Tuesday, 2026-05-12

opendevreviewGrzegorz Grasza proposed openstack/contributor-guide master: docs: add guidelines for tool-generated contributions  https://review.opendev.org/c/openstack/contributor-guide/+/98406109:37
opendevreviewGrzegorz Grasza proposed openstack/contributor-guide master: docs: add guidelines for tool-generated contributions  https://review.opendev.org/c/openstack/contributor-guide/+/98406109:40
opendevreviewSlawek Kaplonski proposed openstack/election master: Add configuration for 2027.1/"I" elections  https://review.opendev.org/c/openstack/election/+/98823810:48
opendevreviewSlawek Kaplonski proposed openstack/election master: Add placeholder for candidates for the 2027.1 election  https://review.opendev.org/c/openstack/election/+/98823910:48
gouthamrtc-members: a gentle reminder that the weekly meeting will be hosted here in ~55 minutes16:04
gouthamrthe agenda is here: https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee#Agenda16:04
bauzasgouthamr: I will probably need to drop, I have to go errand16:05
gouthamrack bauzas 16:06
opendevreviewMerged openstack/openstack-manuals master: install-guide: Add 2026.1 (Gazpacho) release  https://review.opendev.org/c/openstack/openstack-manuals/+/98353416:28
gouthamr#startmeeting tc17:01
opendevmeetMeeting started Tue May 12 17:01:06 2026 UTC and is due to finish in 60 minutes.  The chair is gouthamr. Information about MeetBot at http://wiki.debian.org/MeetBot.17:01
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.17:01
opendevmeetThe meeting name has been set to 'tc'17:01
gouthamrWelcome 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.17:01
gouthamrToday's meeting agenda can be found at https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee17:01
gouthamr#topic Roll Call17:01
dansmitho/17:01
noonedeadpunko/17:02
dansmithlight crowd today17:02
gouthamryeah 17:02
gouthamrcourtesy-ping: tonyb, frickler, spotz[m], cardoe, mnasiadka17:03
gouthamrnoted absence: b a u z a s17:03
cardoeo/17:03
frickler\o17:03
gouthamrlet's get started.. 17:05
gouthamr#topic Last Week's Action Items17:05
gouthamrwe had to resolve the ansible SIG's leadership today/this week17:05
gouthamr#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/NOK24Z4VKHH4XCVD5A7NYKZT3NH3WMDJ/ (Proposal to revive the Ansible SIG)17:06
mnasiadkaO/17:06
gouthamrmnasiadka: o/ in time17:06
gouthamrno update on the thread besides cardoe expressing support 17:06
cardoeSo mnasiadka got added with permissions to merge patches but we don't have permissions to cut releases.17:06
noonedeadpunkI support it as well, but I didn't reply, as text was well written )17:06
cardoePeople are chomping at the bit to get releases into ansible-galaxy so they can install the updates.17:07
gouthamroh, was there a discussion that happened elsewhere?17:07
cardoeNo. Just multiple people asking on patches and in IRC when we'll cut a new release.17:07
noonedeadpunkwell. they technically don't have to use galaxy17:07
noonedeadpunkand can install from git by SHA....17:08
noonedeadpunkso it should not be real blocker, unless they expect it to be part of the `Ansible` package17:08
noonedeadpunktalking about which - there were some updates to collection requirements to remain part of the Ansible17:08
noonedeadpunk* packaged Ansible17:09
gouthamrokay, cardoe and mnasiadka have been added to the core group for ACO:17:09
gouthamr#link https://review.opendev.org/admin/groups/ansible-collections-openstack-core,members (ansible-collections-openstack-core)17:09
gouthamrits worth recording this, but, did this go sideways?17:09
mnasiadkaYes, but we’re still not in the -release group - so can’t release anything17:09
gouthamroh17:09
noonedeadpunktalking about releasing17:09
cardoeSomeone added me so I added mnasiadka.17:10
gouthamr#link https://review.opendev.org/admin/groups/ansible-collections-openstack-release,members (ansible-collections-openstack-release)17:10
noonedeadpunkit's been also couple of years in Zuul to get galaxy publish role to merge17:10
noonedeadpunk#link https://review.opendev.org/c/zuul/zuul-jobs/+/89923017:10
noonedeadpunkas today it's in-repo playbooks, which I think would be highly beneficial to be generally available17:11
