22:15:48 according to russell's stats we have: 22:15:59 New patch sets in the last 30 days: 545 (18.2/day) 22:16:48 We are getting a lot more patches and contributors than before, which also very cool - even if it means reviews are taking longer at the moment (which sucks) 22:17:23 still I see very few core reviews during the last days 22:17:23 yes, which means we need more reviewers, hopefully these new contributors are willing to lend some review time 22:17:45 Personally I've had less time for reviews since the Summit though I hope to be back to normal in the new year - sorry for not helping picking up much of the slack at the moment 22:17:51 I think several new contributors have jumped right in 22:17:57 and some old contributors as well :) 22:18:02 but in general, we'd love to see more folks doing reviews 22:18:03 Yep we should definitely encourage people to review :) 22:18:23 david-lyle, we have been blessed today to make mor reviews, 6 more people from tuskar should do at least a review per day 22:19:01 I think the current review load has no margin of error, that is if a core goes on vacation has a work commitment we fall terribly behind 22:19:04 david-lyle, to make things moving :-) 22:19:05 good news lsmola_ 22:19:18 so, in that light I would love to see a larger review pool 22:19:26 s/review/reviewer/ 22:19:44 so all the patches are perfect by the time a core reviewer gets to it :-) 22:20:00 :-) 22:20:01 ha! I'd love to see it 22:20:11 I believe ;) 22:20:45 So, to the current tuskar folks, do you worry about the latency for your incubating project? 22:21:08 incubating projects require a faster turn around time then we currently provide 22:21:08 we do have blazing review turnaround right now 22:21:10 david-lyle, yes 22:21:13 so it is a concern 22:21:36 how many of you tuskar guys have been core at tripleO? 22:21:50 mrunge: for UI? 22:21:51 is there something that the tripleo team is doing that speeds up their review time? Or is it simply the number of patches vs. reviewers? 22:21:53 david-lyle, we have to make a lot of patches for tuskar-ui and also for horizon in coming months 22:21:54 yes 22:22:10 mrunge, me 22:22:17 mrunge: I think 5-6 22:22:21 me, jtomasek, lsmola 22:22:22 mrunge, jomara, jtomasek 22:22:26 hehe 22:22:40 ok, I see. 22:23:06 in my eperience, not answering to reviewers requests has been an issue in the past 22:23:16 esp. from you tuskar guys 22:23:23 hate to bring this up, but all tuskarUI folks are RedHat? 22:23:28 that slowed the process down a bit 22:23:46 david-lyle: that's correct 22:23:47 yeah 22:23:49 david-lyle: yep 22:23:55 just the tuskar-ui guys, not all tripleo 22:24:22 mrunge, tzumainn, lsmola, jtomasek, jomara + non core akrivoka, rdopieralsky 22:24:23 I ask merely because we try not to have all 3 people, author, 2 core be from the same company, a working principle in general 22:24:49 I'm not sure how that changes for an incubated project 22:24:56 not sure that it bothers me either 22:25:05 I think it's a good rule 22:25:08 I was tuskar core? 22:25:10 *principle 22:25:17 mrunge, (though rdopieralsky have kickass python skills, he just came late) 22:25:32 mrunge: you don't know about that? 22:25:34 mrunge, no the mesage is to you :-) 22:25:35 :) 22:25:35 no 22:25:46 hahaha 22:26:03 mrunge: lsmola_ uses an odd IRC client that puts a comma when he tab completes 22:26:07 mrunge: shame him! 22:26:25 jomara, hehe 22:26:32 +1 on the general concept of reviews not all coming from the same company for any specific patch (although i must admin nebula has been guilty of that in the past, but it was a smaller project then) 22:26:44 +1 ohnoimdead 22:27:02 that's rule I like 22:27:13 Agreed 22:27:23 admin=admit fyi 22:27:28 but in the past, that has not be a problem 22:28:07 stupid question: where does the tusker-ui code currently live? 22:28:13 ohnoimdead: you mean +2 reviews, right? 