13:00:10 <mnasiadka> #startmeeting kolla
13:00:10 <opendevmeet> Meeting started Wed Sep 11 13:00:10 2024 UTC and is due to finish in 60 minutes.  The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:10 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:10 <opendevmeet> The meeting name has been set to 'kolla'
13:00:12 <mnasiadka> #topic rollcall
13:00:13 <mnasiadka> o/
13:00:15 <SvenKieske> o/
13:00:18 <mmalchuk> \o
13:00:19 <bbezak> o/
13:00:22 <mattcrees> o/
13:00:26 <frickler> \o
13:00:28 <mhiner> o/
13:00:45 <r-krcek> o/
13:01:15 <kevko> \o/
13:02:27 <mnasiadka> #topic Agenda
13:02:27 <mnasiadka> * CI status
13:02:27 <mnasiadka> * Release tasks
13:02:27 <mnasiadka> * Regular stable releases (first meeting in a month)
13:02:27 <mnasiadka> * Current cycle planning
13:02:29 <mnasiadka> * Additional agenda (from whiteboard)
13:02:29 <mnasiadka> * Open discussion
13:02:31 <mnasiadka> #topic CI status
13:02:35 <mnasiadka> Seems overall fine
13:03:05 <frickler> there was some memory upgrade for the arm nodes, but I didn't check recent results yet
13:03:14 <mnasiadka> Mainly aarch64 is in the woods, but that's some storage/io problem on the ARM nodes in CI
13:03:29 <mnasiadka> frickler: I think it's getting better, but I don't know if it's getting a lot better
13:04:04 <mnasiadka> #topic Release tasks
13:04:15 <frickler> ok, so let's watch for a bit before enabling stopped jobs again
13:04:16 <mnasiadka> Nothing for us in R-3
13:05:02 <mnasiadka> But 2024.2 coordinated release is coming soon
13:05:18 <mnasiadka> And then some weeks for us since we're cycle-trailing
13:05:25 <mnasiadka> but let's try not to aim for December again :)
13:05:39 <mnasiadka> #topic Regular stable releases (first meeting in a month)
13:05:43 <mnasiadka> Any volunteer to raise them?
13:05:52 <frickler> iirc bbezak wanted to?
13:06:15 <SvenKieske> yes, he said so last week
13:06:24 <SvenKieske> iirc
13:06:27 <bbezak> yeah, I forgot about it :)
13:06:29 <bbezak> will do today
13:06:37 <frickler> +1
13:06:39 <bbezak> on it
13:06:46 <mnasiadka> good bbezak, your colleagues have ratted you out on that
13:06:51 <bbezak> :)
13:07:02 <mnasiadka> #topic Current cycle planning
13:07:26 <mnasiadka> So, on my end - MariaDB 11.4 is probably not getting in, unless they fix WSREP/Galera file renames
13:07:42 <mnasiadka> I need to focus on getting Noble in, and some other OVN related things I'm working on
13:08:11 <kevko> Regarding new ubuntu I asked last week which solution will we apply  .. mnasiadka https://review.opendev.org/c/openstack/kolla-ansible/+/926632 or https://review.opendev.org/c/openstack/kolla/+/926744 ... i would prefer the symlinkink ..as it is more clean and simple ...also ..it will help me with another patch i've already posted
13:08:20 <kevko> but waited for mnasiadka ..
13:08:20 <SvenKieske> will we backport mariadb 11.4 later, then? I guess we need to?
13:08:41 <mnasiadka> SvenKieske: we can always bump in E, 10.11 is supported until 2025 or 2026
13:09:34 <mnasiadka> kevko: since we merged dev refactor, that's the only place we're using distro_python_version - I'm not a fan of the symlinking approach in the image, I don't know if we're not violating some venv architecture rules or whatever
13:09:52 <mnasiadka> But if other cores feel like your solution is better - so be it, I just want to get that finally fixed
13:10:11 <mnasiadka> frickler, SvenKieske, bbezak: do we need to vote? ;-)
13:10:29 <kevko> mnasiadka: why you are not fun of symlinking even if distro owners are doing symlinking ? As example we have new python installed in rocky (or centos ? ) ...where newer python3 binary is symlink to newer python ..while older python3 is still there
13:10:44 <frickler> I would have to look at those changes a bit deeper before voting
13:11:18 <mnasiadka> kevko: symlinking a binary vs symlinking whole directory with subdirectories and a lot of libraries is probably something else
13:11:57 <SvenKieske> me too, as far as I looked both patches seemed to do the trick. in general I hate symlinking, but that doesn't mean it has it's unfortunate usecases :)
13:11:59 <mnasiadka> So let's discuss that in gerrit, if other cores could do an in-depth review of both patches - we can discuss the direction next week (or async in any day)
13:12:16 <bbezak> https://review.opendev.org/c/openstack/releases/+/928908
13:12:25 <kevko> mnasiadka: it's same ..
13:12:40 <frickler> bbezak: thx, added to my list
13:13:31 <mnasiadka> Ok, another thing - it's next cycle planning but I'll ask here
13:13:45 <mnasiadka> I'm basically flying back from Indianapolis on first day of PTG
13:13:56 <mnasiadka> last time we had majority of our sessions on Monday
13:14:18 <mnasiadka> Are the key people fine with moving that to Tuesday or some other day?
