14:00:00 #startmeeting kolla 14:00:00 Meeting started Wed Dec 6 14:00:00 2023 UTC and is due to finish in 60 minutes. The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:00 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:00 The meeting name has been set to 'kolla' 14:00:03 #topic rollcall 14:00:14 \o 14:00:17 o/ 14:00:21 o/ 14:00:25 o/ 14:01:16 o/ 14:01:30 o/ 14:03:43 #topic agenda 14:03:43 * CI status 14:03:43 * Release tasks 14:03:43 * Regular stable releases (first meeting in a month) 14:03:43 * Current cycle planning 14:03:44 * Additional agenda (from whiteboard) 14:03:44 * Open discussion 14:03:52 osmanlicilegi: welcome back ;-) 14:04:01 #topic CI status 14:04:24 mnasiadka: thanks :) I will be back after long and tiring years. 14:04:25 I think it's overall green, with the cephadm-upgrade jobs failing because of the RMQ switch to Quorum queues 14:04:51 but that's being fixed with 902776 14:04:57 #link https://review.opendev.org/c/openstack/kolla-ansible/+/902776 14:05:34 #topic Release tasks 14:05:44 Kolla, Kolla-Ansible and ansible-collection-kolla going for a final Bobcat release 14:05:54 #link https://review.opendev.org/c/openstack/releases/+/902718 14:06:03 Kayobe still waiting for rc1 - last patch merging 14:06:09 jovial: will you create a patch for kayobe repos rc1? 14:07:49 well, I guess he has no other choice - worst case scenario I'll do it 14:07:53 let's move forward 14:07:59 #topic Current cycle planning 14:08:27 So, usually we did have a stability phase - so we did stay with building latest stable release (Bobcat) in Kolla 14:08:33 mnasiadka: sure, I can try after the meeting 14:08:43 just to have some calm weeks 14:09:10 But I'm fine with us reverting to master faster to have a look if this breaks us badly 14:09:20 Any opinions? 14:09:33 I'd be fine with waiting until January 14:09:40 #link https://docs.openstack.org/kolla/latest/contributor/release-management.html#early-cycle-stability 14:09:59 makes sense, let's at least have stable gate over christmas period 14:10:36 should we prepare patches that need reverting in a gerrit topic - so it's easy to revisit them in a month time? 14:11:15 I'm fine with either. a topic seems nice 14:11:45 I think so - if there's a volunteer - I would be happy ;) 14:12:24 ok then 14:12:48 Any other obvious issues that we need to fix in freshly released Bobcat? 14:12:55 Ah, the horizon policy thing? 14:13:15 yeah 14:13:17 #link https://review.opendev.org/q/topic:%22fix-horizon-role%22 14:13:48 https://review.opendev.org/c/openstack/kolla-ansible/+/901761 - seems upgrade jobs are failing here 14:14:11 mnasiadka: yes, the fix needs to be backported first 14:14:34 makes sense 14:14:45 #link https://review.opendev.org/c/openstack/kolla-ansible/+/902659 14:14:48 Looking for another +2 14:15:19 upgrade-cephadm is known for failing now 14:15:58 ok then 14:16:09 those two on the radar and let's see if anything else shows up 14:16:43 I'll finally populate Whiteboard with cycle priorities - so we can discuss next week 14:16:58 I'll be testing things downstream, image builds are fine already 14:17:27 I'd also like to have a go at trying to look through all kolla-ansible reviews and abandon very old reviews and try to review things that would be useful 14:18:17 kolla looks quite nice with only a couple of patches in progress 14:18:28 sad thing is a lot of things likely are indeed still useful and relevant. but yes, that would be a good task 14:18:51 there's a lot that are ready or nearly ready - we just lack capacity to review them 14:19:05 so I'll try to make them tidy and maybe RP+1 those that are easy to get in 14:19:10 and fairly small 14:19:17 happy to do some additional reviews 14:19:36 ok then, let's move on 14:19:45 #topic Additional agenda (from whiteboard) 14:20:07 mattcrees: https://etherpad.opendev.org/p/kolla-ansible-rmq-quorum-queues-proposal 