gouthamrack, good stuff17:11
gouthamrokay, can we talk through the next steps then?17:11
gouthamrmnasiadka: could you submit the governance change to update the SIG leadership?17:11
cardoeWe can look at switching to that.17:11
gouthamras SIG leaders then, you'll can rejig the gerrit memberships as you please17:12
mnasiadkaclarkb: would you be willing to have a look in the zuul-jobs patch linked by noonedeadpunk?17:12
mnasiadka(I have no rights in that repository)17:12
clarkbI guess? its honestly not a big priority for me (opendev doesn't use ansible galaxy at all17:12
mnasiadkagouthamr: will do17:12
clarkbwe just install the meta package for ansible which includes most of that or install things from source when it doesn't iirc17:13
fungiin this case it's about projects being able to publish their modules to galaxy, if memory serves17:13
gouthamrthank you17:13
gouthamr#agreed: lazy consensus on Ansible SIG leadership change has been reached17:13
gouthamr#action: mnasiadka will update the governance repo to reflect the new ansible SIG leadership17:13
clarkband I don't see that changing because galaxy and I don't get along17:13
noonedeadpunkcorrect17:13
clarkbthey change their api every 6 months and I gave up17:13
clarkbfungi: right but as an individual I don't care much if I never install things from the platform and if the platform is difficult to use17:13
noonedeadpunkI am also not using galaxy to install, but would love to publish things there17:13
clarkb(I personally don't think anyone should use galaxy)17:14
noonedeadpunkfor those who do use it17:14
clarkbwe tried multiple times to integrate with galaxy and in the middle of doing so they made breaking backward incompatbile changes17:14
fungiclarkb: right, zuul's inbuilt capability to share roles makes galaxy a lot less necessary for consumption by zuul jobs, so it's hard to find zuul-jobs repo reviewers to look at the proposed publication change17:14
clarkbafter ~2 or 3 times of that happening I basically stopped caring about anything galaxy17:14
fungii suppose it could go in openstack-zuul-jobs instead17:15
gouthamr^ wouldn't that be a better place for this?17:15
noonedeadpunkI can propose it there17:15
noonedeadpunkI mean. depends?17:15
gouthamrright now, our concern is openstack-ansible, and potentially kolla-ansible, ACO etc17:15
noonedeadpunkAs a zuul user I'd love to have some "existing" tool to publish to galaxy anyway17:15
fricklerthat doesn't change the set of (non-)interested core reviewers much, though17:15
clarkbI don't think its a problem to have in zuul-jobs. The problem is finding people with time to care17:15
fungiproposing it to openstack-zuul-jobs increases the chances that i'll feel comfortable single-core approving it, worst case17:16
noonedeadpunkI can propose there as well17:16
gouthamryeah, smaller blast radius if the API stability is our concern17:16
noonedeadpunkpatch was proposed 3 years ago and I checked recently and it still should work 17:17
noonedeadpunkas it's relying on `ansible-galaxy` binary for API interaction...17:17
noonedeadpunkanyway17:17
clarkbright if you use their tool then it will probably contineu to work17:17
noonedeadpunkI can propose to openstack-zuul-jobs17:18
clarkbtrying to proxy galaxy or integrate with other tools is painful17:18
gouthamrah, maybe they fix that more keenly17:18
gouthamr+117:18
noonedeadpunkno, you should not do this indeed....17:18
noonedeadpunkthey don't care about api stability...17:18
noonedeadpunkbut that is still what you have to do if you want to publish things for community17:18
gouthamrwe're kinda derailing here.. still on action items :) 17:19
noonedeadpunkor in case of openstack collections - if you want it to be included in Ansible packaging17:19
noonedeadpunk++17:19
gouthamrwe chatted a lot of RBAC stuff in the past week, and gmaan's threads are getting some attention.. unsure if there's anything to update for this week?17:19
gouthamras i mentioned in the TC's summary, we're not considering the "global reader/auditor" thing a part of the existing SRBAC goal.. and, if the "enforce_scope" patching can be dropped, we're at a position to call this goal done17:20
cardoeYeah we can make a new goal17:21
gouthamr+117:21
