14:01:23 <mnaser> #startmeeting ansible_sig
14:01:24 <openstack> Meeting started Fri Oct  4 14:01:23 2019 UTC and is due to finish in 60 minutes.  The chair is mnaser. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:28 <openstack> The meeting name has been set to 'ansible_sig'
14:01:38 <mnaser> anyone around? :)
14:01:42 <mnaser> i had a few things to dsiscuss
14:02:52 <mnaser> welp, perhaps someone will follow up on this later
14:03:26 <cloudnull> o/
14:03:35 <mnaser> i actually was hoping to discuss https://review.opendev.org/#/c/684740/1
14:03:47 <mnaser> i spoke a bit with mordred at ansiblefest and we thought now that everything is becoming a collection
14:04:08 <mnaser> and that we can host things whereever we want, it might be a good target for it to live htere
14:04:23 <mnaser> and then we can use gerrit and i think most people here are interested in making sure these modules work in some way or another
14:04:50 <cloudnull> thats cool, and a good idea imo
14:05:03 <mnaser> i personally cant stand the current github workflow
14:05:15 <mnaser> its quite unmotivating to do reviews tbh
14:05:54 <mnaser> i think mordred also pushed up an initial patch
14:06:04 <mnaser> https://review.opendev.org/#/q/project:openstack/ansible-collections-openstack
14:06:44 <mnaser> we might need some help in adding some jobs to that, i guesws
14:11:20 <mnaser> i think right now we need at last some base zuul jobs for collections
14:11:25 <mnaser> i think i can try helping with some of those
14:28:00 <cloudnull> sorry in double meetings :)
14:37:14 <mnaser> yeah, maybe we need to revisit the time of this
14:49:50 <cloudnull> I normally dont have a conflict at thisi time
14:50:06 <cloudnull> but ++ It would seem most folks are AFK :(
14:50:30 * mordred waves to mnaser and cloudnull
14:50:39 <mnaser> oh yay hi
14:50:40 <cloudnull> o/
14:50:51 * mordred is excited about our new home for ansible-collections-openstack
14:50:54 <mordred> needs to add jobs
14:51:12 <mnaser> yeah, i think jobs are the thing we need right now, but i dont think there's an established pattern of testing collections
14:51:16 <mnaser> so we may be in uncharted territory..
14:51:53 <mordred> yeah. on the other hand - we do have an test job in sdk for testing some ansible playbooks and stuff
14:52:01 <mordred> so I was thinking of starting with that and seeing how far we got
14:52:11 <mnaser> i guess we have to write something to 'build' the collection and then 'plop it in'
14:52:45 <mordred> yah - ansible-galaxy collection build makes a tarball, then you can ansible-galaxy collection install my-tarball.tgz
14:53:18 <mordred> so it shouldn't be *too* hard
14:53:23 <mnaser> famous last words
14:53:32 <mordred> might be nice to make a role in zuul-jobs for "install this collection in this directory over here"
14:53:43 <mordred> and a basic base job
14:54:24 <mnaser> yes thats kinda what i was hoping to be able to do, a base job for ansible collections
14:54:35 <mnaser> but yeah a role makes sense initially
14:54:50 <mnaser> like "get me an ansible and prepare it to test $this_dir"
14:55:13 <cloudnull> in tripleo-ansible we're using a lot of molecule based testing for our roles, plugins, and playbooks. There may be some bits we can crib from there?
14:57:22 <cloudnull> we're not using the tox-molecule, largely due to it requiring docker. the TripleO-Ansible base job is just a basic runner so just about any molecule driver will work.
14:57:41 <cloudnull> ** the tox-molecule job found in zuul
14:58:32 <mordred> nod. I haven't done much with molecule yet - largely because IIRC it seems to want to manage test ends for you still ... but I should probably learn about it at some poitn
15:07:19 <cloudnull> it seems to work pretty well these days.
15:08:40 <cloudnull> the delegated driver (local test) works really well in zuulci.
15:09:34 <cloudnull> the docker/podman drivers work well too, provide nice coverage for multiple Operating systems in a single test run, but are impractical for types of tests.
15:09:51 <cloudnull> ** for some types of tests
15:17:37 <mnaser> #endmeeting