Tuesday, 2026-11-17

cardoeHello all in about 57 minutes we'll have the TC meeting.16:04
cardoetc-members: in about 56 minutes we'll have our weekly IRC meeting.16:04
cardoethe agenda is here: https://etherpad.opendev.org/p/openstack-technical-committee-meeting16:04
dansmithI have a double conflict this week, FYI. Hopefully this is not a trend an I'll have zero next week and not three16:37
fungiso not a triple-conflict, should be fine right? ;)16:40
dansmithis that like "three wrongs do actually make a right" ?16:41
cardoe#startmeeting tc17:00
opendevmeetMeeting started Tue Nov 17 17:00:43 2026 UTC and is due to finish in 60 minutes.  The chair is cardoe. 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
cardoeWelcome 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
cardoeToday's meeting agenda can be found at https://etherpad.opendev.org/p/openstack-technical-committee-meeting17:01
cardoe#topic Roll Call17:01
gtemao/17:01
cardoeCourtesy ping: noonedeadpunk, frickler, ChristianBerendt[m]17:02
cardoenoted absence: s p o t z, g o u t h a m r, d a n s m i t h, m n a s i a d k a17:02
cardoeToday is also OpenInfra Day North America (Chicago). Attendance may be thin.17:03
cardoewell looks like it might just be you and me gtema.17:04
* gouthamr is watching this chat, but is meant to get on the podium in a few mins17:04
fungibreak a leg!17:04
gtemayey cardoe, we'll have fun. The community "flu" spreads everywhere17:04
cardoe#topic Announcements17:04
cardoeElections kickoff (2027.2)17:05
cardoe#link https://review.opendev.org/c/openstack/election/+/101459217:05
cardoeThis will setup the candidate directories17:05
* gouthamr spoke to ianychoi here - he's going to continue as an election official for 2027.217:05
cardoe#link https://review.opendev.org/c/openstack/governance/+/101485317:05
cardoeThis drops extra-acs and re-proposes the election officials17:06
gouthamr^ he'll also look at the election repo patch and help get it merged17:06
cardoeGood to hear about ianychoi continuing17:06
cardoeslaweq isn't in this channel but will ping as well17:06
gouthamr++17:06
gouthamrwe could onboard a volunteer or two, if there are any17:07
cardoeAbsolutely, it is open to any interested volunteers17:07
cardoeJust speak up! :)17:07
cardoeOpenInfra Foundation Board individual member representative nominations open; board elections Jan 11–15, 202717:08
cardoeSo if you have interest in the foundation board this is a great opportunity.17:08
cardoe#topic Release: R-18 this week17:08
cardoeIndri-1 milestone is Thursday Nov 19; also Cinder Spec Freeze week17:08
cardoeI've seen the release team has a lot of pending items.17:09
gtemayeah, one of those days you leave laptop for 5 minutes and when you come back there are 50+ emails in the inbox17:10
fricklerlate \o sorry17:10
cardoefungi, clarkb: I know there's some pending tooling patches which I suspect you two might want to weigh in on17:10
cardoeNext week (R-16 window Nov 30–Dec 4): 2026.2 "Hibiscus" cycle-trailing release deadline (Dec 3), Cinder Midcycle #1.17:10
fungipointers welcome, we're drowning in notifications too so easy to miss things17:10
fungiif you mean in the releases repo, yes those are already on my agenda17:11
cardoeyeah in the releases repo17:11
cardoeI can link them if it'd be helpful?17:11
funginah that's okay, since i'm also a release manager i'm already aware of which ones they are17:12
cardoeThere's also one in zuul/zuul.17:12
cardoeThis is for dependency group support.17:12
fungioh, that would be a good one to call out then17:12
fungiif it's the one stephenfin has been working on for docs jobs, i think it's nearly there17:12
cardoe#link https://review.opendev.org/c/zuul/zuul-jobs/+/101494817:12
cardoeThat's the one it's for docs jobs.17:13
fungiyep, good, one i already was aware of at least. thanks!17:13
cardoeOverall it should allow us to modernize a lot of our build and get rid of setup.py17:14
cardoeAnd package all the dependencies in one place, pyproject.toml17:14
cardoeAt least that's how it's been pitched to me.17:14
cardoeMoving on.17:14
