19:01:36 <jillr> #startmeeting Ansible Diversity Working Group | Agenda: https://github.com/ansible/community/issues/577
19:01:36 <zodbot> Meeting started Thu Jun 24 19:01:36 2021 UTC.
19:01:36 <zodbot> This meeting is logged and archived in a public location.
19:01:36 <zodbot> The chair is jillr. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:36 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:01:36 <zodbot> The meeting name has been set to 'ansible_diversity_working_group_|_agenda:_https://github.com/ansible/community/issues/577'
19:02:21 <jillr> #topic open floor
19:04:07 <felixfontein> o/
19:04:16 <jillr> #chair felixfontein
19:04:16 <zodbot> Current chairs: felixfontein jillr
19:04:25 <jillr> hola, I think it's just us today
19:04:43 <briantist> I'll join o/
19:04:51 <jillr> #chair briantist
19:04:51 <zodbot> Current chairs: briantist felixfontein jillr
19:04:53 <jillr> yay!
19:05:42 <jillr> we don't have any agenda items, does anyone have things they'd like to discuss?
19:06:27 <briantist> I don't have anything in particular
19:07:55 <thedoubl3j> o/
19:08:04 <jillr> #chair thedoubl3j
19:08:04 <zodbot> Current chairs: briantist felixfontein jillr thedoubl3j
19:08:08 <jillr> howdy!
19:08:36 <jillr> I'll keep it open for a few minutes in case anyone swings by with open floor items they'd like to discuss
19:08:45 <thedoubl3j> I have nothing in particular either.
19:09:19 <felixfontein> I actually do have a question :)
19:10:04 <felixfontein> I recently grepped for some words in some collections I maintain, and noticed that there are several modules in community.general which use terms like 'slave'
19:10:54 <jillr> I think that was an open todo when we made the chnages in ansible/ansible
19:10:57 <felixfontein> this are mainly the c.g.nmcli and c.g.redis modules. the naming comes from the upstream project, and at least for nmcli there seems to be no alternative
19:10:57 <jillr> *changes
19:11:28 <felixfontein> I'm wondering what's the best way to proceed here. create our own alternatives (and risk that the upstream project eventually chooses different ones)?
19:12:25 <jillr> do the upstream projects have open tickets/ongoing discussions?
19:12:53 <felixfontein> I haven't found anything for nmcli (though I also didn't search that long)
19:13:22 <briantist> wow that's tricky..
19:13:32 <felixfontein> for redis I just found https://redis.io/topics/replication, which looks to me like they are avoiding 'slave' (though keeping 'master')
19:14:10 <jillr> we could add reasonable aliases, so people have the ability to choose something else for their playbooks?
19:14:11 <felixfontein> (looks like they are using 'replica' instead of 'slave')
19:15:19 * samccann stumbles in late to the party
19:15:43 <jillr> #chair samccann
19:15:43 <zodbot> Current chairs: briantist felixfontein jillr samccann thedoubl3j
19:16:14 <samccann> as for master/slave etc - there is an entire project here internally at red hat on conscious language choices. I can look for what master/slave should be changed to
19:17:24 <briantist> samccann: the tricky thing is that it's content around some upstream project(s), which still use the problematic terms
19:17:31 <briantist> if I understood it correctly
19:17:34 <jillr> I think the concern is around doing so when the upstream hasn't, especially if they pick something else in the future
19:18:00 <samccann> yeah finally read back on the thread and see the difficulty
19:18:37 <felixfontein> hmm, actually NetworkManager might have some alternatives. after all it's a RedHat project apparently :) (just read that on WikiPedia)
19:20:08 <jillr> ah yeah maybe we can get some info on if that's happening  :)
19:23:01 <thedoubl3j> definitely and if not, maybe help start pushing the cart a bit to start the conversation around it possibly changing.
19:23:04 <jillr> samccann: are you going to those internal WG meetings?
19:23:36 <jillr> or do we know anyone else in Ansible that is, that could pass along questions?
19:27:28 <jillr> maybe we can pass that on to the internal Ansible D&I group and see if anyone else in the org is involved in that one
19:27:50 <samccann> I'm not in the internal meetings, no.
19:28:26 <samccann> But I am on the internal chat so I can ask questions. Is the primary question 'what do we do when an upstream project is still using questionable language'?
19:28:35 <jillr> samccann: yeah that would be great!
19:28:49 <jillr> otherwise it looks like the intranet says to send things to their mailing list
19:29:11 <jillr> or even just, hey is there a roadmap for nmcli to do this?