22:28:18 seems fair to have at minumum 1/3 of (reviewr1+reviewer2+author) not be from the same company 22:28:20 ohnoimdead, as part of tripleo program 22:28:22 jtomasek: yeah 22:28:31 ohnoimdead, https://github.com/openstack/tuskar-ui/tree/master/tuskar_ui 22:28:38 mrunge: thanks! 22:28:39 if I'm not terribly wrong 22:29:20 david-lyle, well that could slow things down in few coming months, though won be a problem later I guess 22:29:47 it will 22:29:52 david-lyle, we are currently rebuilding it due to feedback from the conference 22:30:02 I think we have some logistics to work out, but I would be in favor of adding tuskar ui 22:30:09 so, would it be possible for you tuskar guys to make a plan, how the code shall be merged into our code base? 22:30:29 ... just to discuss this at the next meeting? 22:30:38 is that too quick? 22:30:52 mrunge, can we sit on that tomorrow with rdopieralsky, seems like you already have a plan :-) 22:31:16 I was just thinking about reading a 20k lines of code patch 22:31:22 :| 22:31:22 and I don't like the idea 22:31:39 jpich, I think the main difference here between tuskar and other service UIs is that Tuskar is a more fundamental change to the direction of Horizon 22:31:45 david-lyle, would it be possible to allow to only Redhat people for the Infrastructure tab? 22:32:05 I'd rather not react late to such a scope change 22:32:06 lsmola_, is there a reason for this? 22:32:06 david-lyle, before other people will get to pace of what Tuskar-Ui is 22:32:25 Ismola_: O.o 22:32:49 lsmola_: not sure, I think there may need to be some mechanics before the two core teams merge 22:32:55 I think lsmola mean, is it ok if for a while the +2 reviews happen to come from the same company for this specific part of the code 22:33:07 ah 22:33:16 jpich, yes, thank you for rephrasing :-) 22:33:18 that sounds different 22:33:19 jpich: that would be fine 22:33:47 I was worried about prohibiting other contributors to patch there something... 22:34:12 jpich, :-) thanks for understanding me 22:34:14 There certainly will be a ramp up phase 22:34:38 lsmola: No worries... though I had to read twice too :o) 22:34:57 david-lyle, also itś not entirely easy to run development env for tuskar UI 22:35:07 Is the assumption that the core teams will be merged straight away? 22:35:18 david-lyle, also costs at least 12GB ram, preferably 16GB 22:35:20 I think that needs to be discussed as well 22:35:39 lsmola_: you may be doing it wrong ;) 22:35:56 david-lyle, we are running to many openstacks :-) 22:36:28 david-lyle, itś the minimal required env with simulatin 4 baremetals 22:36:28 david-lyle: Fair enough, re: fundamental change. I'm realising this as I read more :) (/me was thinking, well it's a new dashboard/tab right, just coming with a bit more code than usual) 22:36:48 another stupid question: does tusker work with devstack? 22:36:49 david-lyle, well core merge would be welcome 22:37:08 ohnoimdead: not a stupid question at all 22:37:13 ohnoimdead, no, tuskar is for deployment 22:37:31 ohnoimdead, so it also does what devstack do (tuskar+tripleo) 22:37:49 * mrunge is not using devstack at all.... 22:38:02 gotcha. makes sense. 22:38:02 It sounds like leveraging our new plugin mechanism is also an option, even if the project/repo moves under the horizon umbrella 22:38:29 ohnoimdead, think of openstack as an regular application(like wordpress), you will use openstack to deploy the application (meaning it will deploy itself :-)) 22:38:42 jpich +1 22:38:56 * ohnoimdead needs to go read a lot more about tusker+triple-o 22:39:49 ohnoimdead: you could use devstack to bring up a baremetal/ironic cloud 22:40:02 ohnoimdead: and run tuskar on that to deploy a KVM cloud separately 22:40:41 very cool. i'm a little behind on some of the new os projects. 22:40:58 new being within the last 6 months. :p 22:41:00 lifeless, I think I have tried similar crazines from start :-) but failed 22:41:12 david-lyle, let's make an action item, the tuskar guys should come up with a merge plan ? 