cardoe#topic Deprecate networking-sfc17:15
cardoe#link https://review.opendev.org/c/openstack/governance/+/101496817:15
cardoeThis has now merged and it's deprecated17:15
cardoe#topic Last Week's Action Items17:15
cardoeAmending of the AI policy interpretation resolution17:16
cardoe#link https://review.opendev.org/c/openstack/governance/+/101208017:16
cardoeIt's had some feedback and no votes17:17
cardoeSo it'll likely take another spin17:17
cardoe#topic Propose goal-process change: require group champions17:17
cardoeThis is still opened and we'll need to update this goal-process. Overall there wasn't voices against this approach.17:18
ChristianBerendt[m]o/ Sorry, I lost track of time.17:18
gtemacardoe - I have somehow missed the essence: what is it about?17:19
cardoeIt's okay. I'm moving quick. I'm happy to jump back to any previous topic.17:19
cardoegtema: The concern was that single person goals often struggle. It might be a better approach to require a group.17:19
cardoeIt serves to alleviate the bus factor as well.17:20
gtemagot it, not a fan of it to be honest17:20
cardoeokay, would you like to voice the concerns?17:20
gtemawe do not have 'too many' people, so every time the statement of bus factor comes I read it as - we don't want to give possibility to do the work even to those few left willing to do the work17:21
gtemathis is slowing the progress rather than giving any certainity17:22
gtemathe same was stated towards my proposals, so I see it really as a "regression"17:23
fungii suppose an alternative could be to avoid approving goals or cancel them when it looks like there may not be sufficient capacity to drive them forward17:23
cardoeIt'd be more a of popup team that people can join it17:23
cardoeSure that could work as well.17:23
fungiit took a relatively long time for the tc to decide to move the fips goal from accepted back to proposed17:24
gtemawe should be more free in simply letting things die out naturally than trying to artificially keeping them alive. The natural thing of the evolution is a fast try/fail/try/success cycle17:24
gtemaand so if the goal is not moving - feel free to stop it. It will not stop the work, but the official process (whatever it is actually good for)17:25
cardoeWe would need to be quicker at saying they failed or replacing a champion that's not present.17:25
cardoeWell it sounds like we've got some feedback on how this should move forward.17:26
fungiwhat it's (in theory) good for is helping teams affected by the goal understand that the tc has interpreted community interest in prioritizing that activity across openstack. so once the tc no longer sees a priority attached to a goal, then yes it's misleading or even harmful to give individual team leadership the impression it's still a priority17:26
gtemayes, definitely. Cause when you require a group of people to exist to even start you will initially kill-off multiple things where a single person is really dedicated to it17:26
cardoeIt wasn't a group.17:28
cardoeIt was a pop-up team that can add members17:28
cardoeJust cause I'm doing bad on time I'm gonna move us forward17:29
cardoe#topic Security SIG: sunset or integrate with VMT17:29
cardoeThere was no firm conclusions on this one17:29
cardoeGiven the folks that were discussing it last time aren't here today I'm tempted to punt it to next week17:30
gtemayupp, I do not have a strong opinion on the topic17:30
cardoe#topic migrate-to-privsep goal17:30
cardoewe didn't get an update from ralonsoh last week and none this week17:30
cardoeThis goal is what prompted the pop-up17:31
gtemait's so bureaucracy - different terms with different requirements for the same thing17:32
cardoe#topic Archive the First Contact SIG17:32
cardoe#link https://review.opendev.org/c/openstack/governance/+/101485517:32
cardoeNeeds some more eyes and is eligible to merge but needs more votes17:33
cardoeSorry it's got 4 votes so it has enough17:33
cardoeso tc-members please vote on the above if you can17:34
cardoe#topic Propose resolution: stable branches of retired deliverables go EOL17:34
