15:10:58 #startmeeting ansbile core irc public meeting 15:10:58 Meeting started Thu Apr 4 15:10:58 2019 UTC. 15:10:58 This meeting is logged and archived in a public location. 15:10:58 The chair is bcoca. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:10:58 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:10:58 The meeting name has been set to 'ansbile_core_irc_public_meeting' 15:11:05 woot 15:12:05 #topic https://github.com/ansible/ansible/issues/54131 15:12:14 +1 to document or napalm ini format ... 15:13:21 hey, sorry 15:13:34 +1 to document 15:13:43 yeah, just +1 to document 15:13:52 +2 to have sivel document it 15:13:56 ;-p 15:13:58 +3 for bcoca 15:14:00 ;) 15:14:14 Is this the IRC equivalent of hands-on-the-bat game? 15:14:21 I'll just add the decision to the issue for the time being 15:14:35 open docs issue 15:14:50 or remake as docs issue .. docs team might have time to do it instead of waiting for us 15:14:51 I can convert the existing issue 15:14:54 yeah :) 15:15:09 Ok, that was only 3, but I think that is sufficient for this 15:15:57 yep, i would not even have put this to vote, 'just do it' 15:16:37 I'd be fine with a toggle. But it's definitely `waiting_on_contributor' 15:17:19 not that sure, would have to expand gather_ keywords at play also 15:17:44 ah, module_defaults would work, so ok, as long as no new gather+play keywords 15:17:55 * gundalow waves 15:17:56 #topic open floor 15:18:34 https://github.com/ansible/proposals/issues/160 same sub-group name in group of groups 15:18:59 New proposal that could do with an initial "This makes sense"/ "Try X instead" 15:19:51 iirc responded to that same 'feature idea' in ansible/ansbile 15:20:16 .hello2 15:20:17 maxamillion: maxamillion 'Adam Miller' 15:20:21 its 'ansible is not puppet' 15:20:35 we look at hosts as first class citizens, not class/group 15:20:38 And making such a change would cause backwards compat issues 15:20:44 both focuses are valid, jsut different 15:20:53 and not to mention, that it is just an implementation detail of the YAML inventory format that confuses people 15:20:59 and a fundamental change to ansible that makes it incompatible with existing setups 15:21:04 I am -100 15:21:47 sivel: we had this request before YAML, every user that came from 'group centric' systems to our 'host/task' centric goes through same thing, have many tickets with ini file or json script output as examples 15:22:29 yeah, I think it just those using YAML misunderstand. No other inventory formats support that same sub group format 15:22:32 'constructed' gives them a way to do this by dyamic groups/variables that can be based on existing intersections 15:22:37 so does group_by 15:23:03 sivel: well, they do, it jsut doesnt LOOK as hier as yaml file 15:23:10 plenty of ini examples 15:23:43 Isn't the way to handle this by using a host pattern? `europe:&webservers` 15:24:03 at one point we removed YAML format to avoid this issue, but it kept being reported with ini, so i decided to add back yaml since confusion was there anyways 15:24:08 sdoran: yes 15:24:22 sdoran: excpet vars are not 'intersection aware' 15:24:36 and they expect hosts to 'only have vars depending on how I access them' 15:24:46 I don't like the idea of group hierarchy. Seems overly complicated. 15:24:49 so vars defined in dc2 group applygin to host that is still there 15:24:54 But I am biased as a long-time Ansible user. 15:24:59 it is, but people are used to it 15:25:06 they like 'familiar' 15:25:09 In any case, I don't think we ahve enough concensus necessarily to decide on a proposal 15:25:17 our quorum is too small today 15:25:20 I'm -1 to it for what it's worth 15:25:22 part of ansibe's simplicity is that 'host is always the same' no matter how you access it 15:25:30 -1 also 15:26:02 -1 on this 15:26:05 so we do have -4/nil 15:26:16 or maybe NaN 15:26:22 -4/0 15:26:34 if someone want's to use Hiera or Jerakia with a lookup plugin, then more power to them but this seems like a fundamental change to how Ansible handles inventory and I think it's an anti pattern 15:26:37 we are not romans, we have the concept of 'nothing' 15:26:45 lol 15:27:09 * bcoca side swipe to 2k year old math concepts ... 15:27:21 -1 15:27:39 -5/0 ... giving it another 3 mins, then closing and rejecting 15:28:48 Ok, well 5 is my quorum requirement 15:28:56 so we are good then :) 15:29:43 NEEEEXT! 15:29:49 #topic open floor 15:30:37 Who would like to put some words on #160 and close it? 15:31:21 phrasing 15:31:41 'who would like' vs 'who's lap can i drop this in' 15:32:15 if no one else, i'll add to my queue, ive already close 20+ versions of this same request 15:32:28 thanks 15:32:40 I'll take it 15:32:46 k, if no new busyness, i'll end meeting in 5 15:32:50 gundalow: I'll write something up and have you review 15:33:13 maxamillion++ 15:33:13 gundalow: Karma for maxamillion changed to 6 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:33:14 Thanks :) 15:33:16 maxamillion: can probably just copy paste/ref some of the tickets i already closed with same 15:34:14 bcoca: you know which ones those are? 15:34:29 woo! zodbot karma 15:34:32 no, dont keep those ticket numbers in cache 15:34:45 bcoca: no worries, will poke around 15:35:53 check sivel's ticket page, i think he has 'bcoca being jackass' category 15:36:23 lol 15:36:25 https://ansible.sivel.net/pr/ <= just pretty sure its hidden from my ip 15:37:12 I remember some discription I really liked, about a person not being different because of the color of coat 15:37:20 #endmeeting