Tuesday, 2024-10-15

opendevreviewchenker proposed openstack/governance master: Appoint Ke Chen as PTL for Watcher  https://review.opendev.org/c/openstack/governance/+/93241914:23
opendevreviewchenker proposed openstack/governance master: Appoint Ke Chen as PTL for Watcher  https://review.opendev.org/c/openstack/governance/+/93241914:27
*** bauzas_ is now known as bauzas15:24
bauzasreminder: nova meeting in 12 mins here15:48
bauzas#startmeeting nova16:00
opendevmeetMeeting started Tue Oct 15 16:00:34 2024 UTC and is due to finish in 60 minutes.  The chair is bauzas. Information about MeetBot at http://wiki.debian.org/MeetBot.16:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:00
opendevmeetThe meeting name has been set to 'nova'16:00
bauzashey folks16:00
bauzas#link https://wiki.openstack.org/wiki/Meetings/Nova#Agenda_for_next_meeting16:00
bauzaswho's around ?16:00
bauzastick tock16:01
bauzasoh shit16:02
bauzaswrong chan16:02
* bauzas facepalms16:02
bauzas#undo16:02
opendevmeetRemoving item from minutes: #link https://wiki.openstack.org/wiki/Meetings/Nova#Agenda_for_next_meeting16:02
bauzas#undo16:02
bauzas#endmeeting16:02
opendevmeetMeeting ended Tue Oct 15 16:02:57 2024 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)16:02
opendevmeetMinutes:        https://meetings.opendev.org/meetings/nova/2024/nova.2024-10-15-16.00.html16:02
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/nova/2024/nova.2024-10-15-16.00.txt16:02
opendevmeetLog:            https://meetings.opendev.org/meetings/nova/2024/nova.2024-10-15-16.00.log.html16:02
noonedeadpunk:D16:04
noonedeadpunkyeah, too early :D16:05
bauzasnah, rather too late for our nova meeting :)16:16
bauzasone IRC chan to rule them all, mwahahaha16:16
* gouthamr thought this was a takeover attempt in his absence17:00
noonedeadpunktc-members: gentle reminder that we're meeting here in an hour17:02
gmanntc-members: kuryr-kubernetes and its tempest plugin are ready to retired now https://review.opendev.org/c/openstack/kuryr-kubernetes/+/923074  https://review.opendev.org/c/openstack/governance/+/92250717:27
noonedeadpunk#startmeeting tc18:00
opendevmeetMeeting started Tue Oct 15 18:00:22 2024 UTC and is due to finish in 60 minutes.  The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot.18:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.18:00
opendevmeetThe meeting name has been set to 'tc'18:00
noonedeadpunkWelcome 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.18:00
noonedeadpunk#topic Roll Call18:00
noonedeadpunko/18:00
noonedeadpunk#chair gouthamr 18:00
opendevmeetCurrent chairs: gouthamr noonedeadpunk18:00
bauzaso/18:01
gmanno/18:01
cardoe\o18:01
noonedeadpunkthere're 3 absences that are on a record18:02
slaweqhi, do we have meeting today?18:03
noonedeadpunkwe do if we get qourum and right now on a rollcall :)18:03
slaweqahh, ok, sorry for being late18:04
gmannyou are on right time :)18:04
noonedeadpunkgtema seems to be not on the channel18:04
noonedeadpunkbut we have 5 ppl in there, so technically - it's a quorum18:05
bauzaswhen will be the DST for the US ?18:05
bauzasthis weekend ?18:05
gmannNov 3rd18:05
noonedeadpunkI was thinking about next one though18:05
noonedeadpunkbut yeah18:06
gtemaHey, I am on PTO this unavailable 18:06
noonedeadpunkack18:06
cardoeSo no quorum then?18:06
bauzasgmann: ack18:06
noonedeadpunkwe have 5 ppl exactly 18:06
gmannyeah, we have quorum 18:07
noonedeadpunkgiven slaweq has checked in :)18:07
noonedeadpunk#topic Last Week's AIs18:07
noonedeadpunkfirst action point was on gmann regarding proposing changes to increase coverage for requirements18:07
noonedeadpunkany updates on this one?18:08
gmannI proposed it but there are soem failure regarding quota which I need to debug18:08
gmann#link https://review.opendev.org/c/openstack/python-openstackclient/+/93185818:08
gmannneed to check if those are related to adding new jobs or something else18:08
gmannthis is for osc, I still need to add greande job for requirement though 18:09
noonedeadpunkit does not look intermitent to me right now18:09
noonedeadpunkas failures are exactly same in 2 jobs18:10
gmannyeah, maybe some race on quota tests from multiple jobs but need to debug it18:10
noonedeadpunkyeah, ok, but patch is around, which is good.18:11
noonedeadpunkgoing next18:11
noonedeadpunkgouthamr needed to confirm PTG sessions for TC. 18:12
