18:00:10 <noonedeadpunk> #startmeeting tc 18:00:10 <opendevmeet> Meeting started Tue Dec 17 18:00:10 2024 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:10 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:00:10 <opendevmeet> The meeting name has been set to 'tc' 18:00:21 <noonedeadpunk> 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. 18:00:28 <noonedeadpunk> Today's meeting agenda can be found at https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee 18:00:36 <noonedeadpunk> #chair gmann 18:00:36 <opendevmeet> Current chairs: gmann noonedeadpunk 18:00:46 <noonedeadpunk> #topic Roll Call 18:00:49 <gmann> o/ 18:00:52 <slaweq> o/ 18:00:53 <noonedeadpunk> o/ 18:00:58 <frickler> \o 18:01:25 <gtema> o/ 18:02:36 <noonedeadpunk> courtesy ping cardoe 18:03:15 <cardoe> I’m on. 18:04:26 <noonedeadpunk> awesome, then we can start. Notified absence g o u t h a m r, s p o t z and b a u z a s 18:04:41 <noonedeadpunk> #topic Last Week's AIs 18:05:06 <noonedeadpunk> I've looked through last meting results and I don't see any AIs. Am I missing some? 18:06:06 <gmann> I think nothing other than we have separate topics for a few of the followup 18:06:26 <noonedeadpunk> ++ 18:06:36 <noonedeadpunk> #topic Announcement 18:06:43 <noonedeadpunk> There will be no weekly meeting on 24th Dec 2024 and 31st Dec 2024. Our next meeting after this one will be on 7th Jan 2025 18:06:50 <noonedeadpunk> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/YQHW23JKGQVGUBWPOIU6QTRKIP44AI7L/ 18:07:11 <noonedeadpunk> #topic Update on unmaintained branches 18:07:26 <noonedeadpunk> frickler, any updates from your side? 18:07:40 <frickler> most things are in the review 18:08:05 <frickler> elod has announced that his company want to keep some repos alive, we'll see how that goes 18:08:25 <frickler> no other people interested in opting in to anything afaict 18:09:05 <frickler> #link https://review.opendev.org/c/openstack/releases/+/937515 18:09:38 <frickler> one question maybe: should I keep the patch in WIP state, which might hide it for some reviewers/make it less visible? 18:09:53 <JayF> Adding to that: the Ironic community determined at PTG none of our people who work on UM branches want to keep Victoria. 18:10:02 <frickler> or unwip and mark W-1 instead? 18:10:15 <gmann> ++ I think removing the WIP is good idea 18:10:45 <noonedeadpunk> though it depends on how to address elod's comment 18:10:58 <gmann> I am not sure if elodilles_pto interested to keep all projects repo or a few of them. 18:11:19 <gmann> if not all then this change at least can EOL the not required one 18:11:21 <frickler> iiuc just some yet to be specified subset 18:11:41 <noonedeadpunk> so kind of WIP makes sense then, as potentially no need to ask people for reviews anyway 18:11:51 <noonedeadpunk> (at this point) 18:12:24 <frickler> well not review of the content itself, but review as mentioned in the commit message by stating intent to opt-in if there is one 18:12:38 <gmann> having +1 also a good input there to know people have reviewed it and know V is going to EOL 18:12:43 <noonedeadpunk> as I spotted that people frwquently just miss to re-vote on patches if they updated, once they've voted on them 18:12:49 <cardoe> How can we make it less painful going forward? 18:13:24 <frickler> well creating this patch was easier than I expected 18:13:47 <frickler> the painful part will be integrating the updates that elod is to make and checking the CI status 18:13:47 <gmann> I think all -1 are something we need to address/discuss them. I do not think they will get ignore 18:14:22 <noonedeadpunk> it somehow feels that EM was easier to manage/maintain overall. at lest that is my personal impression so far. 18:14:48 <noonedeadpunk> and eom requires more time from more parties... 