18:00:47 <gouthamr> #startmeeting tc 18:00:47 <opendevmeet> Meeting started Tue Oct 29 18:00:47 2024 UTC and is due to finish in 60 minutes. The chair is gouthamr. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:47 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:00:47 <opendevmeet> The meeting name has been set to 'tc' 18:01:01 <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 18:01:05 <gouthamr> Today's meeting agenda can be found at https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee 18:01:09 <gouthamr> #topic Roll Call 18:01:10 <noonedeadpunk> o/ 18:01:11 <gmann> o/ 18:01:13 <bauzas> \o 18:01:25 <gtema> O/ 18:01:30 <frickler> \o 18:02:03 <slaweq> o/ 18:02:10 <spotz[m]> o/ 18:02:49 <gouthamr> courtesy-ping: cardoe 18:04:51 <gouthamr> let's get started.. 18:05:05 <gouthamr> #topic Past Weeks' AIs 18:05:06 <cardoe> sorry I'm on. 18:05:14 <gouthamr> hey there, cardoe 18:05:39 <gouthamr> Debug quota test failures in the requirements patch and continue work on the grenade-related issues (gmann): 18:06:04 <gouthamr> gmann: seemed like separate issues that you brought up here in the past weeks 18:06:08 <gmann> greande job is fixed now and some project failing it also have fix up 18:06:31 <gmann> quota test failures is in sdk job not in requirement. 18:06:44 <gmann> I still did not get time to debug those, I will look into that this week 18:07:01 <gouthamr> #link https://review.opendev.org/q/topic:%22grenade-venv%22 (Grenade job changes) 18:07:57 <gouthamr> gmann: ty, any gerrit links we should track for that second item? 18:08:07 <gmann> yeah 18:08:11 <gmann> #link https://review.opendev.org/c/openstack/python-openstackclient/+/931858 18:08:13 <gmann> this one ^^ 18:08:39 <gouthamr> thank you! we'll follow up with this next week 18:08:53 <gouthamr> next up, Lead the tracker related to translations (bauzas) 18:09:04 <bauzas> hmmm, sure 18:09:06 <gouthamr> ^ this was sorta actioned during the PTG week 18:09:22 <gouthamr> and probably a work item to track through the release/ 18:09:25 <bauzas> nothing really important to discuss as an AI 18:09:44 <gouthamr> bauzas: ack; worth tracking with the TC tracker? 18:09:45 <gouthamr> #link https://etherpad.opendev.org/p/tc-2025.1-tracker (Technical Committee activity tracker - 2025.1) 18:09:53 <bauzas> we started to discuss with ian and seangsoo about how to work on the AI translation 18:10:35 <bauzas> https://etherpad.opendev.org/p/oct2024-ptg-i18n#L47 18:10:40 <frickler> maybe worth noting that translation updates from zanata are completely broken now 18:11:01 <bauzas> there was just a concern about legal issues 18:11:12 <frickler> since reqs dropped py3.8 pins and thus installing projects on bionic no longer works 18:12:02 <fungi> and the (abandoned) zanata cli tools need a jdk old enough that it was last available on bionic 18:12:08 <bauzas> we said that maybe just a documentation top paragrath would say "it's an automatically translated" 18:12:36 <clarkb> I was thinking about that and maybe a container image for the zanata client would be an "easy" thing to solve that problem 18:12:44 <clarkb> then you can have java 8 + zanata client wherever you are 18:12:52 <cardoe> that would make sense clarkb 18:13:25 <fungi> though also sounds like a good bit of work when people could be spending time finishing the weblate transition instead 18:13:49 <clarkb> ya thats the other thing to keep in mind I guess. zanata is still very dead end and we should move away from it 18:13:52 <gouthamr> is this the job? https://zuul.opendev.org/t/openstack/builds?job_name=propose-translation-update 18:14:06 <bauzas> honestly I don't know 18:14:11 <frickler> gouthamr: yes 18:14:31 <bauzas> that's a different question 18:14:52 <bauzas> (I just discussed with ian for the automatically translations for the service docs) 18:16:04 <gouthamr> it is a different question 18:16:17 <gouthamr> bauzas: can we deal with the legal question as a topic? 