17:07:13 <chouseknecht> #startmeeting ansible-container
17:07:13 <zodbot> Meeting started Mon Dec  4 17:07:13 2017 UTC.  The chair is chouseknecht. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:07:13 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:07:13 <zodbot> The meeting name has been set to 'ansible-container'
17:07:18 <j00bar> 👣
17:07:24 <chouseknecht> w00t!
17:07:47 <j00bar> 0.9.3?
17:07:48 <chouseknecht> ttomecek: anything to discuss with the 'buildah' work?
17:07:58 <chouseknecht> yeah, 0.9.3 too
17:08:07 <j00bar> I still owe him a code review
17:08:48 <chouseknecht> i have spent no additional cycles on the RHEL image build and push to RHCC
17:09:13 <chouseknecht> Perhaps we bump that bit to post 0.9.3
17:09:20 <j00bar> So we haven't tested my patch as far as that?
17:09:30 <chouseknecht> we have not test yo patch
17:09:33 <chouseknecht> tested
17:09:57 <gregdek> +1 to post 0.9.3
17:10:14 <j00bar> Well... It was my hope to take another pass through the issues for nitpicky issues like the ones from Friday
17:10:30 <j00bar> So I propose:
17:10:44 <j00bar> Wednesday, wherever the code lies, that will be 0.9.3
17:10:52 <chouseknecht> +1
17:10:56 <j00bar> House, if you can test it, great. If not, shit.
17:11:01 <ttomecek> chouseknecht, regarding buildah, I think I'm really close to building ansible-container-demo with it; just need to fix the synchronize module in core; which I hope I'll do early this week
17:11:17 <gregdek> W00t!!!
17:11:41 * chouseknecht prepares to delete all docker code from ansible container
17:11:49 <chouseknecht> just kidding
17:11:54 <ttomecek> :D
17:12:01 <ttomecek> we still need to run the containers somehow
17:12:11 <chouseknecht> details
17:12:11 <j00bar> That's awesome, ttomecek!
17:12:48 <ttomecek> j00bar, also +1 on your 0.9.3 proposal
17:12:49 <chouseknecht> yeah, very cool.
17:13:31 <ttomecek> hopefully libpod will be the runtime to pursue for
17:13:52 <ttomecek> https://github.com/projectatomic/libpod
17:14:15 <ttomecek> they are in a process to rename the project
17:14:16 <j00bar> Is that the new kpod?
17:14:24 <ttomecek> correct
17:14:38 <j00bar> But it's not a lib?
17:14:47 <ttomecek> that's the thing! they moved kpod from cri-o to libpod, and libpod doesn't have a release, yet
17:14:55 <ttomecek> j00bar, lib + binary
17:15:02 <j00bar> O.o
17:15:08 <ttomecek> I think that's the reason they want to name it properly
17:15:29 <j00bar> Okey dokey
17:16:24 <j00bar> Alright, so reconvene Wednesday?
17:16:28 <j00bar> Any other business?
17:16:52 <chouseknecht> I'm working on a new K8s module...
17:17:04 <j00bar> Do tell?
17:17:04 <chouseknecht> https://github.com/ansible/ansible/pull/33517
17:17:34 <chouseknecht> One module that handles all the objects (k8s and openshift), and is not generated.
17:17:43 <j00bar> Whoa. Interesting.
17:17:49 <chouseknecht> fully documented.
17:18:11 <chouseknecht> still uses the openshift Python client that fabianvf is maintaining
17:18:24 <chouseknecht> but no need to generate hundreds of modules
17:18:48 <chouseknecht> and no need for code in module_utils
17:19:20 <ttomecek> chouseknecht, what about differences between openshift and vanilla k8s?
17:19:51 <chouseknecht> there are two different Ansible helper classes provided by the openshift Python client.
17:20:17 <chouseknecht> so it's just a matter of instantiating the k8s one, and asking if has the API version + model requested. If not, then try the openshift one.
17:21:00 <chouseknecht> and honestly, in a future iteration, we could just one Ansible helper that did the work for us.
17:22:20 <ttomecek> chouseknecht, right, interesting!
17:25:08 <fabianvf> chouseknecht: that is very cool
17:25:09 <chouseknecht> not as interesting as buildah, but the current modules suck, and i can't look at them any longer.
17:25:24 <chouseknecht> fabianvf: hello!
17:25:38 <fabianvf> chouseknecht: hi! sorry I'm in training this week but we're on our lunch break now
17:25:47 <chouseknecht> no worries
17:25:56 <chouseknecht> anything interesting going on?
17:26:28 <fabianvf> I'm being loaned to the openshift-ansible team, and am onboarding another contributor from the asb team for openshift-restclient-python (david zager)
17:27:00 * ttomecek would argue on the "cool factor"
17:27:25 <fabianvf> I'll still be primary contributor to the modules, just getting that bus number up
17:27:40 <chouseknecht> that sounds good
17:28:06 <chouseknecht> will likely need your helping landing the single k8s.py module
17:28:15 <fabianvf> yeah, totally
17:28:32 <chouseknecht> my hope is that we deprecate the need to generate modules
17:28:52 <fabianvf> there are also some additional changes that need to be made to the module helpers, there's a class of resource that they don't support well
17:29:15 <chouseknecht> does that class happen to include the 'build' object?
17:29:24 <fabianvf> stuff like buildrequest or imageStreamImport, which are more like triggers for actions and don't create a resource of the same kind that can be watched
17:29:24 <fabianvf> yep
17:30:10 <chouseknecht> i tried to generate the modules last night, as i was working on a bug fix. they broke on openshift_v1_build..
17:30:16 <chouseknecht> maybe it was buildrequest
17:30:49 <chouseknecht> we may not even need those
17:31:10 <fabianvf> hmm, shouldn't break module generation, I've only seen them break at run time
17:31:37 <chouseknecht> i may poke at it later today.
17:31:42 <fabianvf> chouseknecht: maybe, I have had a few bug reports for resources like that, so there are people trying to use them at least
17:32:55 <fabianvf> chouseknecht: only other reservation I have are resource name collisions between k8s/openshift APIs. There are a few resource names that exist in both, so defaulting to the first one found in k8s may result in behavior different than the user expects
17:33:47 <chouseknecht> we could add an override, to let the user force one over the other maybe
17:34:01 <fabianvf> I'll get my comments into the PR review though, this discussion is probably better suited for that
17:34:10 <chouseknecht> roger that
17:36:05 <fabianvf> I'm trying to get the module generation working for k8s 1.8 and 1.9 as well, there were some changes to the kubernetes gen repo that we need to update to track. changes should be relatively minor, don't anticipate it being super interesting
17:36:18 <fabianvf> think that's all I have to report
17:38:33 <chouseknecht> cool. thanks, fabianvf
17:38:56 <chouseknecht> anything else, or shall we adjourn?
17:39:25 <fabianvf> chouseknecht: how soon would you anticipate your proposed module being usable?
17:39:40 <chouseknecht> today/tomorrow
17:39:57 <chouseknecht> i'll ping you in the openshift-restclient
17:40:03 <chouseknecht> it'll at least testable
17:40:13 <chouseknecht> ^be
17:40:36 * chouseknecht wishes for an edit feature
17:43:05 <chouseknecht> i guess that's it.
17:43:17 <chouseknecht> thanks, everyone!
17:43:35 <ttomecek> yup, thanks for the chat, have a great day!
17:43:52 <chouseknecht> #endmeeting