14:00:39 #startmeeting kolla 14:00:39 Meeting started Wed Mar 27 14:00:39 2024 UTC and is due to finish in 60 minutes. The chair is bbezak. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:39 The meeting name has been set to 'kolla' 14:00:46 #topic rollcall 14:00:49 o/ 14:01:09 \o/ 14:01:10 o/ 14:01:13 o/ 14:01:22 o/ 14:01:32 o/ 14:01:48 o/ 14:01:49 \o 14:02:35 #topic agenda 14:02:37 * Roll-call 14:02:37 * Agenda 14:02:37 * Announcements 14:02:37 * Review action items from the last meeting 14:02:37 * CI status 14:02:38 * Release tasks 14:02:38 * Regular stable releases (first meeting in a month) 14:02:40 * Current cycle planning 14:02:40 * Additional agenda (from whiteboard) 14:02:42 * Open discussion 14:03:26 #topic CI status 14:03:53 watermelon? 14:04:09 red on the inside? ;) 14:04:11 watermelon? :D 14:04:15 ah! lol 14:04:18 :) 14:04:20 didn't get that 14:04:37 green outside, red inside))) lol 14:04:43 exactly 14:04:52 not so bad nowadays I think 14:05:03 well at least kolla build failrate seems flat since 24.03. 14:05:30 wasn't the dashboard supposed to be fixed? I see "no data" again: https://grafana.opendev.org/d/c0d59dad13/kolla-failure-rate?orgId=1 14:05:55 for k-a source graphs, kolla graphs are fine 14:06:06 I think it was working last time I checked 14:06:14 I didn't look at it in ages, no idea 14:06:32 I'm pretty sure as well it worked, I'm to blame for not looking often enough I guess. :D 14:06:49 ok, let's look into that 14:07:19 I haven't seen too many complaints from kevko recently, so CI can't be too red in general, I'd say ;) 14:07:39 yeah, I think it looks good 14:07:58 :D 14:08:13 centos build is broken 14:08:21 but it's non-voting 14:08:23 mhm, can I ask something about CI? I know we plan to extend ARM support downstream, ARM pipeline still seems rather flaky 14:08:45 and is non voting, if it's in good shape now, could we make it voting? 14:08:58 seems there was an issue with centos mirrors earlier today, jrosser was also complaining, should be better now 14:09:28 for ARM builds we only have a single cloud provider, that's mostly why we consider it not too reliable 14:10:29 #topic Release tasks 14:10:43 we're in R-1! 14:11:05 People running and screaming 14:11:13 :) 14:11:17 https://docs.openstack.org/kolla/latest/contributor/release-management.html#r-1-prepare-kolla-kolla-ansible-for-rc1-stable-branch-creation 14:11:43 bit early for branching, isn't it? 14:11:53 It says prepare 14:12:04 And we will be prepared for next 3 months 14:12:10 let's start to prepare then 14:12:17 I can take a look 14:12:33 well the UCA and RDO stuff got merged afaik 14:12:45 So let’s have a look what is missing 14:12:53 And we need a volunteer to fix Octavia 14:13:10 octavia ? what is with octavia ? 14:13:17 *wrong 14:13:19 do we have a bugreport for octavia? I have only a fuzzy memory what exactly is broken 14:13:36 kevko: CI is failing on cascade delete of LB 14:13:45 aaaaa, i saw this 14:13:47 maybe forgot anything in A-C-K ? again 14:14:41 any volunteer to take a look into octavia in CI ? 14:14:45 now, is this a release task? or just regular bugfix? 14:14:57 Exploratory surgery 14:15:00 regular bugfix needed for release 14:15:01 :) 14:15:02 :D 14:15:08 bbezak: i will probably because i am playing with ci and now everything is almost working.... so i am playing around ... 14:15:21 cool 14:15:29 shout if you need help 14:15:39 ok .. 14:15:48 I got a question regarding the release tasks posted above, it has this gem: "The release notes for each project should be tidied up" and I have no idea what that means even after following the review examples there. 14:16:17 It means read through all renos, fix obvious errors, remove duplicates 14:16:18 it seems there are mainly old release notes being deleted? I'm not really sure, is this obsolete? 14:16:44 So it looks readable for a random person 14:16:53 ah okay. weird. I would think this is part of normal code review, but okay, thanks for the explainer :) 14:17:15 We have sort of a rule to not obsess over renos in code review 14:17:21 And fix it close to release 14:17:36 And mistakes do sneek in :) 14:17:38 To improve the overall merge flow or something like that 14:18:03 some things are also easier to spot with a global view, like consistent formatting, e.g. use ``code`` where applicable 14:18:07 ok, I guess I missed those fixes then the last few releases :) thanks for the explanation, I just didn't know how to interpret "tidied up". 