18:14:49 <gmann> yeah, once we know the list of repo we need to keep. From TC side, a few volunteer can verify that, discuss/ack in TC meeting and give TC input there 18:15:28 <frickler> I'd also question the usefulness of the whole concept if only a single company is interested in it 18:15:43 <gmann> in last meeting gouthamr and I volunteer to do ^^ part. keep monitoring the -1 on EOL changes and verify the requitement to keep those as unmaintained 18:16:20 <gmann> frickler: yeah that is good point but it is hard to deny anyone even single company/maintainer to maintain anything. even it is single company, it is still worth to keep them as unmaintained 18:16:52 <JayF> frickler: One company willing to do the work doesn't necessarily imply there's only one company reaping the benefits. While I don't like a single point of failure, I think it's still meaningful to keep that stuff available if we believe the folks involved will keep their end of the bargain (a pretty good bet w/Elod) 18:16:56 <gmann> I think it is same as single company maintaining our many of projects alone 18:17:24 <JayF> And given we don't even make maintenance promises for UM branches; the SPOF issue is even less of a concern. 18:18:37 <gmann> starting the V is good idea as in this change we can filter out the repo continuing in unmaintained from the other branch EOL 18:18:44 <frickler> anyway I think we're fine with this for now and we'll see how things turn out next month 18:18:53 <gmann> ++ 18:19:00 <noonedeadpunk> and I assume we move on with W and X once V is done, correct? 18:19:08 <frickler> that's the idea, yes 18:19:24 <noonedeadpunk> ok, thanks for the update. moving on then 18:19:25 <gmann> yeah I think till zed 18:19:25 <frickler> W, X, Y and Z 18:19:29 <gmann> yeah 18:19:59 <noonedeadpunk> well. for Y I have a specific concern, but let's probably come to it first... 18:20:11 <noonedeadpunk> #topic Update on election and changes 18:20:56 <slaweq> regarding election, there is patch https://review.opendev.org/c/openstack/election/+/937408 with proposed dates for the upcoming election 18:21:04 <slaweq> please take a look if you want 18:21:43 <noonedeadpunk> was looking for the link, thanks :) 18:21:50 <frickler> would be good to actually merge it this week IMHO 18:21:51 <slaweq> there is also patch https://review.opendev.org/c/openstack/election/+/934908 from gouthamr which needs reviews 18:22:02 <gmann> slaweq: any date you are planning to merge it? I see many TC members reviewed it 18:22:33 <slaweq> gmann I didn't plan but I will sync with ianychoi about it tomorrow morning 18:22:40 <gmann> thanks 18:22:47 <slaweq> and we should merge it this week as frickler said 18:22:52 <gmann> yeah, ++ 18:24:03 <noonedeadpunk> it has majority of TC votes as well 18:24:18 <noonedeadpunk> which is probably fine already 18:24:33 <slaweq> yeah, I just noticed after I gave link here :) 18:24:45 <slaweq> thx 18:24:58 <noonedeadpunk> great, anything else on the topic? 18:26:05 <noonedeadpunk> #topic Status on migrate CI to Ubuntu Noble 18:26:20 <noonedeadpunk> I see a big progress was made in this regard 18:26:27 <noonedeadpunk> #link https://etherpad.opendev.org/p/migrate-to-noble 18:26:46 <noonedeadpunk> So far the most blocking thing is PBR from what I see 18:26:55 <gmann> yeah, overall good progress by many projects except the 3 open bugs 18:26:57 <noonedeadpunk> And then couple of issues with tacker 18:26:58 <clarkb> is pbr blocking anything? 18:27:02 <gmann> and the doc job (PBR one) 18:27:17 <gmann> it is blocking doc job migration to noble 18:27:35 <gmann> openstackdoctheme fail due to pbr bug 18:28:04 <gmann> #link https://bugs.launchpad.net/pbr/+bug/2088360 18:28:43 <clarkb> as mentioned previously does installing setuptools not fix that? 18:28:56 <clarkb> I feel like we keep going in circles on this topic 18:29:08 <clarkb> my understanding is that pbr works fine you just need to preinstall setuptools on python3.12 or newer 18:29:15 <clarkb> has anyone refuted that and found that this isn't the case? 18:29:28 <noonedeadpunk> I've just marked one of tacker issue as resolved 18:29:35 <noonedeadpunk> #link https://bugs.launchpad.net/tacker/+bug/2090859 18:30:09 <gmann> noonedeadpunk: that is not yet resolved, bug is in devstack-plugin-container whihch hongbin is looking into 18:30:10 <noonedeadpunk> I indeed haven;t seen real issues with pbr, but indfeed setuptools is required 18:30:32 <gmann> for tacker k8s job which were failing due to this, I kept nodeset as jammy until this is fixed 18:30:33 <noonedeadpunk> ooops 18:30:43 <clarkb> gmann: also I'm -2 on your suggestion to remove python2.7 from pbr 18:30:59 <clarkb> you will instantly break any software running python2.7 that uses pbr 18:31:00 <gmann> clarkb: I think failure was about distutils usage as explicit ? 