16:00:48 #startmeeting api-sig 16:00:49 Meeting started Thu Jan 23 16:00:48 2020 UTC and is due to finish in 60 minutes. The chair is sshnaidm. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:50 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:52 The meeting name has been set to 'api_sig' 16:00:58 heh, somehow I'm in three meetings at once :) 16:01:07 who is available? 16:01:17 me 16:01:21 if there are conflict we can change time meeting 16:01:33 I'd prefer to, to be honest 16:01:52 and since we're occupying the API SIG slot, we should probably set up a special meeting for our working group 16:01:58 ok, will create a poll in etherpad about a new time 16:02:02 an hour earlier later would work for me, for example 16:02:07 * or later 16:02:25 sshnaidm: you can start a doodle or similar 16:02:34 dtantsur, yeah, good idea 16:02:43 yeah, or we could agree on a SIG meeting time... when trying to get https://review.opendev.org/697278 done I couldn't find an appropriate set time for the SIG to set the schedule 16:03:14 odyssey4me: if we don't use one of the meeting channels, we can probably use any time that works for us 16:03:14 ok, we'll see in doodle which is best 16:03:46 dtantsur: yep, easy enough - I just wanted to make it easy to communicate and get updates 16:03:54 #action sshnaidm to choose better time for sig meeting 16:04:09 So, as you can know from openstack-discuss list the tranisition is in the progress. Tracking of it is in our etherpad: https://etherpad.openstack.org/p/openstack-ansible-modules 16:04:18 In these weeks I'd ask everybody for reviewes asap not to block peoples workflow, please. Take a look: https://review.opendev.org/#/c/703342/ 16:04:28 We have in CI the same job that was running before in Ansible PRs, so we're covered at least the same as before now. 16:04:44 went over all PRs in github for OS ansible modules and left a message about tranistion, but it can be still merges. We'll sync all things when modules will be removed from 2.10 completely. It will happen in next 3 weeks when ansible folks fix their sync of docs from collections. 16:04:44 Meanwhile I'll prepare these PRs in ansible github. 16:05:55 There are some issues with collection names and using them in roles/playbooks, so we'll need to mention the full collection names for now, "FQCN" 16:06:01 gtema, to your question ^ 16:06:54 yupp. But I think once modules are removed in ansible it is fine 16:07:08 gtema, in 2.10 only though 16:07:11 I think we also need to agree to accept new modules WITHOUT os_ prefix 16:07:44 gtema, yeah, so another requirement from ansible team was not renaming till 2.11 16:08:06 since it can cause problems with collection and their usage, which may change any time 16:08:17 also I will revive mordred's patch for implementing base Module class in the module_utils, it makes life easier (did this for my own collection same) 16:08:18 I think we can wait a little with renaming 16:08:43 gtema, yeah, just need to rebase it 16:08:51 sshnaidm: ansuble guys want that in our repo we do not rename modules? 16:08:59 i thought it was ok with BOTMETA 16:09:02 we are good now to develop new modules, we already have a few in reviews 16:09:26 gtema, it's not so easy, turns out, and things are changing fast there 16:10:00 crap, it's like with Openshift - things are changing and thus no release possible. What are you doing there at RedHat ;-) 16:10:11 just talked this week with bcoca, Paul, gundalow on ansible-devel and to be safe it's worth to wait with renaming 16:10:12 drinking 16:10:23 that's for sure 16:10:42 gtema, yeah, it's our main purpose here :D 16:10:51 ok, but then let's agree on accepting new modules without prefix 16:11:15 you drink, and I am dying 4th day already 16:11:18 gtema, I'm not sure if it worth it, when finally renaming we can do it in bulk 16:11:23 sounds sensible to me, I guess the ask for preventing renames for a while is to allow a deprecation period 16:11:56 if the deprecation notice is up, but new modules don't have the prefix - then we're good 16:11:59 gtema, and to have it with os_ is keeping consistency 16:12:20 ok, #agreed 16:12:46 so that all I have from ansible team for now 16:13:03 what about release jobs? 16:13:05 we're good to add more tests, more jobs 16:13:14 gtema, yeah, good point 16:13:16 I would be happy to start consuming collection from my projects 16:13:27 I stole a pushing job from Paul and working on it 16:13:55 need to ping mordred about pushing to galaxy, because he has the access only 16:14:22 maybe it's worth even to push what we have now and then to do it with releases and push-job 16:14:52 I'm only for early usage to see if everything is ok, etc 16:15:18 #action sshnaidm to add a publishing job 16:16:01 #action mordred to publish collection or to share access for galaxy namespace 16:16:05 yeah, the interesting issue I found using collection on galaxy - you can't update/delete pushes 16:16:17 so whatever we push - stay there with this name forever 16:16:22 gtema, yeah, you need a new version 16:16:35 or at least until you redhat guys drink another beer 16:17:00 I know about new ver, I mean before pushing you really need to be sure it works 16:17:19 gtema, you can submit issues for galaxy, btw, I submitted one about editing namespace and it was solved 16:17:27 SURPRISINGLY 16:18:06 oh yes, suprisingly. You also need to open issue to add maintainers to your org, ... 16:18:14 and lot's of things are "undocumented" 16:18:23 gtema, you can add them, I did 16:18:24 so my life is great 16:18:35 no need to open issue for that 16:19:06 I do not see any other candidate users to add, only some "preconfigured" roots or so 16:19:37 gtema, try on #ansible-galaxy 16:19:54 yeah, when I will have time for that 16:20:00 chouseknecht is the guy 16:20:15 thks 16:20:29 ok, is there something about new fancy ansible modules for openstack you wanna talk more? 16:20:52 3 16:20:54 2 16:20:56 1 16:20:58 nope, I would be offering one more for volume backups 16:21:10 gtema, great 16:21:10 I guess sshnaidm is developing ironic stuff? 16:21:22 (and we need to sync on that one day) 16:21:23 dtantsur, right, looking at node cleaning module now 16:21:34 dtantsur, yep, will ping you when have something 16:21:40 I wonder if we need a spec-like or RFE-like process from now on 16:21:55 dtantsur, and I need a way to test it and better w/o devstack 16:21:57 to have some formal discussion about interfaces rather than just landing whatever looks sane 16:22:08 sshnaidm: w/o devstack - tripleo undercloud? 16:22:20 dtantsur, even better just ironic standalone 16:22:26 bifrost? ;) 16:22:47 dtantsur, just container with ironic? like ironic-dev I had 16:23:05 installing ironic is easy, it's harder when you need it to actually DO anything 16:23:05 dtantsur, but I couldn't auth with it and current Openstack ansible auth 16:23:25 the current openstack ansible auth is broken for at least some cases 16:23:53 a VM with bifrost (we have vagrant support) can be more helpful 16:23:55 dtantsur, yeah, so maybe we can find a way, but if not - tripleo standalone job then :) 16:24:15 or that (although it won't allow you to actually do things with VMs) 16:24:19 dtantsur, well, I can't run vm 16:24:36 dtantsur, we don't have nested on upstream clouds 16:24:40 okay, we can discuss it outside of the meetings 16:24:44 I meant your laptop :) 16:24:46 dtantsur, cool 16:24:57 sshnaidm: and FWIW I do use devstack on a VM in RDO cloud. it works. 16:25:14 dtantsur, rdo cloud is different story :) 16:25:30 okay, we can discuss ironic stuffs on #openstack-ironic, I won't hold the meeting 16:25:33 ok, let's take it offline then 16:25:43 ok, if nothing else... 16:25:51 #endmeeting