19:00:11 <gundalow> #startmeeting Ansible Molecule Working Group
19:00:11 <zodbot> Meeting started Wed Mar  6 19:00:11 2019 UTC.
19:00:11 <zodbot> This meeting is logged and archived in a public location.
19:00:11 <zodbot> The chair is gundalow. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:11 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:00:11 <zodbot> The meeting name has been set to 'ansible_molecule_working_group'
19:00:34 <gundalow> #chair themroc pabelanger zbr decentral1se fabianvf
19:00:35 <zodbot> Current chairs: decentral1se fabianvf gundalow pabelanger themroc zbr
19:00:37 * alongchamps here but also in another meeting
19:00:56 <gundalow> Who else is around?
19:01:21 * tima waves
19:01:34 <pabelanger> hello
19:01:50 <decentral1se> hey all
19:02:08 <zbr> hi!
19:02:11 <gundalow> So on the agenda I think we have
19:02:11 <gundalow> 1) Are we ready to call feature freeze on v2.20 and get the pre-release done and advertised
19:02:12 <gundalow> 2) Review other items on v2.20 board, are we pushing them to v2.21? (I guess this is part of (1))
19:02:12 <gundalow> 3) Zuul
19:02:14 <gundalow> What else?
19:02:22 <gundalow> #chair alongchamps tima
19:02:22 <zodbot> Current chairs: alongchamps decentral1se fabianvf gundalow pabelanger themroc tima zbr
19:02:45 <decentral1se> yep, perfect. One more: decision on https://github.com/ansible/molecule/issues/1779 for the change log issue from me
19:03:08 <tima> Review of the ansible verifier PR by @fabianvf?
19:03:19 * tima looks up PR number
19:03:39 <gundalow> OK, lets start with 2.20
19:03:46 <themroc> https://github.com/ansible/molecule/projects/2
19:03:48 <gundalow> #topic Molecule v2.20 pre release
19:03:55 <gundalow> #info https://github.com/ansible/molecule/projects/2
19:04:04 <tima> https://github.com/ansible/molecule/pull/1714
19:04:12 <tima> found it ^^^
19:04:16 <gundalow> #info Q: Has everything we care about been merged (apart from changelog)
19:04:30 <themroc> i believe so
19:04:32 <decentral1se> https://github.com/ansible/molecule/pull/1745 has been agreed to go in since a few WGs ago
19:04:37 <gundalow> #info Q: Are we now ready to cut and announce the pre-release so we can get some real feedback
19:04:48 <decentral1se> and is the only other one we have on the v2.20 train
19:05:29 <zbr> yep, in addition to the 1714 which i hope to see merged today, only the changelog is scheduled for 2.20
19:05:56 <decentral1se> +1 for annoucement and feedback after #1714 is decided on
19:06:10 <decentral1se> can people triage now or what is the feeling?
19:07:15 <gundalow> Q: Could we add`Add Ansible verifier #1714` after pre-release is cut?
19:07:33 <gundalow> Trying to get the balance between getting the release out and feedback vs keep on merging stuff
19:07:49 <themroc> should we add Mitigate against PyYaml CVE-2017-18342? https://github.com/ansible/molecule/pull/1808/files
19:07:49 <gundalow> fabianvf: what are your thoughts?
19:07:58 <decentral1se> yeah, so Ansible verifier was reviewed a bunch and was put on the 2.21 schedule
19:08:07 <tima> i'm fine with 1714 coming later for the reasons you list @gundalow.
19:08:11 <decentral1se> but that is totally up for change
19:08:37 <decentral1se> themroc: #1808 looks fine to leave since we don't have any dodgy `load` calls
19:08:47 <themroc> ok
19:09:08 <fabianvf> I'm here now
19:09:30 <decentral1se> AFAIK, there is a design issue remaining with Ansible verifier
19:09:33 <decentral1se> no?