18:31:14 <clarkb> gmann: yes distutils usage is explicit but setuptools includes distutils 18:31:15 <noonedeadpunk> And I can't change the status back /o\ 18:31:22 <clarkb> if you install setuptools you should be able to import distutils 18:31:28 <noonedeadpunk> I will reach folks about that after the meeting 18:32:15 <gmann> let me try that 18:32:31 <clarkb> to be clear my understanding of pby with python3.12 is that it works fine if you install setuptools 18:32:37 <gmann> clarkb: one question, is it ok to do that in pbr instead of all other places, 18:32:44 <gmann> i mean this one change only 18:32:45 <clarkb> the only reason it does not work is that python3.12 does not include setuptools like every other python before it did 18:32:46 <gmann> #link https://review.opendev.org/c/openstack/pbr/+/924216/8/doc/requirements.txt 18:33:06 <noonedeadpunk> I'm pretty sure clarkb is right here as can recall fixing smth that way already 18:33:29 <clarkb> gmann: you cannot add setuptools as a dependency of PBR itself, but I think adding setuptools to the doc requirements should be fine 18:34:03 <clarkb> the reason for not adding it as a general dependency is that PBR runs as a setuptools script so setuptools must already be installed for it to work 18:34:37 <clarkb> the modern way to do this is via pyproject.toml which is why they removed setuptools from stdlib (or one reason they did it) 18:34:42 <gmann> k, let me try to install it on openstckdoctheme and if nowhere else we need than we do not need in pbr itself. if needed we can do in doc/requirements only 18:34:50 <gmann> yeah 18:35:17 <clarkb> anyway I just want to be 100% clear that I maintain projects with python3.12 support that use pbr and it works as far as I know if you install setuptools. 18:35:36 <clarkb> I don't liek the idea that PBR is blocking anything when I've now asked like 3 times in three different meetings for people to just install setuptools and report back if it works 18:35:38 <noonedeadpunk> ++ 18:35:39 <gmann> k 18:36:06 <gmann> other than pbr, I still did not see response from skyline team 18:36:10 <cardoe> It works if you install setuptools. I agree. 18:36:11 <noonedeadpunk> ok, hopefully we've solved that at last now :) 18:36:17 <gmann> #link https://review.opendev.org/c/openstack/skyline-apiserver/+/935604/2 18:36:19 <gmann> #link https://review.opendev.org/c/openstack/skyline-apiserver/+/935600/5 18:36:43 <gmann> #link https://review.opendev.org/c/openstack/skyline-console/+/935601 18:37:03 <gmann> so not sure if this inactiveness can cause any issue in release and we need to react in advance? 18:37:08 <noonedeadpunk> well, the last one is not in it's best shape from what I see 18:37:16 <gmann> but this is the only project I find with no response on this migration 18:37:35 <gmann> yeah, console is good but there is no review to merge it 18:37:45 <gmann> and apiserver failing gate and no response from team 18:37:56 <noonedeadpunk> frickler: do you have poc within skyline team? 18:38:08 <noonedeadpunk> (point of contact) 18:38:25 <gmann> as long as team ack and response even fix is taking time, i am ok but no response does not seems good 18:38:36 <frickler> not directly, can ask in the channel 18:39:02 <noonedeadpunk> aha, gotcha. used to be tricky to get any attention there iirc... 18:39:28 <gmann> I can send another ML reminder to them as separate ping also 18:39:48 <noonedeadpunk> yes, would be much appreciated! 