18:16:55 <bauzas> gouthamr: maybe we should ask the Foundation for that 18:16:56 <gouthamr> regarding the job failure, I think the i18n team needs some infra help - they mentioned that they lacked the zuul expertise to maintain these translation jobs.. if zanata doesn't work on anything other than bionic, i think we've hit the first schism 18:18:24 <fungi> in that case, it would be good to know what questions they have about automation 18:18:33 <gouthamr> "For the issue, I18n SIG is asking help to upgrade propose-translation-update job with ubuntu-bionic to newer version of Python" 18:18:55 <gouthamr> L44 of the etherpad bauzas linked 18:19:16 <frickler> so that would be the container solution that clarkb suggested. but I don't think the infra team has much time to help with that 18:19:24 <fungi> ah, yes, the counter-suggestion is to finish up the weblate migration instead 18:19:37 <bauzas> I wasn't there when they discussed this 18:19:37 <gouthamr> ^ my hope as well 18:19:38 <clarkb> and I don't think it needs much zuul expertise to switch the jobs 18:19:46 <clarkb> its copy paste and change out the tools inside the jobs 18:20:06 <fungi> since zanata is exactly the reason the jobs are breaking when moving to newer python/platform 18:20:13 <clarkb> I tried to stress this in the calls we had before that it really should be straightforward someone just needs to decide they are doing it and spend a little time understanding the jobs enough to replace the tools in copies of them 18:20:35 <gouthamr> i think, we can live without translations (for hopefully a short time) until the weblate infra is fixed up, and there are new jobs to propose translation updates from weblate 18:20:36 <gmann> yeah, this is not just job migration 18:21:05 <fungi> don't necessarily even need new jobs, just replace the zanata cli commands with weblate cli commands 18:22:16 <gmann> weblate is also licensed, hope we are all good on their subscription payment this year also? Last time it was paid by the foundation but not sure if that was for more than a year and need to renew or not ? 18:22:40 <fungi> i think it was set up to renew anually 18:22:51 <gmann> ok, thanks 18:23:04 <fungi> have the i18n folks indicated that weblate stopped working? 18:23:30 <bauzas> I don't know 18:23:33 <gouthamr> they discussed "Weblate with openinfraid authentication issue" - which they note is now resolved 18:23:36 <gmann> no, I have not heard that but just want to check if that part is ok or not 18:24:45 <gouthamr> don't want to spend the whole hour discussing this; i've alerted ianychoi and seongsoo to peek into this channel.. its possible they could start a ML discussion as well to follow up.. 18:25:02 <gmann> ML is better due to TZ 18:25:10 <gouthamr> +1 18:25:15 <bauzas> +1 18:25:21 <gmann> last time we synced on all these issue in ML and brian from TC leading the discussion 18:25:29 <gmann> that was productive and fast 18:25:51 <fungi> i follow the i18n ml as well as the i18n topic on openstack-discuss and haven't seen these issues brought up 18:26:21 <gouthamr> ++ bauzas volunteered or was voluntold to play the TC liaison role here, i cannot tell, but, its very much appreciated 18:26:32 <gmann> bauzas: ++ thanks 18:26:45 <gouthamr> lets move on to the next AI: 18:26:45 <gouthamr> Leaderless projects, next steps 18:27:07 <gouthamr> we have a couple of pending changes on the governance repository that still need eyes 18:27:37 <gouthamr> #link