15:00:36 <krtaylor> #startmeeting third-party 15:00:37 <openstack> Meeting started Wed Jan 7 15:00:36 2015 UTC and is due to finish in 60 minutes. The chair is krtaylor. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:38 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:41 <openstack> The meeting name has been set to 'third_party' 15:01:01 <krtaylor> anyone here for third-party working group meeting? 15:01:25 <nati_ueno> hi 15:01:34 <nuritv> Hi 15:01:55 <krtaylor> hi nati_ueno 15:02:00 <krtaylor> hi nuritv 15:02:05 <sweston> o/ 15:02:07 <nati_ueno> hi krtaylor 15:02:12 <krtaylor> hi sweston 15:02:22 <sweston> krtaylor: greetings! 15:02:28 <krtaylor> looks like asselin is lurking also 15:02:46 <krtaylor> happy new year everyone! 15:03:08 <sweston> happy new year, krtaylor 15:03:33 <krtaylor> well, so this meeting is the first of the new third-party CI working group 15:03:50 <krtaylor> we will meet every Wednesday at 1500 and 0400 alternating 15:04:09 <krtaylor> the focus will be different than the previous meetings 15:04:21 <omrim> Hello 15:05:06 <krtaylor> here we will focus on CI operators improving the consumability of the CI services 15:05:11 <krtaylor> hi omrim 15:05:56 <krtaylor> the agenda is extracted from the earlier meetings and is here: 15:06:10 <krtaylor> #link https://wiki.openstack.org/wiki/Meetings/ThirdParty#1.2F7.2F15_1500_UTC 15:06:52 <krtaylor> so, we started a bunch of things before the holidays, now it is time to get moving on them 15:07:09 <krtaylor> first up is the CI documentation 15:07:30 <krtaylor> there is an etherpad for organizing, but little movement 15:07:37 <krtaylor> the link is in the agenda 15:07:51 <krtaylor> I was going to push a virtual sprint 15:08:15 <krtaylor> but after yesterday's infra meeting, it looks like the infra cores we'd need are busy till Feb 15:09:04 <krtaylor> I'd like to see progress before that, so maybe we just do this one patch at a time until we get several things proposed 15:09:10 <krtaylor> any comments? 15:09:32 <krtaylor> omrim already has a patch proposed for adding a FAQ section 15:09:42 <krtaylor> with good reviews 15:09:56 <krtaylor> omrim, are you planning a refresh to that patch soon? 15:10:05 <omrim> krtaylor: Yes Sure! 15:10:32 <krtaylor> it looks like it is close, with a couple of minor fixes 15:10:37 <omrim> A new Patchset will be ready tommorow 15:10:44 <krtaylor> great! 15:11:36 <krtaylor> I did get permission to use any bits from jaypipes blogs as we need 15:12:04 <krtaylor> but we'll need some more workers to go through the sections and refresh/rewrite 15:12:50 <krtaylor> I've signed up for requirements, I plan on a patch for that this week 15:13:33 <krtaylor> any takers on other sections? 15:14:20 <sweston> krtaylor: I will check it out later this week, and let you know 15:14:29 <krtaylor> sweston, thanks 15:15:01 <asselin_> hi 15:15:04 <krtaylor> just put your nick next to the section you will work on so we don't conflict 15:15:11 <krtaylor> hi asselin_ 15:15:19 <sweston> ok, will do 15:15:37 <sweston> hi asselin_ 15:16:50 <asselin_> sorry I'm late. fyi, I don't have scrollback. 15:16:53 <krtaylor> I see a real advantage for a virtual sprint for this work, so maybe we can get that done for a final push for approval for the patches 15:17:27 <krtaylor> asselin_, np, it is being logged 15:17:40 <krtaylor> (hopefully) 15:18:03 <krtaylor> anyone have anything else for documentation? 15:18:34 <krtaylor> next up is the CI monitoring and dashboard work and spec 15:18:48 <sweston> I have an update 15:18:53 <krtaylor> the spec is getting a lot of reviews 15:19:01 <krtaylor> sweston, your floor 15:19:10 <sweston> krtaylor: thanks 15:19:16 <sweston> I have completed the new rest api layer in pecan/wsme, and also finished development on the angularjs webclient. 