14:18:23 Ok, I’d like to discuss the ptg schedule if bbezak will permit 14:18:34 I'll think about it 14:18:50 I can’t do any other day that Monday, bbezak is on vacation 14:18:53 ok, granted 14:18:54 #topic PTG planning 14:19:11 So my proposal is to do Kolla/Kolla-Ansible on Monday for 4 hours 14:19:44 And follow up on not discussed topic on the week after if needed 14:20:31 Unless frickler or kevko feel an urgent need to drive the Kolla PTG on Tue 14:20:36 hmm, we have state holiday :/ 14:20:51 but i will probably join .. 14:20:57 kevko: on Monday? 14:21:14 let me check 14:21:28 i am not sure 14:21:29 that would be April, 8th 14:21:39 I'm fine with skipping Tue 14:22:27 okay, we don't :) 14:22:35 Ok, so that’s agreed - Monday only 14:22:40 +1 14:22:45 jovial: want to move Kayobe? 14:22:53 It’s currently planned for Thursday 14:23:09 I would prefer Wednesday if we did move it. But does that work for others? 14:24:08 does somebody adjust the ptg schedule then? so people are informed? 14:24:59 I will adjust and send mail to Ml 14:25:18 all this because we skip doodle service) 14:25:23 jovial: Kayobe is mainly SHPC - just make sure key people will attend and I’ll move 14:25:39 I guess if Kayobe was moved to the Tuesday then we'll avoid the normal kolla IRC meeting 14:25:48 mmalchuk: doodle made no sense because you all chose the same slots we used for ages 14:26:12 maybe or maybe not 14:27:01 anyway we change slots 14:27:39 ok, anything else PTG related? 14:28:37 as a reminder a link to etherpad - https://etherpad.opendev.org/p/kolla-dalmatian-ptg 14:30:02 let's move on 14:30:18 #Link https://etherpad.opendev.org/p/kolla-dalmatian-ptg 14:30:18 #topic Current cycle planning 14:31:27 does anybody has something to discuss about outstanding tasks? 14:32:09 I was overloaded with customer related tasks recently, so no much update from me I'm affraid 14:32:40 I want to ask to merge init-runonce modification to stable branches 14:32:57 :) 14:33:28 well looking at the caracal tasks not that many are done afaik? 14:34:17 anybody has any news on the next ubuntu release? we wanted to build that in for C release 14:34:47 25.4 ? :D 14:34:53 nope... silent 14:34:55 no, 24.04. 14:35:00 kevko: I've seen a bit discussion the other day. I don't think it is a good idea to modify stable branches without doing the same to master? 14:35:08 google saying Ubuntu 24.04 LTS, codenamed Noble Numbat, is scheduled for release on April 25, 2024. 14:35:29 SvenKieske ask again afer 25 apr 14:35:39 after* 14:35:56 bbezak: yes, i want to merge also to master ...same patch ...but it's needed also to merge to stable branches ..as upgrade jobs are checkout ing previous release ... 14:36:02 but why was it put on the agenda then? I mean the release dates of ubuntu are pretty well known in advance? weird. 14:36:37 bbezak: moreover, it's only runonce script which is only suggested to use as per documentation ...normally it's not used I would say ..with end user 14:37:21 SvenKieske: I think it was just to prepare images building for now. And caracal is a good place for that as it is a SLURP release 14:37:28 bbezak: now i hacked ci :D and download script from my github repo to pass upgrade jobs :D 14:37:51 24.04 feature will be probably backported to caracal anyway at some point 14:38:03 ok 14:38:14 btw about CI we have an issue with docs site preview. anyone know this? 14:39:06 meta: can we maybe clean up the whiteboard? it still has listed e.g. train PTG; afaik a link to some archived document would be enough there, no? the Whiteboard is 880 lines long now and hard to navigate.. 14:39:46 sorry, maybe off topic for current cycle planning 14:40:14 mmalchuk: where do you see that issue? 14:40:28 preview didn't work 14:40:39 on any last change 14:40:41 SvenKieske: I'll add it to notes 14:40:49 I'll show you... moment 14:41:21 bbezak: ty, I can also volunteer to do cleanup or create some archive somewhere (different etherpad) as long as I know what can be cleaned up and what should stay. 14:41:28 wfm https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_33b/911417/12/check/openstack-tox-docs/33b87a9/docs/ 14:42:04 mmalchuk: maybe some storage issue? did you get an (HTTP) error? anything? 