Tuesday, 2026-04-28

opendevreviewMauricio Harley proposed openstack/governance master: Add PQC Migration popup team  https://review.opendev.org/c/openstack/governance/+/98206209:48
opendevreviewFrancois Magimel proposed openstack/governance master: Fix numbered list syntax for migrate to privsep goal  https://review.opendev.org/c/openstack/governance/+/98655416:08
gouthamrtc-members: a gentle reminder that our weekly IRC meeting will be hosted here in ~50 minutes: https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee#Agenda16:09
gouthamr#startmeeting tc17:00
opendevmeetMeeting started Tue Apr 28 17:00:29 2026 UTC and is due to finish in 60 minutes.  The chair is gouthamr. Information about MeetBot at http://wiki.debian.org/MeetBot.17:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.17:00
opendevmeetThe meeting name has been set to 'tc'17:00
gouthamr  Welcome to the weekly meeting of the OpenStack Technical Committee. A reminder that this meeting is held under the OpenInfra Code of Conduct available at https://openinfra.dev/legal/code-of-conduct.17:00
gouthamr  Today's meeting agenda can be found at https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee#Agenda17:00
gouthamr#topic Roll Call17:00
mnasiadkao/17:01
gouthamr*kaomoji stretches* 17:02
gouthamr_silence_17:02
dansmitho/17:02
frickler\o17:02
gouthamrcourtesy-ping: noonedeadpunk, spotz[m], cardoe, bauzas17:02
cardoeo/17:03
gouthamrno clue on how tonyb is doing, but, i guess he's busy somewhere17:03
bauzaso/17:04
gouthamrokay, let's get started.. 17:05
gouthamr#topic Last Week's AIs17:05
gouthamrwas two weeks ago at this point, because we're going to digest the PTG AIs across the board for a while17:05
gouthamrnoonedeadpunk meant to send ML notifications for Venus retirement and Vitrage deprecation.. this was done, there have been no updates on the ML posts since17:05
spotz[m]o/17:06
gouthamr#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/QS3TFDCMT5ZBSNFXUVYGLXTKPKIOQDOZ/ ([vitrage][tc] Deprecation of the Vitrage project)17:06
gouthamr#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/PGT5EUDSNEILHHA2GY6DOQ7FRFIXHTC7/ ([venus][tc] Retirement of the Venus project)17:06
gouthamrso i guess the next thing here is to make progress on the patches themselves to tear down test jobs, empty out the repositories, governance change to move deliverables out of projects.yaml etc17:07
gouthamrwill let noonedeadpunk drive that, and we can use our tracking etherpad to catch up on the progress/updates17:08
gouthamrin a similar vein, an infra core still needs to help with the requestsexceptions retirement:17:09
gouthamr#link https://review.opendev.org/c/openstack/project-config/+/979808 (Retire requestsexceptions)17:09
mnasiadkafrickler, fungi: ^^ if you have a minute17:09
gouthamrty, that'll help us with the governance change here:17:10
gouthamr#link https://review.opendev.org/c/openstack/governance/+/979815 (Retire requestsexceptions)17:10
gouthamrcardoe: noonedeadpunk: you had an AI regarding defining the  MariaDB 11.8 collation changes goal.. any updates to share?17:11
cardoeWe didn't sync up.17:12
cardoeI'm testing some settings out locally and will come back with some doc updates.17:12
cardoeDuring PTG there was a conversation about some of the centralized docs vs project docs.17:12
gouthamrack ty, i can move this to the tracker too.. i think in case this would spawn a community goal, we'd need some heads up 17:13
cardoeOne thing I'm thinking is that there might be a chance for us to centralize the DB creation into 1 doc and have all the project docs point to that as the "official" place.17:13
cardoeBecause some projects tell you "run this MySQL command to create your DB" and they've got the collation in the example command (which is wrong)17:13
gouthamrack, like a db operations & best practices guide17:13
gouthamrthat sounds useful17:14
cardoeIt was just a thought and my patches to the various projects would then just be to link to that.17:14
