15:00:36 <noonedeadpunk> #startmeeting openstack_ansible_meeting 15:00:36 <opendevmeet> Meeting started Tue Dec 3 15:00:36 2024 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:36 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:36 <opendevmeet> The meeting name has been set to 'openstack_ansible_meeting' 15:00:40 <noonedeadpunk> #topic rollcall 15:00:42 <noonedeadpunk> o/ 15:05:49 <noonedeadpunk> #topic Dalmatian release 15:05:56 <noonedeadpunk> so RC2 is about to be ready 15:06:11 <noonedeadpunk> I will propose it once https://review.opendev.org/c/openstack/openstack-ansible/+/936810 land 15:06:30 <noonedeadpunk> hopefully it won't fail intermittently 15:06:41 <noonedeadpunk> though, there's already an issue I spotted in it 15:06:59 <noonedeadpunk> which is that rabbitmq_upgrade can't be re-run after upgrade to 4.0 15:07:08 <noonedeadpunk> under some conditions.... 15:07:36 <noonedeadpunk> as non-tls upgrade jobs are failing 15:07:38 <noonedeadpunk> #link https://zuul.opendev.org/t/openstack/build/56f58e441f2240dfb9842092923eb6c1 15:08:10 <noonedeadpunk> will need to check on that 15:08:48 <noonedeadpunk> but anyway - I think such fixes can be backported once we fix 15:11:41 <noonedeadpunk> #topic new/old projects 15:11:56 <noonedeadpunk> so we have couple of things going on with repos 15:12:07 <noonedeadpunk> senlin/sahara/murano are still not fully retired 15:12:26 <noonedeadpunk> we're waiting for TC to vote on removal them from gobvernance 15:12:48 <noonedeadpunk> httpd role repo - infra hasn't merged the change which would create repo for it 15:12:56 <noonedeadpunk> so it's on a very early stage 15:13:19 <noonedeadpunk> watcher repo - I haven't even proposed to re-incarnate the role 15:14:04 <noonedeadpunk> with that, as I'm starting to look into freezer in my free time - I might wanna raise adding freezer-api role one day 15:14:39 <noonedeadpunk> it's a bit tricky, as consist of multiple components... not sure yet if/how I want to handle agent installation - with dib element, role or both 15:15:30 <noonedeadpunk> but let's not dive deep into it, as I'm not sure about even approximate timeline for this to happen 15:15:48 <noonedeadpunk> and that would also require for freezer to be re-established as active project first 15:26:15 <noonedeadpunk> #topic office hours 15:26:37 <noonedeadpunk> I wanted actually to raise one thing. which is - noble backport to Caracal. 15:26:55 <noonedeadpunk> But it's a bit weird topic and I don't know how I feel overall about the confusion around it 15:27:14 <noonedeadpunk> Like first thing, is that openstack caracal is not tested with python 3.12 15:27:28 <noonedeadpunk> but canonical does ship caracal out of the box with noble 15:27:44 <noonedeadpunk> amount of things they do behind doors is unknown 15:30:18 <noonedeadpunk> but then I somehow already want to deploy new nodes on noble :D 15:30:38 <noonedeadpunk> and don't want to have non-SLURP in production 15:41:34 <mgariepy> Epoxy wont run on 22.04 ? 15:42:23 <noonedeadpunk> it will 15:42:26 <noonedeadpunk> I guess 15:42:40 <mgariepy> hmm ok :) 15:42:52 <noonedeadpunk> at least it does not fail right now 15:45:40 <mgariepy> ok hopefully it will :) 15:48:50 <noonedeadpunk> I guess it come from idea that then we could also have an upgrade path for distro 15:49:01 <noonedeadpunk> which we are not obliged to have, but would be nice if it is there 15:49:21 <mgariepy> it's always appreciated to have an upgrade path for the distros. 15:49:58 <noonedeadpunk> and also do kind of upgrade planning for next year here as well :) 15:50:17 <noonedeadpunk> and it appears that otherwise I'd have slot for OS upgrades only in 2026 lol 15:50:21 <mgariepy> ideally being either on .1 or .2 it would be better to always have SLURP for all version . 15:50:29 <mgariepy> haha 15:50:42 <mgariepy> let's build against arch. no more mass upgrade. 15:51:50 <mgariepy> only small bumps every day :) haha 15:52:15 <noonedeadpunk> yeah 15:52:47 <noonedeadpunk> clients not gonna like that if I'd start restarting l3 agent daily 15:53:01 <mgariepy> haha 15:53:05 <mgariepy> ovn 15:53:35 <mgariepy> but yeah. not quite for server work ;) 15:54:43 <mgariepy> Also the 308 issue on python <3.11 is anoying. 15:55:09 <noonedeadpunk> seen you;ve discussed that, but not still sure why it raised 15:55:13 <noonedeadpunk> smth gitlab related? 15:55:30 <mgariepy> gitlab does now redirect 308 for release . 15:55:43 <noonedeadpunk> well. I would blame not python, but nasty gitlab 15:55:46 <mgariepy> python didn't merge the patch in earliers version. 15:55:47 * noonedeadpunk hate gitlab 15:56:08 * noonedeadpunk thinks that gitlab is just broken version of github 15:57:04 <mgariepy> well currently with ansible on jammy if you take the python version it ships with you cannot get a url that has a 308 redirect. 15:57:38 <noonedeadpunk> well, it is applicable to any python-based app 15:57:41 <mgariepy> the tests for ansible does pull all sort of ramdom verison of python via docker which don't really allign with the version that is shipped with the varions distros. 15:58:06 <noonedeadpunk> which kind of proves the point- that gitlab devs are just obnoxious self-centric 15:58:10 <mgariepy> so ansible 2.15 on jammy pass. but if you pip install ansible-core==2.15 it wont. 15:58:40 <noonedeadpunk> as iirc - it's really on python requests level? 15:59:28 <noonedeadpunk> but dunno... I frankly never heard of 308 code before.... 15:59:38 <mgariepy> urllib level from what i seen 15:59:56 <noonedeadpunk> why in the world not 302 15:59:58 <mgariepy> https://github.com/python/cpython/commit/c379bc5ec9012cf66424ef3d80612cf13ec51006 16:00:28 <noonedeadpunk> ah, true 16:00:55 <mgariepy> the pacth for python 3.9 was submited then closed. 16:01:17 <noonedeadpunk> as it's not supported... 16:01:32 <noonedeadpunk> I wonder if anyone raised that to gitlab instead? 16:03:29 <mgariepy> https://paste.opendev.org/show/bADa7HLYxBFpxUL9nLqV/ 16:03:34 <noonedeadpunk> as given the fact that nobody seen missing 308 as an issue until python 3.11 (for how many decades?) - I'd say one can live without them for some more time... 16:04:08 <noonedeadpunk> ah, i see.... 16:04:46 <mgariepy> where do we ping them ? do you have an idea? 16:06:21 <mgariepy> hmm 16:06:46 <noonedeadpunk> um. but what triggers that request? 16:06:55 <noonedeadpunk> oh, we're over time 16:07:00 <noonedeadpunk> #endmeeting