gmaanyeah, new goal is fine  but we should not extend (keep extending) the current goal scope17:21
gouthamr+117:21
cardoeThere's some traction and interest in the use case we proposed on the mailing list.17:21
gmaanI am more focusing on removal of enforce_scope. and adding it in the projects meeting agenda so that we can meet the target deadline17:22
noonedeadpunkI agree that global reader is not part of the current goal17:22
noonedeadpunkdespite how badly I want it to be17:22
gouthamrnext up, Venus retirement and Vitrage deprecation17:23
gouthamri reviewed the changes noonedeadpunk had for these, think we need some more work.. and in case of venus, we've a chicken-and-egg problem where we can't merge the repository nullification changes ourselves17:23
gouthamrwe were chatting here briefly that the TC initiated cleanups like this one should provide the TC access to the repos.. 17:24
gouthamror, we'll have to rely on infra cores to help with the cleanup, like frickler did for monasca ones17:25
gmaanthat is done via acl update I think that is first step in retirement17:25
gouthamrgmaan: not currently, we set the RETIRED.config ACL only after the governance change has merged.. 17:25
noonedeadpunkso I guess it's the last step...17:25
gmaanI think we always did it as part of 'Step 1: End Project Gating'17:26
noonedeadpunkBut what stops us from merging governance change?17:26
gouthamrit's failing because we validate that the repos you're retiring are "empty"17:26
noonedeadpunkdoh17:26
noonedeadpunkI think then we need to move up ACL change17:27
fungii guess the changes to empty the master branch are in need of someone with approval permission?17:27
gouthamryes17:27
gmaanfor example this one. we had no core to merge the repo cleanup https://review.opendev.org/c/openstack/project-config/+/91934817:27
gouthamrgmaan: ack, so a process deviation where you used common sense/executive decision :D 17:27
fungireordering things could solve that, yes17:27
gmaanyes, if no core available then TC takes the ACL and proceed 17:27
gouthamryes, would solve this.17:27
gmaanif cores are there then we go with normal process 17:28
gouthamrokay, worth a minor doc update so we don't get tripped up like this17:28
noonedeadpunkI thnk if cores are there then we usually don't retire17:28
gmaani did not update it in doc (my bad) but we followed the same in past for all such cases. 17:28
gmaannoonedeadpunk: :) yeah17:28
gmaanthis is another example https://review.opendev.org/c/openstack/project-config/+/91937617:29
gouthamrawesome, there's precedence.. thanks for sharing17:30
noonedeadpunkok, I will update the doc and move ACL to a separate patch17:30
gouthamrthank you noonedeadpunk 17:30
gmaannoonedeadpunk: thanks 17:30
gouthamranything else regarding this deprecation and retirement?17:30
* noonedeadpunk has finalized family matter which took last couple of month of my time17:30
gouthamrhope that means you inherited a bunch of money17:31
noonedeadpunkunfortunatelly not.... 17:31
gouthamrnext action item was about AI 17:31
* gouthamr sorry couldn't help17:31
fungihow meta17:31
gouthamrTaCT SIG has "agentic-workflows" now.. 17:32
gouthamrAGENTS.md goal proposal was abandoned17:32
gouthamrbauzas may register #openstack-agentic-workflows soon, so keep the chatter going there17:32
fungithe irc channel, specifically17:33
gouthamryes17:33
gouthamrwe also merged the PQC popup team, there's no dedicated IRC channel here.. but, mharley[m] is the TC liaison if you need someone to direct questions to17:33
gouthamrthat's all the action items that i was tracking, were you working on anything else to share?17:34
gouthamr#topic A check on gate health17:34
gouthamrhow has the gate been faring.. any updates to share?17:35
fungia change merged today that expects release notes jobs to use a "releasenotes" tox env now17:38
fungii guess keep an eye out for any related job failures, though stephenfin seemed fairly confident those should be few and far between17:38
gouthamr+1 ty17:39
clarkbResolute test nodes exist now too. But there isn't mirror content yet so they are a bit awkward to use17:39
clarkbhopefully that changes soon17:39
fungioh, and the ubuntu-ports mirrors (used for arm jobs) have been stale for about 1.5 weeks due to exceeding quota, but looks like they should be back on track any minute now17:40
