16:00:22 #startmeeting nova 16:00:23 Meeting started Thu May 14 16:00:22 2020 UTC and is due to finish in 60 minutes. The chair is gibi. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:24 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:26 The meeting name has been set to 'nova' 16:00:31 o/ 16:00:39 \o 16:00:44 o/ 16:00:44 o/ 16:00:55 ~o~ 16:00:58 * bauzas now sweats after running for being in time 16:01:06 I'm probably breathing too loud 16:01:49 lets get started 16:01:51 #topic Bugs (stuck/critical) 16:01:53 * artom follows the ea/ee switch to its logical end, and concludes that bauzas is sweeting for the meating. 16:02:03 No critical bugs 16:02:09 \o 16:02:16 #link 29 new untriaged bugs (-2 since the last meeting): https://bugs.launchpad.net/nova/+bugs?search=Search&field.status=New 16:02:22 artom: you're meen 16:02:54 is there any bug that needs special attention on this meeting? 16:03:56 if not then moving forward 16:04:18 #topic Release Planning 16:04:23 Ussuri has been released #link https://www.openstack.org/software/ussuri/ 16:04:26 Thank you all for making this release happen! 16:04:56 * gibi is going to have a cold beer for Ussuri after the meeting 16:05:03 fortunately, we aren't on Slack where stupid emojis and GIFs could happen 16:05:17 IRC++ 16:05:32 I can try to copy-paste some ASCII art if needed ;] 16:06:04 bauzas, 🚁 you were saying? 16:06:47 moving on ? :) 16:06:54 is there anything about the past or the coming release to discuss today? 16:07:14 I guess not so then 16:07:16 #topic Stable Branches 16:07:31 lyarwood: do you have any news for us? 16:08:04 gibi: nothing for stable today 16:08:12 thanks 16:08:19 #topic Sub/related team Highlights 16:08:23 API (gmann) 16:08:29 one things to disucss 16:08:56 for policy work we left the deprecated APIs to move to new defaults and scope 16:09:47 I can complete those in this cycle so that when we remove the old defaults we can remove from all 16:10:11 johnthetubaguy point was we should do for deprecated APIs also as they are still supported. 16:10:22 opinion? 16:10:33 deprecated APIs are mostly proxy APIs 16:10:51 We still have those? 16:10:58 I have no objection adding the new scopes for the still supported but deprecated APIs 16:11:02 I thought stephenfin made a huge cleanup 16:11:14 which ones are we talking of ? 16:11:35 bauzas: for nova-network its gone but other are there like sec grp, volume , image 16:11:47 ah rght 16:12:07 so then I tend to agree with johnthetubaguy 16:12:35 gmann: please file a blueprint for tracking purposes 16:12:36 * artom is still reading through the api-ref, but they all look to have been removed entirely in some microversion... 16:12:43 its policy things otherwise we can say no enhancement to deprecated things 16:13:13 agree with gibi and bauzas 16:13:21 gibi: ok, sure. spec less BP right as it is continue of old spec 16:13:21 deprecated means that we support regression bugfixes but we don't provide new features 16:13:32 but here, it's policy things 16:13:42 yeah 16:13:52 gmann: yes, it does not need a separate spec, just link to the old one 16:13:57 ok, thanks 16:14:24 that's all from my side on API today 16:14:33 gmann: thanks 16:14:41 Libvirt (bauzas) 16:14:45 well 16:14:52 we're slowing starting to organize ourselves 16:15:00 aarents reported 2 bugs that are reviewed 16:15:09 another one is being reviewed as well 16:15:13 https://etherpad.opendev.org/p/nova-libvirt-subteam 16:15:46 no other business to mention except maybe kashyap's q35 strategy but he's not around 16:15:49 that's it 16:16:00 * bauzas just needs to do his homework on the review side 16:16:07 bauzas: cool, thanks bauzas 16:16:10 #topic Stuck Reviews 16:16:16 nothing on the agenda 16:16:27 is there anything stuck that we need to discuss today? 16:17:27 #topic Open discussion 16:17:34 Specless blueprint: https://blueprints.launchpad.net/nova/+spec/six-removal (takashin) 16:17:44 Is there any objection approving it ? 