19:00:30 #startmeeting ansible community meeting 19:00:30 Meeting started Wed May 18 19:00:30 2016 UTC. The chair is rbergeron. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:30 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:00:30 The meeting name has been set to 'ansible_community_meeting' 19:00:48 #chair gregdek 19:00:48 Current chairs: gregdek rbergeron 19:00:50 we really need to space these appart, Wed == no code touched today 19:01:25 Your attendance is not compulsory to this one. :) 19:01:33 * gundalow waves 19:01:52 This one tends to be more "community housekeeping". 19:01:54 the good news is: we're doing open floor for this today :) 19:02:19 because also piles of things to do here as well :) 19:04:45 I have Travis update (Spoilor alert: We often have a crazy huge queue) 19:05:03 * MichaelBaydoun interest is perked 19:05:38 #topic travis update 19:05:43 #chair gundalow 19:05:43 Current chairs: gregdek gundalow rbergeron 19:05:49 take it away! :) 19:05:58 https://github.com/ansible/community/issues/47 19:06:03 See the last comment 19:06:28 erm, no 19:06:32 one moment :P 19:06:52 ah, yes Ididn't refresh 19:06:54 https://github.com/ansible/community/issues/47 19:07:18 https://docs.google.com/spreadsheets/d/1VzUPNN_KseqySj5CFRUjU1okgXQN6he57ZHrOwyQRGc/edit?usp=sharing is a graph of the last 15 days of Travis for Ansible 19:07:26 combined ansible, core & extras 19:07:43 Max: 245 queued builds 19:07:43 Min: 1 queued build 19:07:43 Average: 67.94 queued builds 19:07:43 Mode: 6 queued builds 19:07:44 Median: 44 queued builds 19:08:40 Which is a lot worse than I expected 19:09:10 But does explain why we often see huge delays in testing appearing 19:10:19 For ansible/docsite only changes we should maybe add in [ci skip] so Travis ignores https://docs.travis-ci.com/user/customizing-the-build/#Skipping-a-build 19:10:46 thoughts? 19:11:09 e thinks 19:11:28 * rbergeron ... that is 19:11:51 gregdek: any thoughts? 19:12:18 there are a few small things we can do, like putting more dependencies in docker, but doing larger scale stuff requires some actual planning (detect what files are been changed & only run applicable test) an resource allocating to it 19:12:37 and* resource allocating to it 19:12:55 This is the class of problem that we should be able to throw money at. 19:13:19 The fact that it isn't indicates a serious flaw in Travis's business model. :/ 19:13:36 Travis will accept money to change the queue size 19:13:52 "I'll let you review the data and please don't hesitate to send back new questions this might have generated. We could then talk about the paid options we can offer you if you decide to go forward with a paid plan. 19:13:56 " 19:14:15 Ahh, missed that bit. 19:14:23 it wasn't in the ticket 19:14:28 so you didn't :) 19:14:29 So I think we should escalate that internally. 19:14:34 Ah, then that's why I missed it, LOL 19:14:38 So. 19:14:54 Maybe we should assign that one to me to "solve the Travis resource problem" and I'll escalate internally? 19:15:18 gregdek: Should I contact them to find out if we were (and make it clear I'm not, nor can I) what paid options would make sense for our use cases 19:15:41 Yes, that would be great. Maybe make an intro to me? 19:15:44 they tried to sell me things before, but I said I need some hard data showing the problem. Which they've now provided 19:16:29 #action gundalow to contact TravisInc again and introdue gregdek and see if we were to go for paid options what would make sense 19:16:32 next? 19:16:34 thanks 19:18:15 Floor still open! 19:18:25 * gregdek is working on a deadline for other stuff right now... 19:19:27 #topic open floor 19:19:44 rbergeron: Anything from your goodself? 19:20:58 or anyone else? 19:21:21 MichaelBaydoun: ? 19:21:55 10... 19:23:00 3 19:23:01 2 19:23:01 1 19:23:03 0 19:23:05 #endmeeting