16:00:27 #startmeeting sdk_osc 16:00:27 Meeting started Thu May 20 16:00:27 2021 UTC and is due to finish in 60 minutes. The chair is gtema. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:28 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:30 Stephen Finucane proposed openstack/python-openstackclient master: compute: Update 'server resize --revert', '--confirm' help https://review.opendev.org/c/openstack/python-openstackclient/+/792406 16:00:30 The meeting name has been set to 'sdk_osc' 16:00:40 o/ 16:01:04 #agenda https://etherpad.opendev.org/p/openstacksdk-meeting-agenda 16:01:23 the only thing inside is patches for R1 16:01:53 I have spent 2 weeks going through cloud layer and replacing all direct rest api calls with rest 16:01:55 o/ 16:02:12 produced (I know) tons of things to review 16:02:24 and it feels frightening to look at them 16:02:49 I was touching 2-3 functions and ending up in a XL change due to all the dependencies 16:03:01 Fun stuff. 16:03:29 there are still things to complete, but somehow I need to start getting things merged, otherwise building properly chains is not possible anymore 16:03:43 Fair enough. 16:03:46 Makes sense to me 16:04:02 I also want to get stephenfin's stuff from master (fixing tests) into r1 branch also once it merges 16:04:22 outstanding things are the coe (magnum), for which there is no proxy at all 16:04:33 I'll start reviewing those next week (I'm off tomorrow) 16:04:34 and some remainings in other services 16:04:40 sure, thanks 16:04:53 sadly we have no functional tests running so far 16:04:54 Sounds good to me 16:04:59 https://review.opendev.org/c/openstack/devstack/+/791541 16:05:01 Thats unfortunate 16:05:08 all I'll ask is that mechanical stuff is separated out where possible (like the neutron changes where we dropped the json suffix) 16:05:33 I was trying to keep changes as small as possible, trust me ;-) 16:05:38 sweet 16:05:44 Appreciated! 16:05:53 but there are so many dependencies, that it's impossible 16:06:10 mostly unittest are in the form - throw away and start from scratch 16:06:45 actually figured out that clustering (senlin) cloud layer was completely broken and even tests verifying wrong behavior 16:07:04 so after trying half a day fixing it I gave up and proposed to drop it completely 16:07:19 oof 16:07:32 python-senlinclient is using directly proxy layer, so there should be nothing broken 16:08:30 overall I want to stop now feature development in master for now, since rebasing here is a nightmare 16:08:43 Makes total sense. 16:08:56 Wait for things to land before picking that type of thing back up 16:09:05 sure 16:09:43 there are actually aready now some conflicts with master, since there are some not approved/finished changes on block storage and I needed to rework it heavily (implementing missing things) 16:10:13 hope by end of summer we would be ready to merge things back to master 16:10:26 Ah alright. 16:10:40 so that maybe in autumn we can celebrate 16:10:49 LOL 16:10:53 out of curiosity, what was the reason for the feature branch? 16:11:01 so we can keep cutting 0.x releases until it's ready? 16:11:05 breaking of ansible 16:11:18 networking change is breaking ansible 16:11:24 ah, okay 16:11:42 due to the fact that ansible modules are not explicitly converting result to dict and Ansible tries to modify objects 16:11:52 we have surely our reasons to forbid that 16:12:13 so I wanted to finish this massive rework and in the meanwhile I started preparation in ansible modules 16:12:25 so that we would have sdk1.0 and ansible-collections R2.0 16:12:28 gotcha, that makes sense 16:13:00 anything from your side? 16:13:27 nope, I'm hoping to pick up work again on cinder stuff for OSC 16:13:41 and try reach feature parity with cinderclient this cycle 16:13:51 stephenfin, that would be pretty legit 16:14:04 yeah, this actually in my mind also heavily depending on the r1 work on restructure and adding of general features (quotaset, metadata) 16:14:13 I'm still hoping to do so using legacy cinderclient calls and migrate over afterwards, since it's easier focus on one thing at a time 16:14:34 On that note, I also need to continue migrating nova things to SDK, now that we have feature parity there 16:14:50 yeah, that's also great 16:14:51 nova would like to deprecate novaclient entirely sooner rather than later 16:15:10 One less thing to maintain 16:15:15 exactly 16:15:16 sounds like a honey to my ears 16:16:10 I think half of the things are already migrated. Only the most complex things left ;-) 16:17:58 ok. Feels like we have nothing else 16:18:41 zimmery said that he was almost done with the nova microversions 16:18:48 he's got like one bigger one left 16:18:58 and then they are at parity i think 16:19:05 sounds cool 16:19:27 I told him to send an email to the discuss list when he finishes to be like, look! I did a thing! 16:19:42 yeah, would be cool 16:19:49 And just generally update on the state of the sdk wrt nova 16:19:56 sure 16:20:06 Oh, and he is going to keep helping us through the summer 16:20:13 (northern hemisphere summer) 16:20:17 awesome 16:21:27 anything else? 16:22:07 Nope 16:22:14 I think we can end this :) 16:22:29 cool. Thank you both 16:22:34 Thank you! 16:22:39 #endmeeting