noonedeadpunkand this has been done and written down in etherpad18:12
noonedeadpunkPTG is a topic on itself, so let's get the ball rolling18:12
noonedeadpunk#topic TC PTG18:12
noonedeadpunkthere are 2 sessions which were booked18:13
noonedeadpunk#link https://etherpad.opendev.org/p/oct2024-ptg-os-tc18:13
noonedeadpunkSo we do have a session on Monday, Oct 21 from 14UTC to 16 UTC18:14
noonedeadpunkI believe this is designated for interaction with community and other projects18:14
* bauzas notes18:14
noonedeadpunkAnd then also Friday, Oct 18 from 15 to 16 UTC to sum-up passed PTG18:15
noonedeadpunkptg schedule is available as usual at18:15
noonedeadpunk#link https://ptg.opendev.org/ptg.html18:15
gmannOct 18 15 to 17 UTC , 2 hrs right?18:15
gmanngot it. ++18:15
noonedeadpunkOh, yes, sorry18:15
bauzasthat's what I can see indeed18:15
noonedeadpunkmy bad, I was metioning slots :D18:16
noonedeadpunkcorrecting myself18:16
noonedeadpunkMonday, Oct 21 from 14UTC to 17 UTC18:16
noonedeadpunkFriday, Oct 18 from 15 to 17 UTC18:16
noonedeadpunkalso correct timing was mentioned in etherpad linked above18:17
noonedeadpunkthanks for correcting gmann18:17
noonedeadpunkWith that I think we should also cancel our regular IRC meeting on Tuesday next week18:17
slaweq++18:18
gmann++18:18
noonedeadpunkshould I create a poll for that or...18:18
gmannI do not think we need poll, you can just ask if any objection 18:18
bauzas++18:19
noonedeadpunkok, I assume no objections here then.18:20
noonedeadpunkI will send an email with notification about IRC meeting being cancelled next week18:20
noonedeadpunk#topic Leaderless projects18:21
noonedeadpunkFirst - Watcher. 18:22
noonedeadpunkI see they have pushed a patch as was discussed in a ML18:22
noonedeadpunk#link https://review.opendev.org/c/openstack/governance/+/93241918:22
bauzasthanks gmann for the review18:23
noonedeadpunkit's quite a new one - was pushed just today. 18:23
noonedeadpunk++ thanks indeed!18:23
noonedeadpunka good catch18:24
gmannI hope they are planning for PTG also and not blocked due to PTL appointment 18:24
noonedeadpunkAnd next project is Mistral, we still did not got any response in the patch18:24
noonedeadpunk#link https://review.opendev.org/c/openstack/governance/+/92796218:24
noonedeadpunkI think it was on me to write a ML to brng attention in case they've missed comments18:25
noonedeadpunkbut failed so far :(18:25
gmannI think there was one email about it18:26
noonedeadpunk#action noonedeadpunk to write email about Mistral PTL appointment patch needing attention18:26
bauzasack18:26
noonedeadpunkyeah, but it was before appointment patch18:26
gmannohk, I see18:26
noonedeadpunkI will bump the thread 18:26
noonedeadpunkiirc18:26
gmann++ thanks 18:26
noonedeadpunk#topic A check on gate health18:27
noonedeadpunkAnything on that?:)18:27
gmannfor greande things, enabling global venv there and there are few project failing(catching at least for projects we have job in grenade gate)18:28
gmann#link https://review.opendev.org/c/openstack/grenade/+/930507/1218:28
gmannoctavia is fixed but ironic, heat jobs failing18:28
cardoeI believe JayF 18:29
gmannI am getting less time to debug it last and this week but I will continue on this18:29
cardoetyping fail... I believe JayF had some patches pending for that for Ironic18:29
gmannohk, I will check18:29
JayFIronic has not supported global venv, ever18:29
JayFI had a patch up to support that among a fix for other things18:29
gmann#link https://review.opendev.org/c/openstack/ironic/+/93201618:29
gmannthis is change i have up but we need more change i think18:29
JayFer, whatever the change is, the per-proj-venv18:30
gmannk18:30
gmann#link https://review.opendev.org/c/openstack/ironic/+/93077618:30
gmannJayF: ^^ this one?18:30
JayFyeah, I'm trying to do two things at once with that, one of them is the piece you care about :D 18:31
gmannyeah, can you split those so that we can proceed for the grenade things 18:32
JayFI'm more wondering how, in the future, this can be planned as a migration to let folks know so it's not an emergency for ironic18:32
JayFit's only a coincidence I had this half-done18:32
JayFDid I miss a post to the mailing list about the change in grenade?18:33
gmannI have not posted yet as I wanted to see existing jobs passing first and let other project if they are impacted18:33