19:09:39 <decentral1se> https://github.com/ansible/molecule/pull/1714#discussion_r262619088
19:09:45 <gundalow> fabianvf: we OK to push Add Ansible verifier #1714 into v2.21 (we are working on v2.20 at the moment)
19:10:06 <fabianvf> Yes, I think it makes sense to target it for 2.21
19:10:24 <gundalow> #agreed Ansible verifier #1714 for v2.21
19:10:45 <decentral1se> cool. I think we'll make it since we've already got many eye balls on it :)
19:10:57 <decentral1se> and of course, fabianvf is doing good things :)
19:10:58 <fabianvf> Assuming we get our release cadence a little more cadence-y :P
19:11:53 <decentral1se> so, just #1714 then? Are there reservations?
19:12:27 <themroc> fabianvf, +1
19:12:34 <decentral1se> sorry, #1745 I meant ...
19:12:41 <decentral1se> https://github.com/ansible/molecule/pull/1745
19:14:06 <fabianvf> makes sense to me
19:15:56 <decentral1se> +1 from me, as per last comment
19:16:07 <themroc> +1 also
19:16:23 <gundalow> is that +1 for it going in v2.20 or +1 to merge?
19:16:39 <decentral1se> merge
19:17:21 <themroc> both :)
19:18:23 <gundalow> :)
19:18:32 <decentral1se> See final agreement in https://meetbot-raw.fedoraproject.org/teams/ansible_molecule_working_group/ansible_molecule_working_group.2019-02-20-19.01.html
19:19:04 <decentral1se> but there is of course still space for concerns!
19:19:07 <gundalow> OK, I see +1's and green CI. Should I hit merge?
19:19:32 <decentral1se> 💣
19:19:47 <decentral1se> zbr: your head on the chopping block for fixes ;)
19:20:03 <gundalow> MERGED
19:20:04 <zbr> decentral1se: sure
19:20:22 <gundalow> https://github.com/ansible/molecule/projects/2 now only shows release notes, so I guess we are good
19:20:28 <decentral1se> woot woot
19:20:56 <gundalow> bah, why is Travis Red https://travis-ci.com/ansible/molecule/builds
19:21:28 <decentral1se> It's just https://github.com/ansible/molecule/issues/1799
19:21:36 <decentral1se> I've been babysitting the builds against this flakyness ...
19:22:14 <gundalow> ah, thank you :)
19:22:24 <decentral1se> sure thing
19:22:42 <gundalow> So should I tag pre release now?
19:22:53 <themroc> \o/
19:22:59 <zbr> gundalow: yeah.
19:23:05 <decentral1se> yes please
19:23:21 <zbr> we can take case of release notes after.
19:23:42 <zbr> i am curious about the next agenda items, like zuul
19:23:52 <gundalow> #topic Zuul
19:23:57 <themroc> release notes ->  https://github.com/ansible/molecule/issues/1779
19:23:58 <gundalow> pabelanger: all yours
19:24:06 <pabelanger> hello
19:24:32 <pabelanger> I'm mostly here to help answer questions about zuul CI and why we'd want to use it for molecule
19:24:39 <decentral1se> #info https://github.com/ansible/molecule/issues/1770
19:24:49 <decentral1se> #info https://github.com/ansible/molecule/pull/1773
19:24:54 <pabelanger> this was all in context to the idea of moving a way from travisci
19:25:17 <zbr> with travis feature under serious questions and considering that I am zuuling every day for openstack, I am in favour.
19:25:35 <pabelanger> but, bascially, we in ansible-network have access to a zuul, and are in the process of now buiding out own zuul for ansible org, and happy to help see how to onboard molecule to it
19:25:53 <zbr> questions: is zuul-github integration usable as I suspect gerrit would be too much for most people ;)
19:26:06 <gundalow> zbr: done and has been working for a while
19:26:09 <decentral1se> What is the relationship of ansible-network with ansible? What resources are available?
19:26:27 <pabelanger> yes, github works with zuul, we've been using it for some time
19:26:39 <gundalow> #info Zuul is being used for the Ansible Network's repos, see https://github.com/ansible-network
19:26:50 <pabelanger> for resources, we have capacity in 4 regions, 2 openstack cloud providers
19:26:59 <zbr> pabelanger: great, so nothing to worry about. and the best is that we can run them in parallel without problems. if we have infra to run we are good.