22:41:16 lsmola_: yeah, *I* wouldn't try it, but in principle... 22:41:21 hehe 22:41:34 Is there anyone who wants to express a strong opinion against bring tuskar into the Horizon program? 22:41:35 lifeless, yeah I get you :-) 22:42:25 if not, I'd like to see a plan for the merging of code that is not a massive code drop 22:42:38 +1 to bringing it under the Horizon umbrella - although should we still wait for it to be integrated? Can a program have both incubated and integrated elements? 22:42:59 +1 for the merge in smaller pieces 22:42:59 jpich, good question 22:43:00 jpich: a program can 22:43:13 jpich: in fact programs don't have to have any incubated or integrated elements. 22:43:22 that's why I initially was thinking separate code bases 22:43:25 jpich: infra, devstack for instance. 22:43:28 mrunge, yes it should be perfectly doable, as the tuskar part will be hidden by default 22:43:36 Ok! 22:43:49 jpich: the integrated gate is specifically the set of things that get released every 6 months and their direct deps 22:43:55 ish 22:43:56 :) 22:45:08 ok, who wants to own the merge plan? 22:45:26 lifeless: Yes, ok. It just seems programs have had a general set of repos usually that are either all not integrated, or all incubated, or all integrated - to my knowledge 22:45:26 lsmola_ volunteered on the TripleO side already :) 22:45:42 david-lyle, i took it from tripleo side, so me? ;-) 22:45:51 congrats lsmola_ ! 22:45:53 * lifeless throws lsmola_ into the lion pit 22:45:59 thanks lsmola 22:46:07 david-lyle, I will put heads together with mrunge and rdopieralsky to figure this out 22:46:08 hahaha 22:46:17 lifeless, hehe 22:46:18 no, no no 22:46:20 bringing mrunge down with you 22:46:21 ;-) 22:46:36 lol :-) 22:47:15 let's see a plan, I'll work on a plan for core-reviewer merge, no-merge, 3 from a company, but only on the second Tuesday logistics 22:47:30 haha 22:48:13 So, let's just note for the record that we're accepting Tuskar-UI and we're working on the details 22:48:15 +1 to all that including 2ndtuesday 22:48:23 excellent 22:48:28 awesome 22:48:39 great news 22:48:45 Welcome, bring friends :) 22:48:52 :-D 22:48:53 (thumbsup) 22:48:55 and beer 22:48:57 friends who like to review!! 22:49:08 haha 22:49:16 #topic Ceilometer integration (lsmola) 22:49:17 lsmola started to get more folks in (teaching his son django already) 22:49:39 lsmola_ 22:49:41 jcoufal: his son is 3 weeks old, right? 22:49:43 his son can't even sit. 22:49:44 david-lyle, given we dont have much time, the agenda sums it pretty good 22:49:47 lol 22:50:05 lblanchard, please if you could look on 1. and .2, it would be great 22:50:22 lblanchard, we need to have some great wireframes for that :-) 22:50:22 so lsmola_ you do have a bp that depends on a ceilometer bp that is not owned by anyone 22:50:34 lsmola: will do 22:50:37 Is it also on the UX AskBot? 22:50:38 david-lyle, it is 22:50:44 https://blueprints.launchpad.net/horizon/+spec/ceilometer-api-enhancements 22:50:56 david-lyle, and it should land in I2, that shift this to I3 22:50:59 jpich: there is first version to review 22:51:04 * jpich admires the dependency tree 22:51:04 we need to get more focus on that 22:51:07 https://blueprints.launchpad.net/ceilometer/+spec/statistics-order-by-and-limit-for-grouped-query 22:51:08 Cool! 22:51:14 is the ceilo bp 22:51:28 david-lyle, this one expecially https://blueprints.launchpad.net/ceilometer/+spec/complex-filter-expressions-in-api-queries 22:51:40 david-lyle, that will bring big changes to ceilometer 22:51:55 so, do we need the other ceilo item? 22:51:57 david-lyle, they are willing to prioritize things we need in Horizon 22:52:31 confused, is one replacing the other? 