11:03:26 #startmeeting Ansible hackathon 11:03:26 Meeting started Tue Jul 7 11:03:26 2020 UTC. 11:03:26 This meeting is logged and archived in a public location. 11:03:26 The chair is gundalow. Information about MeetBot at http://wiki.debian.org/MeetBot. 11:03:26 Useful Commands: #action #agreed #halp #info #idea #link #topic. 11:03:26 The meeting name has been set to 'ansible_hackathon' 11:03:58 #chair felixfontein abadger1999 gregdek acozine rbergeron 11:03:58 Current chairs: abadger1999 acozine felixfontein gregdek gundalow rbergeron 11:05:14 Anyone hacking on anything today? 11:05:44 #info 1400UTC Docs Hackathon in #ansible-docs 11:06:02 #topic Open Floor 11:06:31 Just starting the meeting so if we spot stuff we can do `#action must do foo` 11:10:51 persysted felixfontein: how are proxysql and digitalocean looking, need any help there? 11:13:08 I'll be "hacking" on the summit videos :P 11:14:00 If anyone is interested in a simple PR. we need `meta/runtime.yml` creating in https://github.com/ansible-collections/collection_template/ with these two lines https://github.com/ansible-collections/community.kubernetes/blob/master/meta/runtime.yml#L1-L2 11:14:02 cybette: woot 11:14:23 gundalow: for proxysql, I've done an initial commit (https://github.com/ansible-collections/community.proxysql) which gives myself and pondix something to build upon, and with felixfontein's help also submitted the corresponding ansible (https://github.com/ansible/ansible/pull/70490) and community.general (https://github.com/ansible-collections/community.general/pull/624) prs. 11:15:03 Great, I'll take a look of those now 11:15:18 cool ok ty 11:17:59 also I mentioned earlier it looks like there's outstanding issues on ansible-lint and molecule when it comes to role testing on 2.10+, so for now I've excluded the 2.10+ testing from the roles workflow 11:21:07 persysted: i am aware of linter issues with 2.10, but not molecule ones. any open bugs or better PRs related to it? 11:25:05 zbr: I _think_ it might relate to https://github.com/ansible-community/molecule/pull/2547, but wanted to do some more testing to confirm, there was this issue that looks to report the same behaviour: https://github.com/ansible-community/molecule/issues/2650 11:31:12 persysted: lets join effort on https://github.com/ansible-community/molecule/pull/2740 11:31:31 also good idea to get in sync on #ansible-molecule channel 11:35:51 #action gundalow merge https://github.com/ansible/ansible/pull/70497 11:36:36 zbr: cool ok sounds good 12:02:31 felixfontein: I found the configuration for Shippable nightly runs. Can you see https://app.shippable.com/subs/github/ansible-collections/dashboard/insights 12:02:31 0700UTC gh/ansible/ansible 12:02:31 0900UTC gh/ansible-collections/* 12:02:31 Working on https://github.com/ansible/ansible/issues/68132 should help this 12:13:56 persysted: have you thought about adding a changelog? f.ex. with https://github.com/ansible-community/antsibull-changelog/ (docs: https://github.com/ansible-community/antsibull-changelog/blob/main/docs/changelogs.rst) 12:14:41 gundalow: I only see stats there 12:15:26 gundalow: community.digitalocean needs the PR merged, I guess then we can do a first 0.1.0 release to get things going; I can do that later if you want, and if persysted and geerlingguy are fine with it 12:16:10 gundalow: persysted: an initial 0.1.0 release for proxysql wouldn't hurt either, then we can get the other PRs merged 12:16:35 I’m fine with it. I’ve been out sick all last night... seeing if I can do a little work today. 12:17:24 geerlingguy: hope you're better now! 12:17:52 geerlingguy: oh, that's no good :( 12:17:53 Just a headache this morning. We’ll see if Tylenol can knock that out 12:22:01 is it the 1st case of remote conference flu ? 12:23:16 afternoon all 12:23:17 I'd like to apologise for the lukewarm coffee that was served yesterday 12:23:40 * gwmngilfen is going to look at that bubbleplot idea from geerlingguy today 12:24:24 gundalow: how do we get at the action logs from yesterday? 12:24:25 gwmngilfen: which idea was that? I do love the bubblepolot 12:24:38 i mean, i can search my chat history, but if there's an easier way ... 12:24:54 gwmngilfen: https://meetbot.fedoraproject.org/ansible-community/2020-07-06/ansible_contributors_summit.2020-07-06-10.57.html 12:24:55 https://meetbot.fedoraproject.org/ansible-community/2020-07-06/ansible_contributors_summit.2020-07-06-10.57.log.html 12:25:26 sweet, ty 12:25:34 ah yes 12:25:43 nps 12:26:05 gundalow: see geerlingguy at 15:18 for what we're talking about, basically activity in modules 12:26:51 geerlingguy: let me try this out for community.general to start, as the code is set up for one repo, and if it looks good I'll generalise it :) 12:27:22 👍 12:27:37 felixfontein: I'll look to add a changelog & initial release, that's cool about the community.digitalocean work, I can start taking a look at that this afternoon. 