gouthamrmakes sense, ty for thinking about this.. will check back next week17:15
cardoeWe've got one project that's got stuff about latin1 saying its the default.17:15
cardoekeystone talks about when Ubuntu 16.04 is released the command should be X but for now use Y.17:16
cardoeSo if centralizing sounds good, we can take an action on that.17:16
gouthamr++17:17
gouthamrour next AI was to make changes to the PQC pop up team proposal.. think mharley[m] updated it and you folks have been reviewing, ty for that:17:17
gouthamr#link https://review.opendev.org/c/openstack/governance/+/982062 (Add PQC Migration popup team)17:17
opendevreviewMerged openstack/governance master: Retire requestsexceptions  https://review.opendev.org/c/openstack/governance/+/97981517:18
gouthamrfinally, a bunch of other TODOs were related to the PTG17:18
gouthamrwhich was well attended, ty all for the great discussions there.. 17:18
gouthamrthe summary is here:17:18
gouthamr#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/XYXFYV4GAVNHHDWFEBYFITZAAYPQQS6K/ ([tc][ptg] 2026.2 "Hibiscus" Technical Committee PTG Summary)17:18
gouthamrif you're like me and you would like to relive the whole thing in the copious amount of free time in your day, the discussions are also uploaded here:17:20
gouthamr#link https://www.youtube.com/playlist?list=PLhwOhbQKWT7VMSRlyvfJ1eBsl8BjPFNpp (Hibiscus PTG)17:20
gouthamr:D that's all i see for AIs, was anyone else working on anything to bring up here?17:20
gouthamroh, missed one, we made a bulk transfer request on launchpad:17:21
gouthamr#link https://answers.launchpad.net/launchpad/+question/824113 (Transfer ownership of 11 OpenStack driver teams to ~openstack-admins)17:21
gouthamrthis was done as well.. 17:21
gouthamrwe have a lot more cruft though, a number of smaller teams - but these are without the security issues risk.. i will try and compile that data for us to opportunistically tackle sometime soon17:22
gouthamrthe smaller teams are for managing repos/tasks that we don't ship, or support within the security bug proces - can share the list next week 17:24
gouthamrlet's move on.. 17:24
gouthamr#topic ansible-collections-openstack releases and committers17:24
gouthamrcardoe: the next three topics are yours, ty for adding these.. what's going on?17:25
cardoeSo folks want more releases. Folks want patches merged. There's no real traction with the project.17:26
cardoeWhen adding folks with more access gets brought up there's a flurry of merges (without a release) but it tapers off.17:26
cardoeWe've hadd mnasiadka, noonedeadpunk and myself offer to get +2 and make releases since we've been authoring a lot of the changes lately without feedback from the existing +2's.17:27
mnasiadkaEspecially Kolla-Ansible and OSA use that deliverable in their code17:28
cardoeAnd releases don't happen with the normal OpenStack release cycle.17:28
cardoeAnd folks install it from Ansible Galaxy, which requires a release and not just patches sitting in a "stable" branch. So we need more frequent releases.17:28
gouthamrquestion: this is about this repo, correct?17:30
gouthamr#link https://opendev.org/openstack/ansible-collections-openstack 17:30
cardoeYes.17:30
fricklerwhere did these talks about more access happen? iiuc gtema as sdk ptl is the main person to contact?17:30
mnasiadkaI don’t know if it should be managed by the release team - but pushing tags is currently permitted for ansible-collections-openstack-release group - which has 4 members, from which I assume only gtema is active (but he expressed he doesn’t have a lot of interest in maintaining that)17:31
gouthamr#link https://review.opendev.org/admin/groups/ansible-collections-openstack-core,members17:31
gouthamryeah, probably gtema's the only "active" person here17:31
cardoeIn #openstack-ansible-sig and #openstack-sdks I've brought it up17:31
cardoeYeah gtema's the only active person.17:31
cardoeHe's said to me to get the other folks to vote to give access. But they're non-responsive.17:32