gmannthis is needed because greande current master setting were failing 18:33
gmann#link https://review.opendev.org/c/openstack/grenade/+/932017/118:34
bauzasyeah this is explained in the commit msg18:34
gmannand moving to venv is the one we need as first18:34
bauzasnow devstack defaults to a global venv18:34
gmannyes18:34
JayFMy question is simply how/when was that going to be communicated to other teams?18:34
JayFIs this that communication?18:35
bauzasI'd say projects need some liaison18:35
gmannI will post on ML soon but I am waiting if octavia, ironic, heat can be fixed and we know how fixes looks like which can be used a ref to other projects18:35
cardoeI think we should have at least had a ML announcement.18:35
JayFThat's basically ^ all I'm asking for, with a week or two heads up18:36
cardoeWhat if the effort to change Ironic was large an invasive and took months to complete?18:36
gmannit was in draft state and experimental to know how things working18:36
gmanncardoe: is this that big change to ironic ?18:36
JayFI'm happy to fix and understand this may not be the first domino in the chain of events, I just feel like the teacher is asking for us to turn in homework that was never assigned :D 18:37
gmannanyways I think I can post it on ML if that is blocking the ironic change to proceed18:37
cardoeIt's not blocking the change to proceed. But it's certainly not been on our radar to work on it or try to review these items.18:37
JayFI'm saying I would prefer a post to the mailing list so it doesn't instantly fall to the one or two Ironic contributors who pay attention to the TC meeting :)18:37
gmannthat is easy and less time consuming than explaining about importance of change to merge 18:37
gmannwell, I was not finished the change yet that is why did not ask for review #link https://review.opendev.org/c/openstack/ironic/+/93201618:38
gmannanyways I will post it on ML18:39
cardoeJust asking in the future post something so there's awareness. I'll try to prioritize it in the coming week.18:39
JayF016 is passing grenade, the failure is apparently unrelated18:39
JayFI'll also split the global venv part of my ironic change; that should be easy enough 18:40
cardoeI added gmann's change to the weekly prio.18:41
noonedeadpunkok, I think we've done with the topic for now18:43
noonedeadpunkmoving next18:43
noonedeadpunk#topic TC Tracker18:43
noonedeadpunkwe do have an etherpad with topics which I believe we also gonna raise during the ptg18:44
noonedeadpunk#link https://review.opendev.org/c/openstack/neutron/+/93149518:44
noonedeadpunkoops, wrong link :)18:45
noonedeadpunk#link https://etherpad.opendev.org/p/tc-2025.1-tracker18:45
bauzasheh18:45
bauzasat least you sent it to the right channel :p18:45
noonedeadpunkhehe18:45
gmann:)18:45
noonedeadpunkit was wrong buffer eventually18:45
noonedeadpunkI don't have much input on TC tracker topic to be frank18:46
bauzasgiven the outcome of the i18n session, I could lead a tracker for the translations18:46
noonedeadpunk++ that would be much appreciated18:47
noonedeadpunkI am also quite interested in translations activity, though lacking time dramatically to catch up with it18:48
bauzasditto here, but if I can help, I'll do it :)18:49
noonedeadpunkAnd I'm very lost where we are with tooling migration and if it's actually a blocker for onboarding new ppl or not18:50
cardoeI'm not sure what's left to discuss on tracker?18:50
noonedeadpunk#topic open discussion18:50
noonedeadpunkwe have around 10 minutes to raise smth that wasn't in agenda18:51
cardoeI have two semi-related items if no one else has any open items.18:51
noonedeadpunkgo on I guess :)18:51
bauzasshoot18:51
cardoeSo the first is around release blobs. clarkb will probably be better able to express this. At least I think it was clarkb was talking about it.18:52
cardoeWe've got two projects building container images and as such what's our support policy around those. Or even security posture.18:53
cardoeIf the underlying OS distro has a CVE for that what does that mean to rebuilding the container.18:53
noonedeadpunkwell, it's a good one, as eventually what I recall, projects officially produce only pip packages18:54
cardoeIt dovetails a little bit into releases being on PyPi because you could arguably "pip install nova" but that's not a recommended approach.18:54