14:20:18 I think we already violated that proposal with turning quorum queues as default in B 14:20:28 question if we need more prechecks 14:20:54 maybe such prechecks, that would check if a queue is non-durable (if it exists) 14:20:58 but mattc is not here 14:21:04 I'll leave that for next time 14:21:39 jsuazo also is reminding us about two reviews - but I would prefer to wait until we get final Bobcat release out until we merge new features 14:22:48 so both topics are left for next week 14:22:50 mattc reckons we need the same style of precheck that we added previously for HA queues. He will try an propose something this week. 14:23:33 Well, basically I would be happy if we could have prechecks that really stop kolla-ansible break services - so happy to review once it gets into Gerrit 14:24:27 seems you have summoned someone :) 14:24:36 o/ 14:24:48 well, that's the PTL role I guess 14:25:20 mattcrees: so, do your planned prechecks plan to check the ,,durability'' of the existing queues before starting services? 14:25:46 currently our cephadm-upgrade jobs are broken, because we're not stopping cinder, and it tends to complain about non-durable queues that were left behind 14:26:11 Yes, the plan is to have a check similar to this: https://github.com/openstack/kolla-ansible/blob/af6e1ca4fd3366c1ae936d225b2835e95a4191fc/ansible/roles/rabbitmq/tasks/precheck.yml#L211 14:26:24 Only for quorum queue type instead of durability type 14:27:05 ok, let's see how that works 14:27:14 So let's move to open discussion 14:27:17 #topic Open discussion 14:27:27 We should be stopping cinder though... https://github.com/openstack/kolla-ansible/blob/af6e1ca4fd3366c1ae936d225b2835e95a4191fc/tests/upgrade.sh#L22C20-L22C20 14:27:27 osmanlicilegi: you can now share what made you miss us ;-) 14:27:43 mattcrees: we are not, because SCENARIO env var is not passed to upgrade.sh 14:28:04 but it will be - https://review.opendev.org/c/openstack/kolla-ansible/+/902776 14:28:12 ah ok makes sense 14:28:47 the wsgi scripts discussion is worth watching, but I think we can sit and wait how it works out 14:28:52 #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/5RQGZZMCKKHAVH4OLSJCLBQNIHYDPGZD/ 14:28:58 yeah, seen that 14:29:09 I think we still have some services to move to wsgi 14:29:13 starting with Neutron 14:29:19 mnasiadka: very long story :) I will share it via dm. 14:29:29 osmanlicilegi: ok then :) 14:29:46 the eventlet issue is also interesting, but even less we can do about it 14:29:52 do you mean wsgi or uWSGI? 14:30:07 but we could use some more review cycles - so osmanlicilegi feel free to do reviews if you have time :) 14:30:24 SvenKieske: it affects both in a way 14:30:30 but let's not merge any features in master until next weekly meeting please 14:32:21 mnasiadka: I will start soon reviewing :) 14:32:59 cool :) 14:33:26 frickler: ubuntu 24.04 is going to be released in April, we might need to add support for it last minute in C... 14:33:57 but I hope not :) 14:34:09 :D 14:34:15 hmm, I'd rather backport that later 14:34:20 is it possible to test pre releases? 14:34:55 not in the opendev ci 14:36:33 frickler: we'll still have time from April + 3 months, so... I reckon last minute addition 14:37:25 ok, but first we need to see if ubuntu keeps releasing on time ;) 14:37:49 yup 14:38:00 I'll add that to the priorities list with a note to "look out for it" 14:38:09 Ok then, let's finish for today 14:38:16 Ah, one last thing 14:38:22 when do we want to stop meetings for this year? 14:39:07 I planned to cancel the 20th and 27th of Dec and 3rd Jan 14:39:15 I'll be around anyway 14:39:30 I'll be around as well, but I think majority of people will be off 14:39:51 So I'll cancel those three and we can just raise things on the channel outside of the meeting 14:40:05 ack 14:40:06 Thanks for coming, talk to you next week :) 14:40:08 #endmeeting