18:40:33 <cardoe> As far as openstackdocstheme and pbr, I've just been advocating us (OpenStack) taking a "less is more" approach. A lot of things were created when there were a lot more folks. e.g. separate docs team. 18:41:36 <cardoe> In a lot of cases, upstream projects have caught up to what OpenStack has done but implemented slightly differently. So where it makes sense and we can lessen maintenance burden it would be a + 18:42:32 <clarkb> yup we could use built in sphinx templates and possibly switch out pbr with other tools if we give up on some features (or find a way to integrate those features into other installation tools) 18:42:58 <clarkb> on the theming side of things in particular that woudl get better mobile device support instantly I think 18:43:25 <JayF> ++ 18:43:38 <cardoe> Yeah that'd be a ++ from me too. 18:44:04 <JayF> I think part of the issue is it's chicken:egg -- doing something like retiring that theme pays off in the long term, but is likely more work in the short term. 18:44:44 <cardoe> pbr wise I think it brings a lot of value to the build side. But from the runtime side, it's mostly a wrapper of functionality that's finally upstream (or at least in another package) 18:45:04 <noonedeadpunk> and then also fully giving up on docs branding... not super cool either 18:46:11 <JayF> noonedeadpunk: the "branding" is already highly inconsistent when you navigate to our docs from the front pages 18:46:12 <noonedeadpunk> I totally see how right now some stuck-with-other-platfrom CTO would look on native sphinx theme when considering migration out 18:46:35 <cardoe> noonedeadpunk: if I had the ability to easily switch between the versions, good mobile support and functional indexes then I'd take docs that were titled "hello world" 18:47:06 <noonedeadpunk> Is it? As I kind of doesn;t see that a lot, except when talking about not-so-well maintained projects lagging behind with docstheme versions 18:47:20 <cardoe> https://docs.openstack.org/neutron/latest/admin/config-routed-networks e.g. where am I wrt to the docs tree there? 18:47:33 <clarkb> many of the themes do allow for adding logos at least 18:47:42 <clarkb> zuul is using a stnadard theme but sticks its logo on the pages 18:47:48 <JayF> noonedeadpunk: I experience this with Ironic; if we're using the wrong version of the docs theme let me know and I'll fix it *today* 18:48:33 <noonedeadpunk> ok, probably I'm just not in the full context of the issue... 18:49:24 <noonedeadpunk> and we also need to move on, but I'd suggest adding this topic for the next meeting to have a focus on it 18:49:41 <noonedeadpunk> #topic A check on gate health 18:50:38 <gtema> noonedeadpunk - it doesn't make sense to re-add a topic that we discussed multiple times and people simply being afraid to touch anything (do not touch a running system, while we, well, have no maintainers left) 18:51:30 <gmann> one issue i know was introduced by Tempest last week which failed cinder testing, is fixed now. 18:51:34 <noonedeadpunk> I'd say that in everything we do, there should be a volunteer to step in and build some POC docs as a showcase 18:51:47 <gmann> other than that I have not observed any new failure in gate 18:51:57 <noonedeadpunk> which may make ppl be less afraid of the change 18:52:07 <gmann> noonedeadpunk: ++ 18:52:09 <gtema> I volunteered and proposed update to openstackdocstheme. What has happened? nothing 18:52:30 <gtema> and we spend quite lot time "discussing" 18:53:08 <noonedeadpunk> gmann: ok, nice that it's fixed in a timely manner 18:53:29 <noonedeadpunk> I don't have anything to report on gate health so far. 18:53:35 <gmann> yeah, we added some cinder tests in tempest gate to avoid these in future 18:53:52 <noonedeadpunk> Horizon plugins has been fixed wrt to noble support 18:54:11 <noonedeadpunk> (not sure if I've already said that last week or not) 18:54:15 <gmann> yes, that is all good now. manila horizon plugin also passing 18:54:32 <cardoe> gtema: I believe I +1'd it. 18:54:35 <noonedeadpunk> #topic PTG AIs and the TC Tracker 18:54:48 <noonedeadpunk> #link https://etherpad.opendev.org/p/tc-2025.1-tracker 18:57:18 <noonedeadpunk> there're quite some action points on gouthamr there 18:58:52 <noonedeadpunk> also eventlet goal was reworked and waiting for reviews 18:58:55 <noonedeadpunk> #link https://review.opendev.org/c/openstack/governance/+/931254 19:00:13 <noonedeadpunk> but at first look feels like huge amount of context was removed from it 19:00:25 <noonedeadpunk> and we are out of time now :( 19:00:34 <noonedeadpunk> #endmeeting