19:27:05 <pabelanger> this budget today is being funded my community
19:27:15 <gundalow> #info Example PR being tested and merged via Zuul https://github.com/ansible-network/network-engine/pull/231
19:27:30 <decentral1se> ah! community funding ...
19:27:42 <decentral1se> I expect molecule builds to only get heavier and heavier
19:27:56 <pabelanger> as we build the new zuul control plane, it has been noted that more projects testing, need more capacity
19:27:56 <decentral1se> so we will surely be expected to contribute at some point in funding, no?
19:28:11 <decentral1se> we'll need to cover integration tests for 5+ drivers or something ...
19:28:16 <gundalow> Molecule's test bill will be covered by Red Hat
19:28:19 <themroc> already something like 5 hours today, without support of any cloud provider :/
19:28:41 <decentral1se> ok, that is good
19:28:44 <pabelanger> right, funding should be coverged by redhat
19:28:53 <zbr> pabelanger: it seems that these are using comments and not the github checks api, not a deal-breaker.
19:28:59 <pabelanger> if we need more capacity, to ensure testing is working and making molecule stable, that is a good problem to have
19:29:19 <decentral1se> sure. what access would we have to build configuration / admin / etc.
19:29:20 <pabelanger> and capacity for running tests in public openstack cloud or aws, is there
19:29:22 <gundalow> zbr: jlk is working on GitHub Checks API (though he's busy with lots of stuff)
19:29:41 <gundalow> decentral1se: all test definitions are in GitHub repos
19:29:41 <pabelanger> decentral1se: you'll have full root access to the vms to do what you want for testing
19:29:50 <zbr> tbh, I do not think that molecule would a huge resource hog.
19:30:00 <pabelanger> https://github.com/ansible-network/sandbox/pull/27
19:30:14 <pabelanger> for example, is 2 molecule jobs I setup to run in zuul
19:30:19 <decentral1se> and what is the social contract on that? Is there some code of conduct?
19:30:22 <pabelanger> https://ansible-network.softwarefactory-project.io/zuul/status
19:30:26 <pabelanger> is where you see them run
19:30:41 <pabelanger> which, should be happening now
19:30:55 <pabelanger> (maybe not)
19:30:58 <pabelanger> there we go
19:31:04 <decentral1se> nice
19:31:17 <pabelanger> but those tests, it is running, are all in tree atop of tox jobs we've build
19:31:24 <pabelanger> so it was very minimal getting them up and running
19:31:25 <zbr> once we have the "seed" jobs I wll be able to provide help with zuul work as I use it a lot.
19:31:30 <themroc> waw, how to transform a bad news in a good one
19:31:41 <decentral1se> ;)
19:31:50 <pabelanger> the cool part of this, is job content is written in ansible
19:32:04 <decentral1se> pabelanger: are there meetings / place of contact to discuss usage of zuul?
19:32:11 <decentral1se> or monitor how resources are being utilised?
19:32:14 <zbr> #zuul
19:32:17 <pabelanger> so, you write your playbooks / role for production, then use different inventory file for testing, and it should work the same
19:32:22 <themroc> i learned about zuul reading about the integration of molecule in redhat, the convergence is awesome
19:32:42 <pabelanger> decentral1se: yes, #zuul is a good place also http://lists.zuul-ci.org/pipermail/zuul-discuss/
19:32:57 <pabelanger> themroc: yah, I have ansible roles, that use molecule, run by zuul
19:33:00 <pabelanger> works well
19:33:07 <decentral1se> ok, hammering questions on here ;)
19:33:13 <decentral1se> but what would a migration look like?
19:33:24 <decentral1se> we can retain our tox configuration?
19:33:41 <pabelanger> yes, your tox.ini should be able to be the same
19:34:03 <zbr> decentral1se: overall, working with zuul jobs is easier than you may think, the only barrier is creation of first set of jobs, after this is not harder than travis.yml
19:34:06 <pabelanger> then, we basically create in-tree configuration like: https://github.com/ansible-network/sandbox/pull/27/files
19:34:34 <pabelanger> from what I seen, tox was the entry point to testing for you, which makes this straightforward
19:34:38 <pabelanger> however
19:34:40 <pabelanger> what we could do
19:34:44 <themroc> looks pretty
19:34:45 <zbr> i will try to keep as much as possible inside tox.ini to allow developers to run tests locally, so zuul jobs will mainly only orchestrate tox executions.