cardoefungi: this was yours17:34
gtemajust left my +117:34
fungiwe discussed this last week17:34
fungii haven't drafted the amendment yet17:34
fungiit's on my to do list17:35
cardoeno rush just going though the items17:35
cardoe#topic TC Tracker Items17:35
cardoehttps://review.opendev.org/dashboard/?title=Open+Reviews+for+TC+Repos&foreach=is:open+-is:wip&Governance=(project:openstack/governance+OR+project:openstack/governance-website+OR+project:openstack/governance-sigs)&Election=(project:openstack/election)17:36
cardoeThere's a number of items there that need feedback still17:36
cardoeeligible to merge today are:17:37
cardoe#link https://review.opendev.org/c/openstack/governance/+/101485317:37
cardoe#link https://review.opendev.org/c/openstack/governance/+/101485417:37
cardoe#link https://review.opendev.org/c/openstack/governance/+/101485517:37
cardoeWhich cover expired extra-acs, reproposing election officials, adding SIG chairs as extra-acs, dropping First Contact SIG17:38
cardoe#link https://review.opendev.org/c/openstack/governance/+/99865217:38
cardoeRust for new IAM which needs votes and feedback17:38
cardoe#link https://review.opendev.org/c/openstack/governance/+/99276517:38
gtemayeah cardoe - including from you17:38
cardoeTested runtimes for clients vs libraries17:38
cardoegtema: yeah I do I had a whole write up pending17:39
cardoeI'll leave the rest unlinked.17:39
cardoe#topic Gate Health17:39
cardoeclarkb: anything to share?17:40
gtemasounds frightening17:40
fungii know we're tracking some network disconnects that are continuing to adversely impact test nodes in rackspace flex dfw17:40
fungiand glance image corruption in rackspace flex sjc causing some jobs to choke on corrupted git repository caches17:41
clarkbwethought it was happening only at the beginning of jobs and causing them to be retried relatively quickly. But now have some evidence that it can happen later in jobs which causes them to fail so I've got a change up to disable the region17:41
clarkbI did send email to rackspace about this last week after it ws suggested by cardoe so hopefully there will be some sort of resolution17:41
cardoeyeah let me ping those folks again17:42
clarkbAnd we started winding down arm64 package mirroring to free space on the openafs fileservers17:42
fungiwe also eliminated arm64 package mirroring to free up space in afs, which has caused some disruption for arm jobs (though those should be non-voting)17:42
cardoeIf only those regions were on my platform17:42
clarkbfungi: and some of those disruptions have been fixed by having bindep do an apt-get update now17:42
cardoeI've noticed that some kolla / bifrost people have started to push arm64 things.17:42
clarkbbut I'm sure there are other places without bindep that need to run apt-get update to get newer indexes without the mirror setup doing it for them17:43
fungimnasiadka also pushed a change to drop debug symbol packages from our debian package mirrors, and has another up to stop mirroring esoteric kernel config packages, both of which should be non-impacting and net us a bunch more space17:43
fungi(ubuntu distributes a shocking variety of variously-configured kernel packages almost none of which make sense on virtual machines)17:44
fungiand the uptick in security fixes for open source projects across the board is resulting in us taking more frequent unplanned gerrit and zuul maintenance windows to roll fixes in on short notice17:45
clarkbCeph is apparently moving their distributed artifacts (UBI based container images?) to require x86-64-v317:45
clarkbwhich our largest pool of test nodes does not support17:45
clarkbfungi: yes and more generally just being swamped under related activities17:45
clarkbits hard to get anything done when you're going from one fire to the next17:46
fungiright, makes us a lot less responsive to requests for help when we're buried under security-related emergencies17:46
frickleryeah, maybe it was a bad idea for devstack-ceph-plugins to move to containers in an idea to be more portable17:46