noonedeadpunkI personally do not also agree about ^ but I'm in minority here obviously18:54
cardoeWell there's more than Python projects though. There's kolla which is building containers and loci.18:54
cardoeYou do not agree with which part?18:55
bauzasnot sure I understand the concern with release blobs18:55
noonedeadpunkthat "pip install nova" should not be recommended. as that's the only thing that nova officially delivers/produces18:55
noonedeadpunkWell, again, I would assume, that kolla does build certain layers only?18:56
noonedeadpunkas layering of container images is actually the way to distribute responsibility for CVE's, isn't it?18:57
cardoeThey build the layers atop an existing OS image and then publish those containers with Zuul to Docker Hub under the OpenStack name.18:57
noonedeadpunkso when you're pulling image it should consist of plenty things, right?18:58
bauzashttps://docs.openstack.org/nova/latest/contributor/project-scope.html#deployment-and-packaging if that helps to understand the nova deliverables18:58
cardoeMy comments are around the fact that if there is no security promise or effort to maintain this, there's no such clear policy in the places it's published. So the OpenStack name / brand is out there and associated with insecure published containers.18:58
noonedeadpunkbauzas: I think you still package Nova as python package?18:58
bauzaswell,18:59
bauzasour deliverable is technically a git commit18:59
cardoeSo that's the case for the nova project, bauzas. But kolla is taking nova and publishing a nova container under the OpenStack banner.18:59
bauzasthen the release management team ships it into a python build that's sent to pypi fwiw18:59
cardoeSimilarly for loci.18:59
noonedeadpunkso you don't care about setup.py or pyproject.toml?18:59
bauzasnoonedeadpunk: see the above link I sent19:00
bauzaswe care about the CI 19:00
bauzashow nova is productized out of the blue is not in our scope19:00
bauzas(upstream for sure)19:00
cardoeIt's not the nova project's scope, which is fine.19:01
cardoeBut I'm talking about an official OpenStack project that then further packages it.19:01
noonedeadpunkit's veeery vague border.... so like - for CI you can copy/paste content under /usr/lib/python3/site-packages technically///19:01
bauzasokay, so we're talking of deployment services19:01
bauzasnot service projects19:01
bauzasI just wanted to clarify the scope of that discussion19:01
noonedeadpunkcardoe: yeah, that is a very good point. I somehow thought though ,that containers layers could be updated separately from each other19:01
noonedeadpunkso if published container image is layered, kolla should be responsible only for their thing19:02
noonedeadpunknot for underlying ones they depend on19:02
noonedeadpunkoops, and we are overtime.19:02
cardoelayers are like git commits. They depend on the layer under them.19:02
noonedeadpunkI suggest to add that discussion to the PTG, as it is totally a valid one from my prespective19:02
noonedeadpunk#endmeeting19:03
opendevmeetMeeting ended Tue Oct 15 19:03:07 2024 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)19:03
opendevmeetMinutes:        https://meetings.opendev.org/meetings/tc/2024/tc.2024-10-15-18.00.html19:03
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/tc/2024/tc.2024-10-15-18.00.txt19:03
opendevmeetLog:            https://meetings.opendev.org/meetings/tc/2024/tc.2024-10-15-18.00.log.html19:03
slaweqo/19:03
noonedeadpunkgood morning :)19:03
bauzasthanks19:03
noonedeadpunkthanks everyone for attending!19:03
gmannquick one to review kuryr-kubernetes repo content removal https://review.opendev.org/q/topic:%22retire-kuryr-kubernetes%22+status:open19:04
gmannif 2nd reviewer can check and merge it and we can proceed on governance change19:04
gmannthanks19:04
noonedeadpunkdone19:05
cardoedo it again :(19:06
gmannthanks again19:06
noonedeadpunkcardoe: ok, yeah, I was somehow under full impression that it's way easier to swap update container as interally was promoted a lot some way of doing so. But reading docs now, I feel being wrong19:09
cardoeSo the two projects provide scripts to repackage your own.19:09
noonedeadpunkeventually, I think we have kinda simmilar thing with ansible collections and terraform provider19:10