12:28:47 persysted: felixfontein I've merged runtime.yml changes into ansible/ansible:{devel+stable-2.10} 12:28:47 felixfontein is https://github.com/ansible-collections/community.general/pull/624 good now? 12:29:35 https://github.com/ansible-collections/community.proxysql/releases/tag/0.1.0 12:30:06 cool thanks gundalow! 12:32:59 persysted: That should appear in Galaxy soon, I can see a `release` job running (middle column) on https://dashboard.zuul.ansible.com/t/ansible/status 12:33:20 gundalow: it is, but it should only get merged once the collection is published 12:35:34 felixfontein: persysted it lives https://galaxy.ansible.com/community/proxysql 12:41:47 gundalow: in that case, feel free to merge :) 12:42:01 done \o/ 12:43:01 felixfontein: I'm makig a backport for DO runtime.yml now https://github.com/ansible/ansible/pull/70483 13:13:34 gundalow: thanks! 13:40:26 awesome thanks gundalow! 13:47:57 gundalow: persysted: geerlingguy: I'll merge the PR for community.digitalocean and will try to create a first published version 13:49:16 Ace, thanks 13:49:39 awesome 13:50:40 * gregdek hullos 13:50:53 Always nice when irccloud decides to log you out and doesn't tell you, lol 13:53:34 gundalow: persysted: zuul picked up the tag and is (hopefully) building and publishing the collection: https://dashboard.zuul.ansible.com/t/ansible/stream/f45a3c1d0e2c473db7a0a3e4adbbcf9b?logfile=console.log 13:58:29 gundalow: persysted: geerlingguy: https://galaxy.ansible.com/community/digitalocean 0.1.0 is published! 13:58:46 gundalow: I rebased https://github.com/ansible-collections/community.general/pull/622 feel free to merge once tests pass 13:58:57 Nice! 14:00:54 w00t!! 14:03:13 * acozine waves 14:03:20 * resmo waves 14:03:30 * gwmngilfen joins in 14:03:34 hello 14:03:43 hey :) 14:03:47 Hey, are we docsing? 14:03:52 yes! 14:04:00 good to see things get sorted out and community.general is getting smaller 14:04:04 as long as we aren't doxing, that's fine 14:04:16 DING DING DING #ansible-docs time 14:04:20 felixfontein: no doxing, nope 14:04:46 oh, oh, yikes 14:05:03 yeah, i won't use that combination of letters again 14:05:25 Anyway. I stand ready to be led! :) 14:06:21 anyone who wants to play with or work on docs should join us in the ansible-docs channel 14:06:34 gregdek: `/join #ansible-docs` 14:30:27 DING DING DING We are talking about docs and guidance needed before we can close ansible/ansible issues & PRs. If this interests you please join us in #ansible-docs 14:35:53 it's off topic, but just out of interest, if I'm creating some new roles (where there's not currently any related plugins), is there still any benefit to creating them in a collection? or if it's just a set of roles is using a collection just overkill? 14:42:37 IMO using a collection is better, since it works better with other collections (f.ex. when you depend on modules/plugins from collections) 14:42:48 geerlingguy: ^ you might have better perspective on that than i do 14:43:00 and you can have multiple roles in one collection :) 14:44:20 Role-only collections are annoying in that the folder depth and reference (requiring FQCN) is super annoying. 14:44:44 But if you want to have a “bundle” of roles for redistribution it’s definitely the best solution 14:45:57 For roles local to one playbook/project I don’t think I’ll use collections, but if I have any groups of related roles I may explore using collections at some point, assuming the “migrate a role to a collection on Galaxy” issue is solved 14:47:39 geerlinguy: "that the folder depth and reference (requiring FQCN) is super annoying" out of interest, do you think that still applies to any role in a collection..? (is that something that would ideally be made more user friendly generally) 14:49:12 for instance if I have a role that's not dependent on plugins in the collection, but _could_ be bundled given it's similar content, would it still be more user friendly to upload that outside of the collection..? 14:51:19 I'll have a play and form some opinions of my own (I'm still pretty new to collections), but am interested in yours and others perspectives.. 14:58:20 hi, i'm a bit late. sorry. a quick question: what are you going to do, to close all collection related issues/prs in ansible/ansible? will they be open in correspoinding collections (i.e. moved) or just closed in ansible-base and people should copy them manually if they are interested? 