https://review.opendev.org/c/openstack/governance/+/927962 (Adding Axel Vanzaghi as PTL for Mistral) 18:27:57 <gouthamr> #link https://review.opendev.org/c/openstack/governance/+/933018 (Add watcher DPL for Epoxy) 18:28:17 <gouthamr> thanks for adding your RC votes and comments, but if you haven't done so yet, please do 18:28:43 <gouthamr> i'd like to land both these changes this week since they have spent sufficient time on review 18:29:13 <gmann> ++, done 18:29:25 <gouthamr> these do complete our "leaderless" tracking post the 2025.1 election 18:29:33 <gouthamr> #link https://etherpad.opendev.org/p/2025.1-leaderless (2025.1 Leaderless projects) 18:29:34 <slaweq> I just added my RC +1 18:30:22 <gouthamr> thank you 18:30:37 <gouthamr> alright that's all the AIs I was tracking 18:30:55 <gouthamr> did any of you have any other items that you were working on that were actions from past meetings? 18:30:59 <bauzas> I think we can merge those, we have quorum right? 18:31:13 <bauzas> (when looking at r-c label) 18:31:25 <bauzas> (r-*v*) 18:31:26 <gouthamr> bauzas: yes, but we wait until 7 days after the last change.. 18:31:40 <bauzas> ack thanks for explaining 18:32:08 <gouthamr> bauzas: as a minimum :) i've learned to wait longer in case something needs more eyes, as JayF gmann and others used to.. 18:32:51 <gouthamr> we discussed dropping past TC chairs from the tech-committee-chair group (sad panda emoji) 18:33:00 <gmann> I think bauzas should know about our magic script which tell us about minimum (just minimum) requirements to merge the changes. 18:33:13 <gouthamr> soooo - the side effect is that i can tag the tc in changes taht need attention :) 18:33:24 * gouthamr please don't ignore these mass reviewer tags 18:33:35 <bauzas> gmann: I did read the TC docs but I forgot about when merging :) 18:33:37 <gmann> bauzas: https://github.com/openstack/governance/blob/master/tools/check_review_status.py 18:33:41 <gmann> cool 18:34:02 <gmann> ++ on tech-committee-chair cleanup 18:34:10 <gouthamr> sounds like no further AIs to catch up on.. 18:34:11 <gouthamr> lets move on 18:34:12 <bauzas> gmann: thanks for helping ;) 18:34:15 <gouthamr> #topic Goal Progress: Migration to noble 18:34:47 <gmann> not much update but I started working on this, created the plan and sent on ML 18:34:51 <gmann> #link https://etherpad.opendev.org/p/migrate-to-noble 18:35:06 <gouthamr> i have a minor-ish concern regarding these changes 18:35:15 <gmann> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/JOMDY26TCW7OX3NXRGOYQCIDXNNJ4E25/ 18:35:30 <gmann> sure 18:35:34 <gouthamr> i really don't like keeping a test job variant on Jammy 18:35:40 <gouthamr> gmann explained the reasoning here: 18:35:43 <gouthamr> #link #link https://review.opendev.org/c/openstack/manila-tempest-plugin/+/932956 18:36:16 <gouthamr> but i believe that we'd like to keep some testing on Jammy for trunk/main branches because it would allow python3.9 integration testing to continue 18:36:22 <gmann> just to note, this is requirement of this section in testing runtime 18:36:24 <gmann> #link https://governance.openstack.org/tc/reference/runtimes/2025.1.html#additional-testing-for-a-smooth-upgrade 18:37:10 <bauzas> yeah we need to continue to support Jammy until next release at least 18:37:22 <gmann> yeah, at least till next release 18:37:24 <bauzas> because $slurp 18:37:37 <gouthamr> until? 18:37:42 <gmann> what I need to update is to run jammy job with python 3.9 18:38:02 <bauzas> until when we change the PTI values for F, G or some other cycle :) 18:38:04 <spotz[m]> So upgrade 3.8 to 3.9 but stay on Jammy? 