cardoeUnfortunately loci/openstack-helm-images has some hardcoded IPs and docker configuration that's locked away in some openinfra zuul magic repo and makes doing so impossible.19:10
noonedeadpunkwhere thinkgs being published under openstack name as a SIG19:10
cardoeSo ya either use the provided ones or write your own build mechanism like vexxhost has.19:11
cardoeWhich interestingly uses zuul.19:11
noonedeadpunkwhich does fall of governance, and openstack-infra does not have access to these namespaces, afaik19:11
noonedeadpunkthey are managed by indviduals who are member of sig19:12
cardoeBut my concern here is that things are published at https://tarballs.openstack.org or docker hub or quay (the later two using the OpenStack logo)19:12
cardoeSo it seems "semi-official"19:12
noonedeadpunkonly https://tarballs.openstack.org  does kinda19:12
noonedeadpunkdocker hub or quay for me very much alike to galaxy.ansible.com19:13
cardoeTo a random person on the street who Google searches "OpenStack nova container" and lands at Docker Hub or Quay at an Organization called "OpenStack" with the official OpenStack logo.... that's going to seem official to them.19:13
noonedeadpunkI'm not saying you're wrong19:14
cardoeAnd when they have a bad experience with it. They're gonna say "openstack sucks!"19:14
noonedeadpunkI'm not sure there's a good solution to that though. Except to move ready images publishing to some kind of SIG 19:15
noonedeadpunkor be more explicit that OpenStack owns only things on pypi.org which would also corner any rust initiatives, for instance19:16
cardoeWe've got an official Rust initiative?19:17
noonedeadpunkthere're talks around client-side of things at least19:17
cardoeI'd certainly be interested if you know those folks.19:18
noonedeadpunkyou should talk to gtema then19:18
cardoeMy other topic was around projects that use tooling but give off bad user experiences because they fly in the face of those tools best practices or common conventions.19:21
cardoeSo interestingly crates.io has "openstack" and "openstack_sdk". The former is owned by dtantsur and the later is owned by gtema. Unfortunate there was no collaboration to have just 1.19:22
noonedeadpunk"bad user experiences" can be hardly measured, imo. and "best practices or common conventions" are usually unrelated to UX, as end-users hardly care about backend style practises19:23
noonedeadpunkand neither of them are official despite having `openstack` on them19:24
cardoeNot saying they're official just making a note of the possible collab miss.19:25
noonedeadpunkI kinda wonder what olso-policy is there for19:26
cardoeMy tooling best practices comment can be best summed up by a project making a release 0.3.16 back in July and publishing it to tarballs.openstack.org But that depends on a number of other dependencies, some are internal to the project (living in other repos on opendev.org). It turns out that the release is rebuilt on every commit to the child dependencies as well. That rebuilt package then overwrites the tarball.19:26
cardoeThe result is a change in any of the child dependencies gets rolled out into the latest release.19:27
cardoeThe response to the bug? The PTL tells you the project is a framework. If you care about stable releases then you should build them yourselves.19:27
cardoeBut that's once again not expressed in the project's documentation.19:28
cardoeAnd all the docs suggest using tarballs.openstack.org19:28
noonedeadpunkMy personal prespective on that is such kind of things are to be sorted with the project19:28
noonedeadpunkAnd that's why there're elections of PTLs19:29
noonedeadpunkSo I guess I would try to communicate with the community around the project and raise their awareness as well. And propose changes to approaches, or run for PTL of that's in my direct zone of interest.19:32
gouthamrnoonedeadpunk++ thank you for chairing :)19:37
noonedeadpunkgouthamr did amazing job of prepearing everything, so it was a pleasure 19:37
*** elodilles is now known as elodilles_pto19:41
spotz[m]Shoot forgot to put me as absent as I’m at NA20:31
*** bauzas_ is now known as bauzas21:51
*** bauzas- is now known as bauzas22:04
*** bauzas- is now known as bauzas22:12
*** bauzas_ is now known as bauzas23:33
*** bauzas_ is now known as bauzas23:44

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!