cardoeI'll admin I stink at finding groups in review.opendev.org and didn't know Kevin Carter was on that list so I never contacted him.17:32
gouthamroh, from what i recall, we struggled to get gtema access17:32
fungiif the tc asks me or another gerrit admin, we can add someone17:33
cardoeBut I'll go yell at Kevin right now.17:33
gouthamrwait17:33
cardoeI contacted the others via IRC and via email.17:33
gouthamri think there's a bigger problem here.. is your intent to revive the Ansible SIG?17:33
gtemaI can also add whomever you say me to, just again expressing the concern about nobody really commited to the project17:33
cardoeMy intent is to not have OpenStack contributors pinging me to ask me if this project is dead. :-D17:34
cardoeHowever the TC wants to structure the maintainership of that project I'm happy to help.17:34
gtemaand giving release right to somebody who do not understand all the relations is not cool17:34
cardoeI'd just like to make frequent releases to Ansible Galaxy to fix issues that users run into17:34
gouthamryes.. i feel like adding people wouldn't solve the underlying issue that the repo was shepherded by a SIG and we'll just be sidestepping that 17:34
mnasiadkaWell, a-c-o is Ansible SIG deliverable, so we should keep the SIG healthy if we want to maintain a-c-o17:34
gouthamryeah, perhaps.. i feel if we asked cloudnull, sshnaidm - they'd say they're no longer maintaining the SIG and would be happy to hand that over to people that are interested 17:35
gouthamrmaybe we use this opportunity to drive that?17:36
frickleraccording to the wiki, mnaser is also a chair still17:36
frickler#link https://wiki.openstack.org/wiki/Ansible_SIG17:36
mnasiadka#link https://opendev.org/openstack/governance/src/commit/32999aa5b600354153c8000da076a8d27e1aeded/reference/sigs/sigs.yaml#L217:36
mnasiadkaThat’s probably more accurate, but that one points to cloud null and sshnaidm17:37
gouthamryes, this should be the source of truth:17:37
gouthamr#link https://governance.openstack.org/tc/reference/sigs/index.html 17:37
cardoeI've emailed Sagi a number of times asking about the state of the project. I never get back an answer. But usually within ~48 hours of me doing that any of my proposed changes get merged.17:38
gouthamrthe release management for this has been external too? 17:38
mnasiadkaI assume Ansible Galaxy consumes tags17:39
mnasiadka#link https://opendev.org/openstack/ansible-collections-openstack/src/commit/885fadb31e71551087e242dde94985f5f0ba2ec5/.zuul.yaml#L36917:39
gtemayes, releasing is manual17:39
gouthamrseems like there's some magic in-repo and tools that you run by hand to tag and push to galaxy17:39
mnaseransible-galaxy gets a .tar.gz pushed over an api that it publishes17:39
gtemanope, it's automatic, but the tag pushing is semi-manual17:39
gtemahttps://opendev.org/openstack/ansible-collections-openstack/src/branch/master/docs/releasing.md describes the procedure17:39
mnasiadkaYou just push a signed tag from what I see, not something unusual for other repos in openstack namespace that are not handled by release team17:39
gouthamrack.. makes sense for me to point sagi and kevin to this discussion and ask them if they'd be okay with handing the SIG's governance over to you folks.. 17:40
gouthamri mean, ask(again) about what their plans are.. if you don't hear back, make the proposal and the TC can help fix it up.. 17:40
gouthamrif it helps, we've used a lazy consensus approach quite a lot, and here's another place it can help17:42
mnasiadkaShould that another attempt be CCed to openstack-discuss, so it’s easier for TC to track if the answer happened?17:42
gouthamryou don't hear back within a reasonable amount of time, we can assume acceptance and move on.. 17:42
gouthamryes please17:42
cardoemnasiadka: do you wanna send one? I'll reply?17:43
mnasiadkacardoe: I can send one, but next week - I’m off from Wed this week17:44
gouthamrokay, looks like we have next steps here.. 17:45
gouthamr#topic pbr runtime vs importlib runtime17:45
mnasiadkagouthamr: looking at sigs.yaml that might require some cleanup in general…17:46