16:17:58 ahem 16:18:15 this would make sense to work on this now that we only support py3 16:18:26 +1 16:18:33 this being said, it's a pain to both work on it and review such large changes 16:19:04 in the past, we discussed those kind of 'nice-to-have' large series during the PTGs so that the whole nova core team would agree on this 16:19:16 six isn't mock 16:19:25 It's boring work, but is it really hard? 16:19:43 and I'm afraid we could generate some bugs during reviews that would appear trivial 16:20:06 so at least we need to come up with a consensus 16:20:09 my major complaint is that it generates churn, conflicts, and backport annoyance 16:20:20 dansmith: that's my point 16:20:23 for no actual gain 16:20:45 backport anyways needs to be anyways py2 compatible during backporting 16:20:54 it's a major source of frustration for both the owner and the reviewers 16:21:03 mean for new changes we anyways will not support six right 16:21:08 for the owner, that will mean a constant rebase 16:21:30 for the reviewers, that will mean some code reviews that aren't honestly trivial 16:22:08 gmann: dansmith was referring to any backport from a py3-only file 16:22:20 mox removal was basically in the same camp, except it was limited to tests, and also had the goal of getting us off a technically unsupported library (even though it has never been a problem) 16:22:21 not the fact we would start changing ussuri or what other release 16:22:26 this is conflicts everywhere 16:22:36 dansmith: I exacly had mox in mind 16:22:56 I'm not opposed to the work 16:23:07 yeah py3-only file backport can be conflict things 16:23:13 but I just feel it's a massive effort for zero gain 16:23:17 dansmith, so when would you propose we remove it, if ever? When Ussuri is the oldest supported stable branch, so that we know backports don't have to care about Py2 16:23:18 ? 16:23:35 IMHO, this is a mental cleanliness ticky mark that is a bunch of work t make us feel better, for no gain. But I'm quite sure it'll just happen anyway 16:23:56 artom: I honestly feel nobody thought about backports when we started writing py3 only code 16:24:10 artom: if it were me I would spend no time on it pretty much ever, but waiting until py2 releases are out of support is at least something 16:24:20 just letting it die by attrition would be fine with me 16:24:34 honestly, tox is a wrapper 16:24:42 whoops 16:24:49 * bauzas facepalms 16:24:50 bauzas: that is why (due to backport) these cleanup was not part of py-2 drop goal. 16:25:06 OK I don't see a consensus to approve the bp. I will not this down in the bp linking back to this discussion. 16:25:13 I will note* 16:25:29 gibi: maybe having takashi around when we discuss it could be nice 16:25:32 hence a PTG thing 16:26:02 b/c I'm not sure he's full onboard with what it means in terms of patches carry over 16:26:03 bauzas: sure 16:26:32 but agreed with dansmith, letting it die by attrition sounds good to me 16:26:37 I'm OK to discuss it on the PTG or in a separate ML thread stared by takashi 16:26:41 six isn't exactly large 16:27:23 it's also not like mox where the whole *approach* to the code has to be different 16:27:41 thus it's just a little syntactic sugar in places 16:27:46 yup 16:27:53 again, just a wrapper 16:28:39 I was personnally offering to discuss this at PTG time because I do care of stable branches 16:29:07 and I don't want us to be too much py3-only pedantic on ussuri and later branches if that goes starting to be backported 16:29:24 somehow continuing to use six would have my favor :) 16:29:24 are we done? 16:29:33 done 16:29:38 nothing more on the agenda 16:29:42 I think so, was thinking out loud 16:29:43 is there anyithing else to discuss? 16:30:47 then thank you all for being here! For EU it is beertime! 16:30:50 #endmeeting