19:35:13 <pabelanger> for the next step, if interested, is add github app to molecule and run some example jobs on PRs
19:35:26 <pabelanger> you'll get repos back on how things work
19:35:41 <pabelanger> then have another discussion about how to cut over to have zuul gate your repo, if wanted
19:35:48 <themroc> i am looking forward the first functional tests running on zuul !
19:36:02 <pabelanger> the main things with zuul gating, is it becomes the the only one to merge code
19:36:11 <pabelanger> so, humans would have to stop doing that
19:36:15 <themroc> note that some tests have been specially disabled when on travis
19:36:41 <decentral1se> that all sounds really nice then
19:36:42 <pabelanger> themroc: yes, you'll have full root access to vm to setup the way you want
19:36:48 <decentral1se> thanks for so much information
19:36:54 <pabelanger> np
19:37:05 <decentral1se> one question is how to prioritise this then
19:37:12 <decentral1se> but maybe we can't answer that now
19:37:34 <pabelanger> right, so on my side, we are actually building a new zuul control plane, so busy with that
19:37:39 <decentral1se> we can agree to continue the investigation with your proposal though! :)
19:37:52 <pabelanger> however, for the step about about having you demo PRs with zuul, we could do that this week
19:37:59 <pabelanger> on your side, it is just adding a github app
19:38:17 <decentral1se> well that seems like the thing to do then
19:38:25 <pabelanger> but yes, happy to set aside some time to help more investigation
19:38:37 <zbr> i just made a request to enable  softwarefactory-project-zuul app on molecule project, not sure who is approving it.
19:38:47 <decentral1se> and is there some further back and forth needed with ansible-network?
19:39:06 <decentral1se> I mean, is there still questions of feasability from that side?
19:39:10 <pabelanger> decentral1se: yah, we'll need to onboard you into our zuul tenant, I can be point for that
19:39:33 <pabelanger> i think the step here are, show it working well with molecule
19:39:43 <pabelanger> make sure mgmt is okay with bring you into our zuul
19:39:45 <pabelanger> profit
19:39:53 <pabelanger> we can do first step now
19:40:08 <pabelanger> #2, I don't see an issue becaue awx is also doing zuul today
19:40:29 <decentral1se> well then, bring on the zuul hegemony
19:40:31 <gundalow> Once ansible-network is setup with the new hosts etc we can look at this
19:40:47 <pabelanger> gundalow: yah, i think for final cutover that makes sense
19:41:11 <pabelanger> which on your side, will be a different github app to allow and revoke sf github app
19:41:24 <pabelanger> all the configuration in-tree will be the same
19:42:00 <gundalow> zbr: Please don't try and install Zuul in our GitHub org
19:42:18 <zbr> gundalow: i cannot do it anyway.
19:42:32 <decentral1se> final concern: what is the roadmap of zuul?
19:42:46 <decentral1se> ansible/* are banking on it being around for a while, I assume ;)
19:42:51 <gundalow> yup
19:42:58 <gundalow> and it's an important part of OpenStack :)
19:43:10 <decentral1se> ok, all clear, thanks again pabelanger
19:43:41 <pabelanger> np!
19:43:58 <pabelanger> yah, zuul is a top level project in the openstack foundation now
19:44:07 <pabelanger> only going to get larger IMO now
19:44:11 <zbr> pabelanger: thanks! this was a really productive meeting.
19:44:45 <zbr> pabelanger: yep, via https://opendev.org/ which is still in its early stages.
19:45:04 <decentral1se> Dare I squeeze in one final topic?
19:45:37 <gundalow> pabelanger: Thanks
19:45:41 <gundalow> decentral1se: sure do #topic
19:46:14 <decentral1se> cool
19:46:16 <decentral1se> #topic https://github.com/ansible/molecule/issues/1779
19:46:24 <decentral1se> can we move ahead with towncrier for change log?