15:00:28 sorry, mixed up with ansible-docs:) 16:22:55 gundalow: Regarding vmware_rest; I vaguely recalled talking to akasurde before.... Dug up this commit message: https://github.com/ansible-community/ansible-build-data/commit/dc6094a668f5a5676e82f542fcda64c3ce072a9e 16:23:13 gundalow: So I think we should just remove vmware_rest from acd.in for 2.10. 16:23:23 gundalow: Is that okay with you? 16:31:02 abadger1999: hum, I think it's in stable-2.10's runtime.yml. Maybe that's a mistake 16:31:57 gundalow: Yeah... a by product of the migration script but the vmware team don't want that/aren't ready for it for 2.10? 16:32:16 * gundalow -> food 17:00:56 gundalow, Do you know someone from DigitalOcean who can help me out with developer account ? 17:06:27 persysted: Hi, I've given akasurde (part of the team here at Ansible) admin on community.digitalocean repo. He contributed a lot to the original modules 17:07:10 cool ok, sounds good 19:19:28 Is there any module akasurde hasn’t made better?? 19:20:56 Also sorry for non-responsiveness today. I have some sort of bug... just got up the strength to check email and such. Hopefully whatever it is clears up today 19:21:24 geerlingguy: I don't have it to hand anymore, though when we were deciding collection structure we looked to who had contributed to 2+ directories under lib/ansible/modules, akasurde was in the top 5 19:21:44 akasurde++ 08:14:47 out of interest, what's the best way to interact with the mysql working group, to look to get the modules moved out of community.general? 08:16:57 probably create an issue and ping them 08:17:17 cool ok 08:17:26 if you're lucky and they are responsive enough, there's still a chance to do it this week 09:08:06 https://github.com/ansible/ansible/issues/70517 09:10:20 gundalow: would you be ok to create community.mysql? I'm happy to go ahead and prepare the pr's, then hopefully we can get a +1 before Friday and pull the trigger to migrate 13:04:59 * gundalow waves 13:06:03 Hey gundalow, where can I track the progress on getting Ansibot onto the AWS repos? 13:06:16 (collection repos) 13:08:14 tremble: https://github.com/ansible-collections/overview/issues/81 13:08:33 tremble: last time I looked I got a load of backtraces when I tried to run the bot 13:09:13 Ok 13:09:28 (updated with list of repos) 13:09:38 I guess it's somewhere on the back-burner right now? 13:12:54 tremble: Aye, it's high priority with everything else 13:13:12 * tremble laughs, know that feeling too well 13:32:26 action from Monday, claiming `builtin` in Galaxy has been done https://github.com/ansible/galaxy/issues/2432 13:37:43 hi gundalow! 13:40:25 felixfontein: Hi :) 13:48:02 @gundalow: while we fix things on galaxy, Make it possible to delete deprecated collections after a period of time? 13:48:22 immediately would be also okay for me 13:58:03 gundalow: 2 questions on bubbleplots of collections. I'm starting with just community.general for now. 13:58:35 1) are we just interested in plugins/modules, or the whole of plugins/ ? 13:58:50 2) ) i notice there's a lot of symlinks in plugins/modules to deeper in the tree - should I ignore these? 14:04:09 resmo: not currently possible to delete a deprecated collection. Though maybe raise an issue under gh/ansible/galaxy 14:04:38 gwmngilfen: is this bubbleplot of contributors? 14:05:34 activity is the goal - that could be unique contributors, file mtime, etc 14:06:19 gwmngilfen: I think to start with plugins/modules is OK 14:06:31 right 14:06:40 * gundalow -> afk 14:07:42 * gwmngilfen still wonders about the symlinks 14:24:50 a coworker informed about ldap_search module (new in community.general v0.2.0) seems not to be found (builtin path) in ansible-2.10.0a2, trying to reproduce 15:01:49 resmo: so `community.general.ldap_search` did not work? 15:16:55 felixfontein: just tested: `ansible -m ldap_search localhost` fails with "The module ldap_search was not found in configured module paths" but `ansible -m community.general.ldap_search localhost` is found 15:17:33 resmo: no surprise, since ldap_search wasn't in ansible/ansible before the split 15:18:30 that means every new module in community.general must be used with FQCN? 15:19:41 resmo: any new plugin in any collection 15:20:30 only plugins that existed in 2.9 will have 'short name compat' built in 15:21:17 so users have to know when modules were added... 