15:19:32 <sweston> The code is in the stackforge repository #link https://github.com/stackforge/radar 15:19:53 <krtaylor> yes, I just started to look through that 15:19:59 <sweston> and online live at #link http://dashboard.triniplex.com 15:20:39 <sweston> this week I'll get started on getting some interesting information in there, and will update the spec 15:21:05 <sweston> that's all I have for today 15:21:24 <krtaylor> sweston, jhesketh had some good comments about the spec 15:21:34 <krtaylor> about the scope 15:21:50 <sweston> yup, I saw them .. have been too swamped to respond yet 15:22:27 <krtaylor> I was somewhat in agreement, that maybe it grew too big too fast 15:22:32 <sweston> thanks everyone who commented, asselin_ had some good input as well 15:22:35 <krtaylor> what are you thinking there 15:23:15 <krtaylor> I think initially everyone was just looking for radar to be fixed, then enhanced 15:23:27 <sweston> I agree 15:23:44 <krtaylor> maybe it was overwhelming the features that were in the initial drop 15:24:33 <sweston> I think that querying the gerrit api is a good first goal 15:24:45 <krtaylor> agreed 15:24:48 <asselin_> agree 15:24:58 <krtaylor> do you think we shoud split the spec as jhesketh suggests? 15:25:29 <sweston> posting something in the review comments would be preferable, that way we create as little interference with the current process as possible 15:25:43 <sweston> I think it's a good idea 15:26:04 <sweston> best for this to be a phased approach 15:26:24 <sweston> we'll see what works, and what folks are comfortable with, and then iterate 15:26:27 <krtaylor> ok, maybe a phases division in the spec? 15:26:53 <sweston> yes, I think that is best 15:27:04 <krtaylor> ok, good 15:27:24 <krtaylor> so you'll have a new spec soon? 15:27:40 <krtaylor> no rushing you, just wondering 15:27:49 <sweston> yes, i'll try to get to it in the next few days 15:27:59 <krtaylor> great 15:29:00 <krtaylor> ok anything else on monitoring/dashboard? 15:29:11 <sweston> that's it from my side 15:30:00 <krtaylor> ok, next was the puppet modules split out work 15:30:16 <krtaylor> that was discussed yesterday at infra meeting 15:30:36 <asselin_> yes, we got approval to 15:31:07 <krtaylor> and a mini-sprint the end of the month right? 15:31:39 <asselin__> (got disconnected) 15:32:09 <krtaylor> ah, just was saying there is a mini-sprint planned 15:32:29 <asselin__> yes, very excited about that 15:32:42 <krtaylor> that will help get it done 15:33:01 <asselin__> sweston, I posted some comments on your script with fixes. did you get a chance to look at those? 15:33:17 <sweston> asselin__: not yet, will check them out 15:33:20 <asselin__> the auto-subtree script 15:33:41 <sweston> asselin__: thank you for the comments 15:34:03 <asselin__> comments include a few fixes that I tested on my system 15:34:30 <sweston> yup, will push a new patchset today 15:34:50 <asselin__> so please review https://review.openstack.org/#/c/140523/ 15:34:53 <asselin__> #link https://review.openstack.org/#/c/140523/ 15:35:02 <asselin__> so we can get that merged asap 15:35:19 <sweston> will do it asap 15:35:20 <asselin__> we can still do splits before the sprint 15:35:49 <asselin__> and I think that's a good idea to figure out any remaining issues 15:35:59 <sweston> agreed 15:36:01 <krtaylor> agreed 15:36:08 <asselin__> so the sprint day we can be smooth execution 15:37:19 <krtaylor> asselin__, what day is the sprint? was it the 30th? 15:37:28 <asselin__> Jan 28, wednesday 15:37:37 <rprakash_> Neutron SErvice VM? 15:38:07 <krtaylor> 28th, ok making note on my calendar 15:38:42 <asselin__> I think that's it, unless there are othe rquestions 15:39:23 * asselin__ added another agenta topic 15:39:39 <krtaylor> asselin__, I saw that you pushed a new patchset for in-tree