22:52:38 david-lyle, yeah, but itś not a blocker I guess, we can have non-optimal version just with that one 22:53:17 david-lyle, the others are more about optimization of queries. With more data i will bomdard the server with queries 22:53:24 since the non-owned one is not likely to land, until someone actually owns it, will you split your bp to reflect the work for each 22:53:44 makes release planning a bit easier :) 22:53:44 lsmola: I envisioned that Admin users would see a similar Overview page as other users... 22:53:47 david-lyle, yes, I should change the deps a bit 22:53:57 lsmola_: thanks 22:54:01 david-lyle, things ae changing so fast :-) 22:54:15 david-lyle, ok, we can go to next 22:54:42 #topic update from I18N (amotoki) 22:55:06 amotoki can't attend, but the translation update for stable was a success, thanks everyone 22:55:25 excellent 22:55:25 lblanchard, yes, could be, though they might be interested in diferent stats. Risght now the overview pages differ 22:55:48 Most patches got merged a bit late but the stable-maint team has agreed it makes sense to have a freeze exception for translations in general, 3 days after the stable freeze 22:55:49 thanks to all who provided reviews 22:56:05 #topic Open Discussion 22:56:13 few questions here 22:56:24 This is enough for stable updates which have few strings, the process for new release tbd 22:56:24 do we start forcing people using angular? 22:56:50 MaxV, I just noticed the Jasmine target got removed from the agenda, apologies 22:56:56 MaxV, I guess only when they want to write javascript libs 22:56:58 s/target/topic 22:57:03 thiswas my second question 22:57:34 MaxV: force seems strong 22:57:54 i think force is reasonable, if the option is "large pile of jquery" vs angular 22:58:11 I think there is going to be a bit of time getting most of the existing coder base up to speed on angular 22:58:15 ie, dont let people write new interactionsw ith a bunch of jquery 22:58:24 new functionality makes sense, but fixes etc, not so much 22:58:31 ya 22:58:32 david-lyle: +1 22:58:38 agreed 22:58:39 lsmola: agreed…I will look into these both further and we can discuss more this week about designs! 22:58:40 exactly 22:59:00 lblanchard, excellent, thank you very much 22:59:19 We do have at least two whole views that are primarily javascript 22:59:26 MaxV: were you going to talk about jasmine too? 22:59:32 my second question was about jasmine 22:59:43 I'm not sure forcing a rewrite to augment the network topology screen makes sense either 22:59:52 eventually 23:00:07 to correctly test angular we need a javascript mocking library 23:00:13 at least 23:00:24 david-lyle, agreed, same for the heat topology 23:00:39 lsmola_: yes that was the other of the two 23:00:42 :) 23:00:56 :-) 23:01:09 I spent a little time looking into Jasmine, I didn't see a big issue with it 23:01:17 Did anyone have any concerns? 23:01:23 maybe we can keep qunit 23:01:30 for old things 23:01:37 jpich: did you do the qunit implementation? 23:01:39 and force jasmine on angular testing 23:01:54 +1 for jasmine. it's dope. 23:01:54 MaxV: id happily rewrite my tests in jasmine 23:02:00 david-lyle: It was already there when I came on board, I believe 23:02:16 jpich: ok 23:02:37 david-lyle, as far as I can remember, it's good old code, at least a year old, probably more 23:02:47 more 1.5 years now 23:02:52 :-) 23:03:12 qunit stuff has been around forever, since we started 23:03:16 ohnoimdead, might know it 23:03:20 I don't have a problem moving forward then. If we are going to have more javascript then we need to have better testing of it and since we're moving to angular, let's optimize for that 23:03:21 ha ;-) 23:03:34 I propose to keep qunit for old testing 23:03:43 I'd love to unify all that stuff 23:03:50 i'm less enthusiastic about angular, but that's a different topic. :p 23:03:54 +1 to have jasmine for angular 23:04:24 alright, let's get the Jasmine patch reviewed and let the testing progress 23:04:29 overtime 23:04:41 Thanks everyone! 23:04:41 Thanks everyone 23:04:46 david-lyle, thank you! 23:04:51 good to see you ohnoimdead 23:04:53 o/ 23:04:57 #endmeeting