gouthamr++ great, i was curious with the chatter on #opendev:opendev.org17:40
gouthamrwhen is the gerrit update planned?17:41
fungii don't think we've announced an exact date yet17:42
mharley[m]Hello, hello. It seems I was summoned. 😊17:42
mharley[m]That’s right, gouthamr. I received the instructions on how to create the channel and I’ll do that this week (my homework). 😉17:42
gouthamrack fungi 17:42
clarkbmy goal is late May early June. Probably more like early June at this rate given everything else I've been distracted by recently17:42
gouthamrgood stuff, ty 17:42
gouthamralright, let's move to 17:43
gouthamr#topic Open Discussion17:43
gouthamrmnasiadka: regarding python-tempestconf, was the decision to leave it under "openinfra" and let the existing maintainers continue there?17:44
fungii'm not aware of any urgent requirement for it to vacate that namespace17:45
gouthamrack, good to know17:45
fungimainly we wanted to retire the refstack and interop repos that are no longer used17:46
gouthamri contacted the Red Hat folks privately to test the waters. they continue to use/maintain it and don't have opinions on where it should live.. this group intersects little with "openstack-qa".. 17:46
fungii didn't include python-tempestconf with the others because i understood it was still being used and maintained by people outside its original mandate17:46
gouthamryes17:46
gouthamrthere's still some misconception that "x/" can be used... 17:47
gouthamri was going to point them to: https://review.opendev.org/c/opendev/infra-manual/+/95857117:47
gouthamrbut, that's moot if we don't need to change the namespace17:47
gouthamrfrickler: gentle nudge regarding the resolution to get project cores AC status17:49
fungigouthamr: also https://review.opendev.org/c/openstack/project-config/+/956592 "Do not allow new x/ namespace projects"17:50
gouthamrah nice17:50
gouthamrwow there are 582 already in that graveyard17:50
mnasiadkagouthamr: I think then we need a new group in Gerrit config (instead of refstack) and a new acl only for that repository? (re python-tempestconf)17:51
gouthamri'm unsure what the "group" property is used for17:52
mnasiadkafungi: are groups in project-config for gerritbot?17:54
fungi"group" when storyboard is in use defines the storyboard project group to include it in. "group" when launchpad is in use sets the corresponding project name in lp17:55
mnasiadkaah17:55
mnasiadkaSo we would need to create a new one for python-tempestconf?17:55
gouthamrhttps://storyboard.openstack.org/#!/project/91217:56
gouthamrcan't tell if the maintainers are using this, last bug there was nearly a year ago17:57
mnasiadkaOk, I have no clue how storyboard works17:57
gouthamryeah, i don't know how to find out the group there either :D17:58
mnasiadkaWell, there’s project groups on the menu bar on the left17:58
mnasiadkahttps://storyboard.openstack.org/#!/project_group/refstack17:59
gouthamroh, you're right17:59
fungiyeah, groups are just a many:many mapping sort of useful as aggregate dashboards17:59
fungia project can belong to more than one group17:59
fungior none17:59
* gouthamr wonders now how storyboard cleanup happens18:00
fungibut also the maintainers could decide to switch to doing defect and task tracking somewhere else18:00
fungistoryboard cleanup is someone asking me to delete things in the database, at the moment18:00
gouthamryeah, it's good to let them know.. or you'll just have orphaned resources18:00
gouthamrfungi: haha, okay, good to know18:01
gouthamri suspect launchpad is the same way, if everything was handed over correctly to ~openstack-administrators18:01
gouthamri wrote this and could use some reviews:18:01
gouthamr#link https://review.opendev.org/c/openstack/project-team-guide/+/987433 (Add project tracker management guide)18:01
mnasiadkaUm, so what’s the plan? Can we just set groups: to python-tempestconf and that will map to the project in storyboard?18:01
gouthamr+1 i like that approach18:02
fungiyou would just drop the groups list entirely18:02
gouthamrsorry, but we are two minutes over.. will wrap up this meeting.. please continue the chatter18:02