cardoeSo I threw that one on there. We've got pbr as our build tool. But a lot of things also import it but not ALL in OpenStack and they use it for pbr.version.VersionInfo("mypackage")17:46
gouthamryes, and tonyb offered to pitch in a while ago.. might need a new owner, and probably a piecemeal approach17:47
cardoeIn the past there was a lot more that pbr did at runtime17:47
cardoeNow days that's a big wrapper around importlib.metadata.version("mypackage")17:48
gouthamri see, is your ask to stop using pbr.version.VersionInfo at runtime and switch to importlib.metadata.version()17:48
cardoeJust wondering if we should take this as a chance to slim things down.17:48
fungithe one thing i know of you lose is the ability to report the git commit id info17:50
fungiwhich can sometimes be useful when installing a rando git checkout17:50
clarkbbecause that is loaded from a specific pbr metadata file in the package that importlib doesn't know about.17:50
fungiyeah, the pbr.json package metadata file holds that reference17:51
fungi(importlib could report it, you just need to know to tell it to load that, and would effectively be reimplementing pbr.version.VersionInfo yourself)17:51
cardoeDoes it?17:53
cardoeCause I see it calling pbr._compat.metadata.get_version()17:53
cardoeWhich the plumbing if you follow it is using importlib.metadata.version if its available17:54
gouthamrtime check, sorry.. this is probably a mailing list discussion, cardoe? you'll have other pbr maintainers pitch in as well.. 17:54
cardoeworks for me.17:54
fungiyes, we can follow up later17:54
* fungi is in a conference session right now17:55
gouthamr#topic Secure RBAC: manager role definition (keystone only or wider scope)17:55
gouthamri think a bit of this discussion happened here on #openstack-tc yesterday17:55
gouthamrleading to gmaan sending this over17:55
gouthamr#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/ZEKOWEMLY6F2RFFXVD37QRQPMB35H5PR/ ([policy][tc] RBAC global auditor (reader))17:55
cardoeSo I spoke with gmann a bit. My concern is that the docs say its a specific role for keystone to implement management of projects.17:55
cardoeBut then other projects have started to adopt it as well.17:56
cardoeShould we not lift the wording restriction on it being special to keystone then?17:56
gmaancardoe: it is for wider scope and i think its just doc bug which can be improved17:56
cardoeokay I can submit patches to that effect.17:56
gmaan++, I can review that. thanks17:57
gouthamrty for raising that.. project teams have been following this:17:57
gouthamr#link https://governance.openstack.org/tc/goals/selected/consistent-and-secure-rbac.html#implement-support-for-project-manager-personas 17:57
cardoeAnd my last item was that we had auditor but haven't had traction on it and as you posted gmann made a mailing list link.17:57
gmaanyeah, our goal doc says that but I agree to update the keystone doc also17:57
gouthamrack17:57
gouthamrcardoe: yep! if you/someone you know can help, that sounds like a really nice improvement we want, but don't have someone driving the effort for17:58
cardoeYep. Working on writing up the use cases for it.17:59
gtemaI raised my hand on the ML on the topic18:00
cardoeSo the one example I can think of that I've read that the Rackspace Flex team has wanted was to be able to see the hypervisor that a VM is on for example.18:00
gouthamrgreat stuff, ty!18:00
cardoeRight now that requires role:admin but they don't necessarily want to give that level of power to that account.18:00
cardoeSo what we've been looking is a read only user that can see some of the privileged metadata.18:01
gmaangtema: thanks, did not get chance to reply to that but I have marked it unread to respond sometime later today18:01
gouthamrwe're past the hour18:01
cardoeAnother example would be member/reader cannot see the VLAN assigned in neutron to a network only role:admin can see it.18:01
gouthamrlet's wrap up here, we can continue the conversation though18:01
