21:00:44 <strigazi> #startmeeting containers 21:00:45 <openstack> Meeting started Tue Jan 8 21:00:44 2019 UTC and is due to finish in 60 minutes. The chair is strigazi. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:00:46 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:00:49 <openstack> The meeting name has been set to 'containers' 21:00:51 <strigazi> #topic Roll Call 21:00:57 <eandersson> o/ 21:00:57 <strigazi> o/ 21:01:00 <schaney> o/ 21:01:06 <imdigitaljim> o/ 21:02:05 <strigazi> welcome back imdigitaljim 21:02:32 <strigazi> #topic Announcements 21:02:38 <imdigitaljim> ty 21:03:03 <strigazi> Devstack will run by default with python3 soon 21:04:13 <strigazi> I pushed this patch to switch but I think we miss a couple of things 21:04:23 <strigazi> mostly in the test code, I think 21:04:37 <strigazi> #link https://review.openstack.org/#/c/629126/ 21:04:51 <strigazi> PS 1, reveals the errors 21:05:08 <strigazi> PS 2 needs a fix in .zuul 21:06:10 <strigazi> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-January/001465.html 21:06:56 <strigazi> have a look if have a moment 21:07:20 <strigazi> #topic Stories/Tasks 21:09:19 <strigazi> From me the only item is about nodegroups that I spent time with my colleague this week: 21:09:22 <strigazi> #link https://review.openstack.org/#/q/status:open+project:openstack/magnum+branch:master+topic:magnum_nodegroups 21:09:58 <imdigitaljim> for me been catching up from my break but also working towards inplace upgrades (not just kubelet swaps) 21:10:00 <strigazi> Have a look in the API and the structure of the heat templates. 21:10:29 <imdigitaljim> comments about that were to keep the ngs conditional so they arent required in the drivers (at least immediately) 21:11:10 <strigazi> existing drivers will work as they are now 21:14:30 <imdigitaljim> https://review.openstack.org/#/c/629273/1/magnum/conductor/handlers/cluster_conductor.py 21:15:26 <imdigitaljim> wouldnt this have odd behavior if the driver doesnt support ngs? 21:15:29 <strigazi> this fakes the ngs. We have ngs at the moment but only tow. 21:15:33 <strigazi> this fakes the ngs. We have ngs at the moment but only two. 21:15:50 <strigazi> it woulds say, not implemented. 21:15:55 <strigazi> it would say, not implemented. 21:15:59 <imdigitaljim> oh okay 21:16:05 <imdigitaljim> ill keep watching (i know its still WIP) 21:16:11 <imdigitaljim> just wanted to mention :) 21:16:42 <strigazi> sure, I think it is mentioned it in the spec 21:17:20 <imdigitaljim> can you link spec again 21:17:25 <imdigitaljim> i misplaced that link 21:18:09 <strigazi> https://review.openstack.org/#/c/607363/ 21:21:01 <imdigitaljim> also we'll look to introduce the centos driver for review after we finish some tests and tweak for upstream consumption 21:21:29 <strigazi> sounds good :) 21:21:48 <imdigitaljim> and hopefully can converge on requirements for acceptance :) 21:22:41 <imdigitaljim> nondriver-level magnum only needs the cinder support PR i have open 21:22:43 <strigazi> I'm sure it will 21:22:52 <strigazi> link? 21:22:57 <imdigitaljim> https://review.openstack.org/#/c/615592/ 21:23:04 <imdigitaljim> ill fix the merge conflict thats there now 21:23:17 <strigazi> cool 21:23:40 <strigazi> imdigitaljim: you purge/delete volumes? 