14:42:20 oops... fixed. looks like infra knows and fixed 14:43:07 it was an issue with listing 403 forbidden with banner 14:43:26 anyway fixed... lets move ahead 14:43:33 #topic Additional agenda (from whiteboard) 14:43:44 I can see mostly a review requests there 14:43:58 anything to add apart of that? 14:44:09 yes :) I already asked this week but I guess if you don't highlight usernames nobody has time :P :D 14:44:21 yes, plus Ivan still has that issue with Docker version on Debian 14:44:22 mhiner, r-krkcek 14:44:35 sorry :) r-krcek 14:45:17 other than that, it's just reviews from me 14:45:26 so I'll explicitly mention it here: mnasiadka, frickler: https://review.opendev.org/c/openstack/kolla-ansible/+/855498 please come to a solution where to test ovn-exporter, currently I made it work in the ovn scenario and I honestly would prefer if it stays there, because these tests are pretty brittle 14:46:08 meaning it didn't work at all in prometheus-opensearch scenario for many reasons. 14:46:20 For me its the reviews and one question, but wanted to ask that in the open debate. 14:47:00 #topic Open discussion 14:47:17 r-krcek: seems you need to rebase the "add sysctl role change" https://review.opendev.org/c/openstack/kolla-ansible/+/912351 14:47:19 I noticed that we have jinja filter for finding out that service is enabled and mapped to host. However, it looks like it is used only with certain services. Is there a specific reason for not using it everywhere or was is just because of lack of time to do it? The filter: https://review.opendev.org/c/openstack/kolla-ansible/+/665922 14:47:58 lets discuss https://bugs.launchpad.net/kolla/+bug/2042954 14:48:14 SvenKieske: oh, sorry. Ill do that right after the meeting. 14:48:19 imho https://review.opendev.org/c/openstack/kolla/+/913584 not needed and https://review.opendev.org/c/openstack/kolla/+/904781 need revert 14:49:26 the issue appears in kernel 5.18+ but we have 5.4 and 5.15 maximum 14:50:04 the issue not reproduced on Focal and Yoga 14:50:26 mmalchuk: case in point: I know people running HWE kernels on jammy, which bumps you up til 6.5 kernel :) 14:50:40 HWE is a 5.15 14:50:42 Roman Krček proposed openstack/kolla-ansible master: Update cell0 database connection https://review.opendev.org/c/openstack/kolla-ansible/+/910924 14:51:08 22.04-hwe is 6.5 14:52:13 should we support HWE on Jammy in Yoga? Jammy only for transition to Zed 14:52:32 yoga is unmaintained, we should not care at all 14:52:37 so 14:52:41 frickler: maybe we don't have this issue because we have newer openstack? 14:52:51 didn't look into it what the actual issue is 14:53:03 I've dig deeper 14:53:20 the issue only on upstream kernel with old libs 14:53:20 funnily enough the bugreport doesn't mention what the bug is, just that there is one. 14:54:00 ah this is the original bug: https://bugs.launchpad.net/ubuntu/+source/neutron/+bug/1995735 14:54:22 yep 14:55:00 I've tested all the libs/packages for issue and provide paste links 14:55:35 even if we planed to merge https://review.opendev.org/c/openstack/kolla/+/913584 we need fix issue with pyroute2.protocols 14:55:55 we install it in the https://review.opendev.org/c/openstack/kolla/+/904781 via packages 14:56:08 and it can't be uninstalled via pip 14:58:41 r-krcek: let me get back to you on your services/filter question, I need to check this feature 14:58:49 lol? 14:59:07 according to https://openstack-ci-reports.ubuntu.com/reports/cloud-archive/yoga_versions.html the package is not called "python3-pyroute2" but "pyroute2" 14:59:25 but it's installed via "python3-pyroute2" https://review.opendev.org/c/openstack/kolla/+/904781/3/docker/openstack-base/Dockerfile.j2 14:59:40 the other versions in the CI also report this package name afaik 14:59:51 there is no issues with packages 15:00:07 SvenKieske: pyroute2 is source package name 15:00:19 in current images we have pyroute2 on both venv and system 15:00:22 SvenKieske: python3-pyroute2 is binary package 15:00:42 kevko: ah nice, madness.. 15:00:58 I've publeshed an edit for https://review.opendev.org/c/openstack/kolla/+/913584 15:01:20 mmalchuk: understood, seems you are right (beside the HWE kernel version, but those are a moving target anyway) 15:01:32 SvenKieske: standard 15:02:02 I guess we are in overtime now. 15:02:03 ok guys, I need to jump to some call. However please continue the discussion of course :) 15:02:12 thank you all 15:02:14 #endmeeting