19:46:37 <decentral1se> what are the concerns. I'm happy to defer to webknjaz in this regard
19:47:55 * gundalow is happy to use whatever people think is easy
19:49:40 <gundalow> Just seeing if webknjaz is around and can join
19:50:01 <gundalow> ah, nop, he's busy in another meeting at the moment
19:51:26 <tima> parking lot on that one then?
19:51:34 <zbr> i am not very keep about towncrier just because i do not know it and because is not used by any project that I know. so newtech, bit pessimistic.
19:52:07 <decentral1se> any ansible core lurkers? Can someone provide support for Reno?
19:52:08 <tima> does anyone know what the other ansible projects use?
19:52:19 <decentral1se> Reno apparently ... zbr, you mentioned this?
19:52:39 <decentral1se> support = a comment here and there ;)
19:53:21 <tima> I have a bias for using what the rest of the Ansible community projects are using unless the is a really strong case to deviate.
19:54:09 <pabelanger> I can help give a demo on reno / zuul integration
19:54:13 <pabelanger> we actually have a job for that
19:54:43 <pabelanger> https://zuul-ci.org/docs/zuul/releasenotes.html is example sphinx integration
19:54:45 <decentral1se> tima: https://github.com/ansible/ansible/blob/3433ca286db87ff50b9c93fa9330b6318386af7e/docs/docsite/rst/community/development_process.rst#creating-a-changelog-fragment
19:54:56 <tima> @pabelenger: how is reno? meaning, does it work well? stable? buggy? missing features.
19:55:01 <zbr> at least in openstack reno is used. i am just trying to avoid diverging the toolset, so my comments are limited to risk mitigation and not the quality of the tools themselves.
19:55:07 <decentral1se> oh yes, sphinx integration is good
19:55:31 <decentral1se> ok, I'm +1 on Reno, ansible/ansible using it and pabelanger assisting!
19:55:37 <pabelanger> tima: yes, works very well. Born out of openstack project and use across all the projects there
19:55:38 <tima> ugh. sorry @pabelanger. need coffee.
19:55:39 <decentral1se> and zbr swaying the tide
19:55:44 <pabelanger> easy to work with too, imo
19:55:50 <zbr> clearly support is a big plus
19:56:10 <tima> agreed @zbr
19:56:15 <pabelanger> I can mock up a demo for next week if you'd like
19:56:25 <decentral1se> heroic
19:56:48 <tima> I'd be interested in that.
19:57:17 <themroc> i am agnostic on this topic, i learn
19:57:30 <tima> Has anyone here used towncrier to compare it to Reno?
19:57:36 <tima> (For the sake of comparison)
19:58:30 <decentral1se> I've not, was just going on webknjaz report ...
19:59:13 <decentral1se> thing is, we need this to finalise v2.20 (we need to build the change log)
19:59:54 <decentral1se> or well, manually compiling is probably less work right now ...
20:01:04 <zbr> decentral1se: for 2.20 lets do it manually. much easier.
20:01:40 <decentral1se> ok, let's get the Reno demo / try out then and take it from there?
20:01:44 <tima> You mean holding up a release over how we generate a changelog is not a blocker? ;)
20:02:05 <decentral1se> hehe
20:02:25 <decentral1se> ok, happy to leave this now then
20:02:42 <tima> +1 leaving this for now
20:03:46 <gundalow> #agreed we will manually create the changelog fr v2.20
20:03:57 <gundalow> pabelanger: Thanks :)
20:06:11 <tima> Anything else today?
20:06:23 <decentral1se> think we're good
20:06:23 <tima> I'm going to have to drop off really soon.
20:06:38 <tima> @gundalow?
20:06:43 <gundalow> I think we are good
20:06:50 <gundalow> I had to poke quay.io
20:07:07 <decentral1se> oh, finally, just to be clear: we code freeze until pre-release feedback?
20:07:18 <gundalow> decentral1se: I think so
20:07:25 <decentral1se> grand job
20:07:33 <tima> agreed @decentral1se
20:07:41 <themroc> great day
20:07:52 <decentral1se> solid stuff, I'm out the door
20:07:56 <gundalow> Thanks as always everybody!
20:07:56 <decentral1se> thanks all, great stuff
20:07:59 <gundalow> #endmeeting