fungigroups in storyboard are for grouping multiple related projects18:02
gouthamrthank you all for participating18:02
gouthamr#endmeeting18:03
opendevmeetMeeting ended Tue May 12 18:03:02 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.2026-05-12-17.01.html18:03
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/tc/2026/tc.2026-05-12-17.01.txt18:03
opendevmeetLog:            https://meetings.opendev.org/meetings/tc/2026/tc.2026-05-12-17.01.log.html18:03
gouthamrfungi: ack, so ungrouped projects can exist.. because like in this case, it doesn't make sense to have a 1-1 project-group 18:03
fungiright18:03
gouthamrgood stuff, ty for doing the heavy-lifting here mnasiadka 18:04
mnasiadkaOk, I’ll drop the group and create a dedicated all for python-tempestconf18:08
mnasiadkahttps://review.opendev.org/c/openstack/project-config/+/96211618:08
mnasiadkaI added javascript unit tests in https://review.opendev.org/c/openstack/openstackdocstheme/+/987166 just like hberaud wanted, if anybody else can review it - it would be nice18:34
opendevreviewMerged openstack/openstack-manuals master: Community support - update until 2026.1 (Gazpacho)  https://review.opendev.org/c/openstack/openstack-manuals/+/98353618:35
opendevreviewOpenStack Proposal Bot proposed openstack/security-doc master: Updated from openstack-manuals  https://review.opendev.org/c/openstack/security-doc/+/98832718:38
opendevreviewMerged openstack/security-doc master: Updated from openstack-manuals  https://review.opendev.org/c/openstack/security-doc/+/98832719:02
cardoeSo not really sure where or how to best share but something else I'm driving towards is small "reference" containers if you will via OpenStack Helm.. https://opendev.org/openstack/openstack-helm-images/src/commit/977951e9969442e2945d0589beb6fd0a22d077d1/ironic/Dockerfile right now there's more to that than what I'm looking for but more work coming23:09
cardoeI'd like to eventually only install dependencies that the project calls out in their bindep.txt23:10
cardoestephenfin's work to support PEP-735 really makes the ability to do light weight stuff with uv.23:11
clarkbcardoe: that is similiar to opendev's python-builder and python-base container images23:11
cardoeWell I should check that out an collaborate there.23:11
clarkbpython-builder installs the bindep deps for building the wheels then it does an install of all the wheels to build them. Then you run python-base which installs the previously built wheels and any runtime bindep deps23:12
clarkbthe idea is to separate the build process from the actual final product to avoid polluting your containers with useless layers23:12
cardoeWe run all of our stuff in k8s so I've been looking to make a devstack like but via k8s (or really k3s / kind).23:12
clarkband there is a script called assemble that drives the magic23:13
cardoeBut exactly what you said is my goal... less and smaller layers.23:13
cardoekolla's got a lot of great machinery for deployment so that's a different target than me.23:14
clarkbhttps://opendev.org/opendev/system-config/src/branch/master/docker/python-builder/scripts/assemble you can see it in use with a non trivial project in zuul: https://opendev.org/zuul/zuul/src/branch/master/Dockerfile23:14
clarkb(note zuul does extra stuff for reasons, but you shouldn't need that if you can express what you want via bindep and python deps)23:15
cardoenice23:15
clarkbthose reasons being installing ansible to dedicated venvs and some random tools like skopeo23:16
fungicardoe: out of curiosity, how does what you want to build differ from loci's goals?23:19
fungiwhat you're describing sounds like what loci was trying to be, but since i gather that's what osh was using already i'm probably missing some subtle distinction23:20
cardoeI mean its loci23:20
fungioh! you're looking to improve loci then23:21
cardoeThey wanted to do the containers in OpenStack-helm-images instead of there.23:21
* cardoe shrugs.23:22
cardoeBut yes loci but then using that for devstack-like23:23
cardoeclarkb: where do you push the containers?23:34
fungicardoe: https://quay.io/organization/opendevorg23:43
fungiassuming you mean the container images23:44
cardoeYes23:44

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