18:38:08 <gmann> gouthamr: in next cycle, we can drop. this will help to have slurp 2024.1 -> 2025.1 migration 18:38:42 <gmann> gouthamr: I updated in nova change where we can run jammy on py3.9 so that we do not need to keep up the things for py3.8 18:38:44 <gouthamr> gmann: wouldn't this mean the py39/jammy job will continue to run on stable/2025.1 until we sunset that branch? 18:38:48 <clarkb> I wrote down the migration path for centos 9 to centos 10 in an mailing list response and I think ti is similar for jammy to noble 18:39:10 <gmann> gouthamr: yes, even in stable/2025.1 also but in 2025.2, we do not need to run, we will keep only noble jobs 18:39:22 <gmann> clarkb: yes 18:39:33 <gmann> we did same in past also when we moved from bionic to jammy 18:39:47 <bauzas> https://devguide.python.org/versions/ says that 3.9 will be EOL next year 18:39:58 <gmann> #link https://governance.openstack.org/tc/reference/runtimes/2023.1.html#additional-testing-for-smooth-upgrade 18:40:19 <gmann> bauzas: yeah, hopefully we will be good to move to py3.10 as min by then 18:40:47 <noonedeadpunk> isn't jammy 3,10 by default? 18:40:53 <bauzas> I just wanted to say we have ~ 1 cycle (ie. G as the last one) for 3.9 18:41:09 <noonedeadpunk> and it's el9 that's shipped with 3.9 iirc 18:41:11 <bauzas> but if we can drop it for F, that's cool 18:41:20 <gouthamr> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/FJO4GPM25C4UKHDBOQBEIG7YVWTOWL4H/#FJO4GPM25C4UKHDBOQBEIG7YVWTOWL4H (suggested upgrade path for centos/rhel/rocky/alma by clarkb) 18:41:59 <noonedeadpunk> have someone looked into devstack for el10 already? 18:42:02 <gmann> noonedeadpunk: ah right. it is python 3.10. I was wrong on py3.8 things. 18:42:12 <noonedeadpunk> as I can imagine modular libvirt could be slightly different 18:42:23 <gmann> so we can run jappmy on either py3.10 or onb py3.9 which is our min python version 18:42:29 <gmann> jammy 18:42:53 <frickler> one can also update and use newer python on el9. kolla does this due to ansible reqs 18:43:11 <frickler> so not 100% necessary to wait for el10 18:43:36 <noonedeadpunk> well, there's difference of ansible host and rest 18:43:52 <bauzas> I don't know what the el9 plan post 3.9 EOL 18:44:01 <noonedeadpunk> not sure what specifically kolla does, in osa we use 3.11 python for EL only as "deploy" host version 18:44:29 <frickler> I think we use it for services, too, but I need to doublecheck 18:44:38 <noonedeadpunk> the problem with non-default version, is that there're no python bindings for things, like ceph, libvirt, selinux, etc 18:44:53 <noonedeadpunk> so kinda makes things tricky 18:45:03 <gouthamr> i think my initial question was well answered through this discussion, thanks everyone.. i do wish we highlighted our upgrade path somewhere 18:45:05 <frickler> hmm, then maybe I misunderstood 18:45:17 <gmann> gouthamr: is your concern about Jammy job resolved? This is implementation of what we have defined in 2025.1 testing runtime. 18:45:26 <clarkb> gouthamr: yes I called that out with the relese team after sending that email I think 18:45:52 <gouthamr> would a "Reasoning" section under the tested run times help? 18:45:55 <clarkb> I also updated a diagram explaining slurp in release docs. If even we are confused about it then we're not documenting it properly and I suspect part of the issue is that text is not sufficient for communicating these ideas clearly 18:45:59 <clarkb> need more illustrations 18:46:29 <gmann> gouthamr: also I tried to described it in this section https://governance.openstack.org/tc/reference/project-testing-interface.html#upgrade-testing 18:46:31 <gouthamr> #link https://releases.openstack.org/#releases-with-skip-level-upgrade-release-process-slurp (Release upgrade illustration) 18:46:31 <gmann> #link https://governance.openstack.org/tc/reference/project-testing-interface.html#upgrade-testing 