ci 15:39:51 <asselin__> yes, that's the next topic :) 15:40:01 <krtaylor> hehheh, yes I see the refresh now 15:40:17 <krtaylor> I was going to mention it anyway, no worries 15:41:47 <asselin__> so I'm interested to hear what ppl think is the best way to refactor the puppet scripts to be reusable 15:42:22 <asselin__> and where they should go / how to structure it 15:45:18 <krtaylor> asselin__, this new module will change documentation, I am wondering if we push here first before documentation rewrite 15:45:43 <krtaylor> or just fix (or add a new section) with this in mind 15:46:00 <asselin__> I hesitate to delay documentation 15:46:15 <asselin__> I prefer to fix/add a new section 15:47:02 <krtaylor> agreed 15:47:02 <ja_> +1 to fix/add ... before too long I have to start standing up a new CI system from scratch 15:47:20 <sweston> agreed, as well 15:47:45 <asselin__> and it's not clear how long it will take 15:47:52 <krtaylor> asselin__, have you had any feedback from infra ci on refactoring 15:47:56 <ja_> ... maybe the refactoring itself can/should be staged/sprinted ala the earlier discussion 15:48:34 <asselin__> some discussions, but nothing concrete. 15:48:49 <asselin__> everyone supports the idea, just not clear the best solution. 15:49:19 <asselin__> so I'd like to get some ppl thinking about what the end result should look like and how to get there 15:49:42 <asselin__> there are some infra comments on the log server split out attempt 15:49:54 <asselin__> #link https://review.openstack.org/#/c/138913/ 15:49:54 <krtaylor> ok, we'll see how the reviews go, I very much like the idea 15:50:31 <asselin__> in that case I was refactoring everything and leaving it in openstack_project 15:51:09 <asselin__> another option is to put the refactored portions in the puppet-* projects directly 15:51:43 <asselin__> but i'm not sure what ppl think about that. i.e. have an openstack-zuul.pp inside puppet-zuul 15:52:29 <krtaylor> from some of the template ideas we kicked around, that was the direction from those discussions IIRC 15:52:38 <asselin__> my goal is to have thirdpartyci and openstackproject use the same puppet scripts 15:53:02 <krtaylor> +1 15:53:11 <sweston> asselin__: +100 15:53:12 <ja_> +1 15:54:16 <asselin__> the other approach I thought was to get a thirdpartyci solution created 15:54:38 <asselin__> then, from there, refactor out the puppet classes out of that and openstack_project. 15:55:08 <asselin__> the advantage is you get something working sooner, but not sure how easy it would be to execute 15:55:26 <sweston> that's interesting 15:55:39 <sweston> might work better than refactoring things as they are 15:55:44 <krtaylor> yes, and at some point, maybe a total fork is the right approach 15:55:59 <asselin__> but it would be clear what's needed 15:56:41 <ja_> how onerous is the creation (and, if goal is met, eventual deletion) of a new 'solution' ? 15:57:35 <krtaylor> it really is a question of getting critical mass on one effort, but that is why we are here 15:57:51 <ja_> ...although, given that there will always be a certain degree of upstream churn, maybe it's only creation. if its contents someday become empty, until the next Major Churn, that's not automatically bad. 15:58:50 <krtaylor> really good discussion, but we are almost out of time 15:59:09 <asselin_> hi sorry got disconnected again 15:59:29 <krtaylor> I wanted to remind everyone that next week will be Wednesday at 0400 UTC 15:59:39 <krtaylor> I will get the agenda updated 15:59:56 <krtaylor> that will be late Tues for those of us in the US 15:59:57 <asselin_> krtaylor, could you send reminders to the mailing list? 16:00:07 <krtaylor> yes, will do 16:00:09 <sweston> thanks everyone 16:00:15 <krtaylor> thanks all! 16:00:18 <asselin_> thanks 16:00:24 <krtaylor> #endmeeting