gouthamrdoes anyone want to note anything else for the minutes today?18:01
gouthamrsorry we didn't get to the standing topics.. but, cognitive overload from the last week is real. I'll make changes to the meetings repo to officially drop the APAC friendly meeting18:02
gouthamrslot18:02
gouthamrand we've finalized this time for at least until 2026.2 wraps up18:03
gouthamrthank you all for participating18:03
gouthamr#endmeeting18:03
opendevmeetMeeting ended Tue Apr 28 18:03:27 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)18:03
opendevmeetMinutes:        https://meetings.opendev.org/meetings/tc/2026/tc.2026-04-28-17.00.html18:03
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/tc/2026/tc.2026-04-28-17.00.txt18:03
opendevmeetLog:            https://meetings.opendev.org/meetings/tc/2026/tc.2026-04-28-17.00.log.html18:03
dansmithgouthamr: I'm sorry for even saying this, but could I convince you that "AI" is longer an allowable acronym for "Action Item"?18:04
bauzasgouthamr: do I need to specifically add a topic for https://review.opendev.org/c/openstack/governance/+/984958 ?18:04
dansmithevery time you use it it shocks me to my sou18:04
gouthamr:( 18:04
* dansmith wonders who that frowny face was for18:05
gouthamri know, noonedeadpunk said that once too.. i'll stop saying AI.. at this point we need a physical thingy that we can "nail down to the wall" and keep yelling at (/me invokes sean-k-mooney) 18:05
gouthamrdoesn't take me more time to type "Action Items" :P ty for noting it, dansmith 18:06
dansmithah, okay glad I'm not the only one.. "item" is only two more characters than "AI" and I think it's plenty descriptive in most contexts :)18:06
gouthamrbauzas: yes.. i think you've bootstrapped the project and people asked questions on it being in the "openstack" namespace on the ML18:07
gouthamri need to revisit the thread if we noted the specifics.. we lack a namespace for "unofficial but related" work i think18:07
gouthamrand we can't abuse "x/" as fungi and clarkb told us.. 18:07
bauzasgouthamr: this is why we have a note in the README18:08
gouthamrthat's an excessive burden on you though to define before you make progress18:08
gouthamrif anything, let's discuss taht bit further?18:08
gouthamrthe readme in the repo18:08
gouthamr?18:08
bauzasbut hopefully, we could accept https://review.opendev.org/c/openstack/governance/+/984958 and then this souldn't be a problem18:08
JayFleaningtowerofstack/some-unofficial-repo when :D 18:09
sean-k-mooneyo/18:09
bauzasgouthamr: see https://sbauza.github.io/ai/openstack/agentic-workflows-week-one/#finding-a-home18:10
gouthamri don't know what the "stackforge" story was, but i'm afraid to ask now and be reprimanded18:10
sean-k-mooneyit was a place to build unoffical project18:10
bauzasgouthamr: and see https://opendev.org/openstack/agentic-workflows#openstack-agentic-workflows 18:10
sean-k-mooneybefore the big tent18:10
bauzas"This repo is not an official OpenStack deliverable"18:10
sean-k-mooneywhere thign could incubate and graducate18:10
* sean-k-mooney is refing to stackforge18:10
bauzasanyway, this is too late for me 18:11
bauzas20pm here18:11
bauzas8pm sorry18:11
bauzas(for discussing)18:11
gouthamryes, please, we can indeed add a topic to next week, bauzas 18:11
* bauzas stops18:11
mnasiadkathis would be an interesting clock system that would have 20pm18:11
JayFstackforge lasted long enough for it to be seen as semi-official18:11
bauzasgouthamr: noted, will add18:11
JayFwhich I think is part of the problem :D 18:11
JayF(any unofficial collection of things becomes official if it sticks around too long)18:12
gouthamryes.. we had a concept of "experimental" apis in manila, where APIs sat flushed out for a decade and the team treats them like any other part of the API.. so, users are like, wth, this is just official, isn't it.. why the weird naming18:13
gouthamrso i get that18:13
spotz[m]I know OSA started in Stackforge and then graduated18:44

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!