18:46:41 <gmann> if anything missing there, feel free to update 18:47:19 <gmann> this subsection 18:47:21 <gmann> #link https://governance.openstack.org/tc/reference/project-testing-interface.html#extending-support-and-testing-for-release-with-the-newer-disto-version 18:47:21 <gouthamr> ++ thank you; i'll read it 18:47:51 <gmann> that is where we started building those doc when SLURP things came and when we bump the distro version 18:48:17 <gmann> for smooth upgrade, testing the old distro one job for a one more cycle is migration path we can provide 18:48:31 <clarkb> as a side note the container based deployments actually care about this a lot less 18:48:43 <gouthamr> containers need OS too :D 18:48:44 <clarkb> because you can update 2024.1 to 2025.1 containers running the latest python version for each 18:48:49 <clarkb> gouthamr: yes but you don't need overlap 18:49:03 <clarkb> because you can replace 2024.1 python3.9 with 2025.1 python3.12 in the same step 18:49:16 <gouthamr> yes 18:49:25 <noonedeadpunk> clarkb: you do if you don't use containers 18:49:34 <clarkb> noonedeadpunk: correct I'm explaining the difference 18:49:41 <noonedeadpunk> or we're saying that openstack outside of docker images is unsupported? 18:49:45 <clarkb> you need overlap if not using containers. If using containers then you don't 18:49:54 <clarkb> no one is saying that I'm trying to respond to the discussion about osa and kolla above 18:50:00 <noonedeadpunk> ++ 18:50:01 <clarkb> they are actually orthogonal since they don't have the problem 18:50:44 <gouthamr> alright; anything else on $topic? we've 10 mins 18:50:58 <gouthamr> there're only regularly scheduled topics next 18:51:27 <noonedeadpunk> I'd say overlap is potentially always good to have 18:51:44 <noonedeadpunk> just to give a breather for ones who are not risky changing multiple things at the same time 18:51:47 <gmann> ++ 18:52:07 <gouthamr> these next topics (Gate Health and TC Tracker) will probably be expanded if we spend some time next week.. 18:52:09 <gmann> that was one of the key reason of our testing runtime and take care of such overlap when needed 18:52:19 <gouthamr> any objections moving to Open Discussion for these ~8 mins? 18:52:27 <gmann> one thing for gate 18:52:34 <gouthamr> #topic Gate Health 18:52:37 <gouthamr> go for it gmann 18:52:41 <gmann> Tempest broke for py3.8 drop and it is fixed now 18:52:43 <gmann> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/FOWV4UQZTH4DPDA67QDEROAESYU5Z3LE/ 18:52:56 <gouthamr> ah \o/ thanks; you're encouraging pins for unmaintained branches 18:53:00 <gmann> but this fix where tempest master is used for testing means stable/2023.1 to master 18:53:36 <gmann> yes, for unmaintained branches, maintainers need to pin Tempest. which should have been done as soon branches move to unmaintained status 18:53:49 <gmann> like we did for EM branch status change in past 18:53:50 <cardoe> I'm ++ on that. 18:54:13 <gouthamr> ^ ah good place for a reminder that release folks will transition stable/2023.1 to unmaintained/2023.1 in the next couple of weeks likely 18:54:20 <gouthamr> (their target was end of Oct) 18:54:37 <frickler> yes, elod is on pto this week, will happen next week 18:54:46 <gmann> ++ 18:54:47 <gmann> that is all on gate thing 18:54:54 <gouthamr> thanks gmann 18:54:58 <gouthamr> #topic Open Discussion 18:55:02 <frickler> still need to decide how to handle existing unmaintained branches 18:55:03 <bauzas> thanks gmann 18:55:21 <gmann> frickler: I will say if no one fix then it is good signal for EOL 18:55:24 <frickler> IMHO they are in a bad state and should be eoled mostly 18:55:25 <gouthamr> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/UDQAC7SR5JAQJE5WBAG54A2MTBVBTJ44/ ([PTL][release][stable] 2023.1 Antelope transition to Unmaintained) 18:55:33 <gmann> frickler: ++ 18:55:36 <gouthamr> frickler++ yes 18:56:08 <gouthamr> totally in support of pruning the long list we've accrued - and we can't do it selectively no matter what we tell ourselves :( 18:56:23 <gouthamr> selectively = one project at a time, at the request of the project team 18:56:50 <noonedeadpunk> I have 1 thing to raise with releases team about that actually 18:56:57 <fungi> note that unmaintained branches are supposed to be eol'ed by default if nobody volunteers as a caretaker 18:57:06 <noonedeadpunk> As I'd really love to have trailing projects to be also trailing for moving to unmaintained 18:57:20 <gmann> Tempest pin can be one good check here if we have anyone maintaining them or not 18:58:01 <fungi> well, even if tempest jobs are working on them, if nobody volunteers to take care of them they're supposed to go to eol anyway 18:58:02 <noonedeadpunk> There're multiple issues otherwise. First, I need anyway to wait for rest to move to unmaintained to reffer their tags for EM tag of osa, for instance 18:58:37 <noonedeadpunk> then, trailing projects have not released 2024.2 yet, and instead of focusing on release have to focus on unmaintained 18:58:48 <frickler> fungi: yes, but how is this volunteering to be happening, who tracks it, who proposes the actual eol patches? 18:58:51 <fungi> and the volunteers are supposed to renew their intent every 6 months 18:59:25 <gmann> noonedeadpunk: I see your point here, make sense. 18:59:43 <gmann> noonedeadpunk: just wondering how it was when we had EM thing? 19:00:05 <fungi> "By default, only the latest eligible Unmaintained branch is kept. When a new branch is eligible, the Unmaintained branch liaison must opt-in to keep all previous branches active." 19:00:15 <fungi> #link https://governance.openstack.org/tc/resolutions/20230724-unmaintained-branches.html#unmaintained-branches 19:00:28 <noonedeadpunk> I was always ... annoyed that it was not having trailing policy for EM as well :) 19:00:46 <bauzas> time check nope ? 19:00:49 <gmann> in that case, we should do the same trailing policy 19:00:54 <frickler> fungi: well technically the first sentence is wrong ... by default, nothing changes 19:01:01 <gouthamr> a few minor updates from me: 19:01:01 <gouthamr> - the TC PTG summary is still being written; so, please, in case you've any more thoughts to share on the discussions, go for it without any further delay :) 19:01:01 <gouthamr> - there's an "unofficial" official OpenStack website page for the TC: https://www.openstack.org/community/tech-committee ; the content seems to be broken, please fix your foundation profiles.. but also, rest assured, this page will die soon.. because the OpenStack website is getting overhauled 19:01:01 <gouthamr> - next week's meeting will be our Video Meeting! VMWare Working Group folks will be joining us to present their findings, and seeking opinions for organizing work items wrt the large number of new users coming over from VMWare now, and in the near future 19:01:03 <noonedeadpunk> and if you check history - osa was the latest for quite a while to produce em/eom tags 19:01:05 <fungi> frickler: so if there is no unmaintained branch liaison, there is nobody to opt into keeping the branches open, and they should go to eol 19:01:10 <gouthamr> that said, we're a minute over.. 19:01:20 <gouthamr> we can keep chatting.. but i'll end this meeting 19:01:21 <gmann> noonedeadpunk: I mean, good to have trailing policy for unmaintined 19:01:25 <gouthamr> thank you all for attending 19:01:27 <noonedeadpunk> ++ 19:01:29 <gouthamr> #endmeeting