fungialso there's a big gerrit upgrade on the horizon in a few weeks (3.13->3.14)17:46
gtemaspeaking of OS, kernel, etc: I spent 2 days chasing weird failures in github ci, which ended up being caused by switch to 26.04 by default and (depending on kernel) started facing https://github.com/moby/moby/issues/54099 - whenever ubuntu 26.04 is used with docker this might be relevant. You may want to keep it as an open tab since side effects will very unlikely hint you to the proper direction17:47
clarkbwe use podman now and already arn into problems iwth apparmor signal handling rules there on noble17:48
gtemalol17:48
clarkbinteresting that the kernel is playing a role now17:48
fungitrading one set of problems for another, yep17:48
clarkb`sudo podman kill -s HUP container` -> broken `sudo kill -s HUP $PID` -> works17:50
clarkbits the sort of security by obscurity thing that makes people feel happy I guess17:50
gtemaindeed17:51
cardoeWell I think we've transitioned to some open discussion items since this is past the gate stuff17:51
cardoeSoo lemme do that17:51
cardoe#topic Open Discussion17:51
gtemanothing from me17:52
cardoeAny items folks want to bring up?17:52
fungican't remember if i mentioned last week, but sqlparse (an openstack dependency) announced they're stopping development because they can't keep up with the rising volume of security bug reports17:52
cardoethe security fire hose will likely continue to keep blasting us all in the face until morale improved unfortunately.17:52
cardoeor Sam Altman gets all that RAM that he wants17:52
fungiwe'll probably have to address the situation with depending on sqlparse relatively soon one way or another17:53
cardoeoh yes good point fungi 17:53
cardoeWe should likely look at retiring its usge.17:53
fricklerafaict it is only a transitive dependency? so the task would be a) to find out which libs depend on it and b) nag their maintainers if needed. or replace the usage of those libs if possible17:54
fungiright, not straightforward, needs to be on someone's radar though17:54
cardoehmm okay17:55
fungiand be on the lookout for that starting to happen more often in the coming days, probably17:55
fungithere's well over 500 python package dependencies alone tracked in upper constraints17:56
gtemayes, I see this now often also in my other projects17:56
cardoeyeah I used to offer to others to take over my open source projects that weren't something I used anymore.17:57
cardoeBut with people hijacking repos for nefarious purposes and the such, I'm instead shutting them down myself and letting others make a fork under a new name.17:57
cardoeI had a number of really early Rust crates for example.17:58
fungii was gonna say, jia tan is volunteering to take over maintaining your more popular projects17:58
cardoeYep I bet they are.17:58
cardoeI'm going to close the meeting but the discussion can continue.17:59
cardoeThanks all for participating.17:59
cardoe#endmeeting17:59
opendevmeetMeeting ended Tue Nov 17 17:59:22 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)17:59
opendevmeetMinutes:        https://meetings.opendev.org/meetings/tc/2026/tc.2026-11-17-17.00.html17:59
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/tc/2026/tc.2026-11-17-17.00.txt17:59
opendevmeetLog:            https://meetings.opendev.org/meetings/tc/2026/tc.2026-11-17-17.00.log.html17:59
clarkbit is an interesting social experiment that we drop a ton of work on people out of the blue, some of it critical for systems we all rely on day to day, and it comes with little more than a good luck have fun!17:59
fungiclarkb: by the way, good luck and have fun! ;)17:59
mnasiadkaGood luck, have fun, don’t die?19:01
cardoeclarkb: I am told dfw3 networking should be sorted tomorrow night21:02
clarkbcardoe: thanks for the heads up. We can probably avoid turning that region off for now then asi I suspect we can manage that long21:04
fungithanks cardoe! i'm going to credit you with fixing it ;)21:05
cardoeI just did the flogging21:05
fungicredit you with the flogging21:06

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