19:29:41 <samccann> ok I'll ask. I think the problem we have is equivalent to having this in a GUI or CLI command, right? (these won't be people familiar with collections etc so want to describe our situation clearly)
19:30:14 * jillr thinks..
19:30:49 <felixfontein> for nmcli, this is actually a CLI command we're using
19:31:07 <jillr> "There's an Ansible module for nmcli. Ansible is trying to eradice this language in our stuff and we see there isn't yet an alternative term in the NM project.  Ist hat something on the roadmap for that project that we can plan around?"
19:32:09 <jillr> presumably it would change in both the NM GUI and cli and we'd want to use the same term, and roughly at the same time (in a perfect world)
19:32:38 <samccann> what's the actual command felixfontein?
19:32:59 <felixfontein> actually there might be an alternative term, but it's not so easy to find if there is
19:33:33 <jillr> maybe the request is just for someone from that project to come talk to us?
19:34:02 <jillr> or we could go talk tot hem, either way. but having a contact from that project
19:34:38 <felixfontein> there are several network types (bond-slave, bridge-slave, team-slave) that use slave
19:35:32 <felixfontein> interestingly on https://developer.gnome.org/NetworkManager/stable/nmcli.html these are all marked as deprecated, but i'm not sure what "deprecated for ethernet with master" means
19:35:58 <samccann> yeah noticed that as well felixfontein... couldn't figure out what it meant
19:36:25 <jillr> that's one that goes pretty deep, it might even be the term used in the LAG RFCs
19:37:44 <jillr> actually rfc7424 and 7130 appear to use member
19:38:03 <jillr> the kernel docs use the other terms though
19:39:12 <jillr> so maybe it's a combined approach, reach out to the network manager project and ask, do y'all have plans to take this on?
19:39:26 <samccann> is that an upstream project?
19:39:38 <jillr> and consider adding aliases to our modules for primary member or some such
19:39:54 <felixfontein> samccann: nmcli is a part of network manager
19:39:56 <jillr> samccann: yes, but one we're either a significant sponsor of or the primary sponsor of (I forget which)
19:40:29 <jillr> it's a lot of tools that basically control all the network interfaces on a linux system
19:42:14 <jillr> step one either way is probably, "hey red hat representatives at $project, is there a plan?"
19:42:21 <jillr> then we can figure out what we want to do from there
19:43:24 <samccann> yeah if they plan on changing the terms, then we wait for that to propagate
19:43:56 <jillr> ok so officially it's a Gnome project, but we're heavily involved in it
19:46:27 <jillr> is there anything else we think we should cover on the topic before we know if there's any RH contacts involved in the project we can talk to?
19:47:08 <jillr> oh, felixfontein do you have any issues already opened for these?
19:47:14 <jillr> we should put this on the project board
19:47:15 <felixfontein> looks like redis first didn't want to change terms (2016), but then later (2018) they did rename at least slave
19:47:29 <felixfontein> no, not yet, though right now I'm working on a PR for redis
19:47:35 <jillr> thanks!
19:48:01 * jillr files one for nmcli
19:48:09 <felixfontein> cool, thanks!
19:52:32 <jillr> #link https://github.com/ansible-collections/community.general/issues/2866
19:53:39 <jillr> thanks samccann!
19:53:59 <jillr> oh and you can always just point them to that GH issue  :)
19:54:08 <felixfontein> #link https://github.com/ansible-collections/community.general/pull/2867
19:54:12 <felixfontein> for redis :)
19:54:28 * jillr reads
19:55:21 <jillr> felixfontein: how would you feel about deprecating slave in ansible?
19:55:41 <jillr> is it deprecated in redis, or just aliased?
19:56:18 <felixfontein> I think it's aliased; the value 'slave' is also still returned instead of 'replica' for backwards compatibility
19:56:24 <jillr> ah
19:56:55 <felixfontein> I thought about deprecating, but I would only do that at least one major release later, so there's a period where both names work without deprecation warnings
19:57:12 <jillr> that makes sense
19:57:42 <felixfontein> I'm not sure whether deprecating is a good idea if the upstream project doesn't deprecate the name
19:57:49 <jillr> yeah
19:58:09 <jillr> leave it deprecated forever to nudge people  ;)
19:58:13 <jillr> j/k
19:59:39 <jillr> anything else we want to cover today?
20:01:39 <jillr> #endmeeting