15:21:26 yes 15:21:31 ansible-doc -l should show 'usable names' 15:21:38 they are mentioned in the changelog, though :) 15:21:38 playbooks and roles can define a list of `collections` to search if a short name is used 15:21:55 (for modules) 15:22:00 'collections' keyword can help, but its just another way of doing fqcn 15:22:22 resmo: why i thought those collections should be 'static' and new content go elsewhere .. but that ship sailed 15:22:49 I see, I guess this is or will be written in the migration to 2.10. (what's new for users) 15:24:59 maybe it makes sense to reorganize the ACD changelog to collect all new modules/plugins of all included collections in one place, instead of listing the in the section for every collection 15:27:54 felixfontein: currently the changelog for new modules uses plugin_names (not fqcn), probably worth considering to use fqcn for new plugins because users have to use fqcn 15:28:34 resmo: for the ACD changelog, definitely. for collection changelogs, I guess this could be an option; not all collections might want to use that 15:28:45 resmo: feel free to create an issue for that ;) 15:29:45 the fqcn should be implicit in collections .. shortname should really only be there for backwards compat 15:34:08 ::@:qq 15:34:08 :qq 15:34:10 :qq 15:34:13 qq:qqqq!! 15:34:16 :QQ 15:34:18 :qq 15:34:20 :qq!! 15:35:03 rruunnttii 15:35:07 ##:qq!! 15:35:08 ccdd ....//.... 15:35:10 q!! 15:35:30 :q! 15:35:45 child or cat? :D 15:35:55 child chasing cat ! 15:36:09 (or laptop on laptop? a colleague of mine managed to do that this morning...) 15:36:16 vviimm 15:36:20 :sseett nnoosspp 15:36:26 ://mmoodduullee__uuttiillss 15:37:05 ok, vim beats both child and cat... 15:37:22 with a mode stick 15:37:44 :) 15:37:51 quitting vim probably. 15:38:05 i bet on computer freeze 15:38:17 oh, ffs 15:38:18 sorry 15:38:23 let's laugh at him later 15:38:28 aw 15:38:49 never leave for later that which you can do now! 15:38:53 ;-p 15:38:53 apparently I have hit something on my keyboard so keystrokes are being duplicated 15:39:12 https://www.youtube.com/watch?v=eOifa1WrOnQ 15:39:12 probably 'sticky keys' or some other function like that 15:39:45 gundalow: it's ok, it was somewhat amusing to watch, and you didn't seem to paste passwords or other sensitive stuff :) 15:40:34 felixfontein: maybe he tried to quit his browser thinking he was in vim. and now his facebook password is : ":qqqqqqqqqqqq!" 15:43:14 :D 15:44:34 (at least no passwords) ;) 15:45:48 gundalow - what are your thoughts on migrating mysql from community.general to community.mysql? I created this issue to notify the working group: https://github.com/ansible/ansible/issues/70517, and created some pre-emptive pr's https://github.com/ansible/ansible/pull/70523 & https://github.com/ansible-collections/community.general/pull/633.. ..at this point, if you're happy to create community.mysql, I'd be fine to go ahead to migrate 15:45:48 everything across, and to confirm I'm happy to join the existing folks in looking to manage this going forward. 16:06:58 persysted I'll let gundalow make the call on migrating the mysql modules, but thanks for stepping up and offering! 16:07:52 * abadger1999 back online after irccloud disconnected all night. 16:40:58 abadger1999 welcome back, and join the club 16:41:24 a bunch of folks ran into the IRCcloud outage 16:50:44 Here's the playlist for Contributor Summit videos: https://www.youtube.com/playlist?list=PL0FmYCf7ocrbmgUsXJFqJrIKQRg9hz02h 17:34:33 persysted: wrt community.mysql: Once of the advantages of moving this to `community.mysql` is we can have every PR run against a set of different MySQL server versions. 17:35:24 yep would be cool 17:39:26 persysted: out of interested, what prompted you to want to split out `community.mysql` 17:39:34 Do we have agreement, if so I'll create the repo now 17:45:36 so, now I'm back :) 17:45:55 I love the concept of dedicated collections - I think they can more aligned to the working groups and will lower the bar when it comes to new folks wanting to contribute. I like also that we can use generic tooling like github actions, rather than have to learn custom tools and processes. I also think it helps wrt role discovery, I think one minefield can be the wide array of roles available, and typically I see folks just gravitate to 17:45:55 wards the most downloaded - including a set of community backed roles that folks can get behind might help with that. 17:48:54 oh, i guess I should #endmeeting before the proper Community Meeting starts in 15 17:50:15 I think we have agreement from those that have responded, if you want to create the repo, I'm happy to start migrating content across - we can always leave it another day before we merge the prs for any others to respond 17:51:27 there has been no response from the WG leader though 17:52:19 but it is probably still ok 17:54:09 It lives https://github.com/ansible-collections/community.mysql 17:56:01 Awesome ty! 17:58:20 DING DING DING community meeting in #ansible-community IN TWO MINUTES! 17:59:26 #endmeeting