21:23:57 <imdigitaljim> were adding the ability to 21:24:20 <imdigitaljim> for us, most of the time if a cluster is being deleted all the volumes can be deleted as well 21:24:27 <imdigitaljim> but were making it toggle-able 21:24:40 <imdigitaljim> default (dont delete) 21:24:42 <strigazi> ok 21:25:15 <imdigitaljim> otherwise we'd have tons of dangling volumes 21:25:16 <imdigitaljim> over time 21:29:11 <imdigitaljim> also @strigazi our organization is almost approved (on our side) to upstream to Kubernetes so we'll probably be making some openstack-cloud-provider changes that may help out all drivers too 21:29:39 <imdigitaljim> all drivers that use the openstack ccm at least 21:29:51 <strigazi> in which part of the occm? 21:30:54 <imdigitaljim> a few sections, fixing some keystone auth bugs/logging, adding some loadbalancer edge cases, and add some annotation based features to the loadbalancer 21:30:59 <imdigitaljim> are a few that come to mind 21:32:07 <imdigitaljim> nothing major so far 21:32:22 <openstackgerrit> Merged openstack/magnum master: containers: clean-up build code https://review.openstack.org/625996 21:32:22 <strigazi> imdigitaljim: have you tried to tweak the keystone authz part? 21:32:24 <imdigitaljim> we'd like to help the migration process for volumes from intree to out-of-tree as well 21:32:52 <imdigitaljim> a little bit yeah 21:32:58 <imdigitaljim> wrt anything particular? 21:33:03 <strigazi> imdigitaljim: eg use ldap for groups and not keystone projects? 21:33:15 <imdigitaljim> oh, not really no 21:35:06 <strigazi> imdigitaljim: authz with gitlab? 21:35:20 <strigazi> if you use it 21:35:53 <imdigitaljim> no we dont 21:36:03 <imdigitaljim> but that sounds like a useful addition tbh 21:36:33 <strigazi> not sure how we can have many 21:36:56 <imdigitaljim> specifically i added some missing fields to use and tempered the logging 21:36:57 <strigazi> maybe keystone with the webhook and gitlab with dex? 21:37:04 <imdigitaljim> and like one other bug fix 21:37:11 <strigazi> cool 21:37:23 <imdigitaljim> (i just checked) had to refresh my memory sorry 21:37:49 <imdigitaljim> so yeah no major keystone changes yet 21:37:57 <imdigitaljim> wrt occm 21:38:48 <strigazi> oh, and one more thing, I opened an issue with use client-keystone-auth with kerberos, not sure how is to implement. https://github.com/kubernetes/cloud-provider-openstack/issues/406 21:40:50 <imdigitaljim> i think we've settled on not requiring the plugin file and just require the OS_TOKEN instread 21:40:54 <imdigitaljim> instead* 21:40:58 <imdigitaljim> internally 21:41:10 <strigazi> fair enough 21:41:56 <imdigitaljim> we dont like either solution 21:42:20 <imdigitaljim> but at least not using an additional script seems more intuitive 21:44:09 <imdigitaljim> but yeah thats all for me 21:44:22 <imdigitaljim> glad to be back, hit me up if you need anything @strigazi @flwang 21:44:34 <strigazi> thanks imdigitaljim 21:45:20 <strigazi> anything else, anyone? 21:45:33 <jakeyip> o/ 21:46:10 <strigazi> jakeyip: hello 21:46:43 <jakeyip> I've noticed magnum-tempest-plugins reviews are not getting much love, anyone here able to help push things along? https://review.openstack.org/#/q/project:openstack/magnum-tempest-plugin 21:47:02 <jakeyip> people are submitting duplicate reviews which is a waste of time... 21:47:36 <strigazi> jakeyip: will take a look tmr, your patches first :) 21:48:50 <jakeyip> strigazi: thanks! I've gone and +1 / -1 some, hope that helps. 21:49:22 <strigazi> that's great 21:51:17 <strigazi> eandersson: you recommend this approach? https://review.openstack.org/#/c/597994/1/magnum_tempest_plugin/common/client.py@a36 21:51:48 <eandersson> It's fine, as-is, but it's the more general approach to this issue 21:51:54 <eandersson> for OpenStack 21:52:49 <strigazi> eandersson: I thought the comment was old, did see the date 21:52:53 <strigazi> jakeyip: ^^ 21:53:04 <eandersson> haha yea comment was ~30s ago :p 21:53:44 <eandersson> decoding in py27 is fine as well, but will convert it to a unicode obj 21:53:50 <eandersson> which is probably fine 21:53:54 <jakeyip> thanks I'll take a look! 21:55:53 <strigazi> thanks 21:56:25 <strigazi> let's wrap the meeting, thanks everyone 21:57:09 <strigazi> #endmeeting