13:14:54 <frickler> can't you join from the plane? that would be a first I guess? ;)
13:14:58 <mmalchuk> Tuesday ok
13:15:03 <SvenKieske> without looking at any calendars: sure :)
13:15:08 <frickler> but seriously, fine with moving
13:15:12 <kevko> last thing what i want to say regarding this is -> if we really hate symlinking ..does it mean that we are doing something bad in every service container image when we are symlinking the source code and that point pip to install from symlink ?
13:15:12 <kevko> https://github.com/openstack/kolla/blob/c4b12b1ace851b719d63d696ae6a221d535876fb/docker/glance/glance-base/Dockerfile.j2#L40C1-L40C40 ?? cmon ..
13:15:37 <kevko> because it's maybe worse ..
13:15:57 <mnasiadka> ok, thanks, I'll prepare an etherpad and schedule some slots - we can discuss them next week
13:16:30 <mnasiadka> kevko: that's something frickler has also pointed out - we should rather try stop having all sources in /
13:16:52 <mnasiadka> I'd be happy to move them somewhere to /opt/kolla/sources or something like that and stop that madness
13:17:32 <mnasiadka> ok, let's go to additional agenda from the whiteboard
13:17:38 <mnasiadka> #topic Additional agenda (from whiteboard)
13:17:55 <kevko> why not, but i really don't understand why there is so offence against symlink if it is regular tool in linux ...also update-altenatives works with symlinks
13:18:53 <mnasiadka> I don't have an offence, but we discussed that many cycles ago, and some former cores and PTLs were against it :)
13:19:19 <mnasiadka> So trying to get all cores together to form a decision which one is better, that's it - I'm not the resident Python expert here
13:19:41 <SvenKieske> kevko: I was not talking about a specific usecase in kolla, more in a general sense, you might want to watch this video if you want to know why I hate them: https://www.youtube.com/watch?v=4JrY-DntoyU&list=PLbw4szFfveGoMAryhcaaWVOjqU5eBwLoY&index=10 if you like text, I encourage you to read https://lwn.net/Articles/899543/ instead
13:19:57 <mnasiadka> Ok, I have a sort of a problem - SvenKieske has overloaded the Additional Agenda with RMQ
13:20:11 <jovial> For those not in the know, What is the removal of distro_python fixing? Not much in the commit message.
13:20:20 <SvenKieske> mnasiadka: ah wait, I guess that can be removed, we discussed this last week
13:20:21 <kevko> SvenKieske: thanks, i will check it ....thanks
13:20:39 <mnasiadka> jovial: basically every time we upgrade Ubuntu/Debian/Rocky (choose one) we need to bump distro_python_version
13:20:56 <jovial> Ahh, I see - thanks
13:21:05 <mnasiadka> so if we drop that variable completely - we don't need to do the dance with passing new version from Kolla jobs to Kolla-Ansible jobs, then changing it in Kolla-Ansible, and then reverting the changes in Kolla
13:21:50 <mnasiadka> mattcrees: looking at your comments with packages - I'm pretty sure we don't use PackageCloud for long time
13:22:20 <mnasiadka> mattcrees: we moved to CloudSmith, and then to some weird RMQ mirror on somebody's computer/laptop, anyway - it all works now, so I doubt we're affected
13:22:22 <kevko> and for example for me if i am patching python sources ..i don't need to provide several same patch files with different headers for every distro for every customer ..just one ..because path is /var/lib/kolla/venv/lib/python3
13:23:17 <jovial> In some circumstances do you not get the venv/bin/python3 link then?
13:23:53 <mnasiadka> binaries yes, but we point wsgi scripts to /var/lib/kolla/venv/lib/python{version}/something
13:24:27 <mnasiadka> while symlinking binaries seems to be a sort of a standard, I haven't found anybody that does lib/python{version} symlinking to lib/python ;-)
13:24:48 <jovial> got you - thanks for the clarification
13:25:12 <mnasiadka> let's go further
13:25:18 <kevko> "One day, I would really like to see that kolla-ansible can use any image from kolla... sometimes it's good to take ovs from ubuntu, apache from centos, and so on... distro_python_version sometimes throws a wrench into that.
13:25:22 <kevko> yep
13:25:27 <mnasiadka> mhiner [2024-09-11]
13:25:27 <mnasiadka> please review KOLLA_SERVICE_NAME in Podman fix https://review.opendev.org/c/openstack/kolla-ansible/+/927997
13:25:58 <kevko> i will check it after meeting ..i've already checked and it looked good for me ..just want to have second round
13:26:03 <mnasiadka> sure
13:26:03 <mattcrees> mnasiadka: good to know, I'd foolishly assumed our docs were correct ;) https://github.com/openstack/kolla/blob/stable/2023.1/doc/source/contributor/versions.rst Looks like the rabbit versions haven't been updated since Zed, I should fix that
13:26:27 <mnasiadka> r-krcek [2024-09-11
13:26:27 <mnasiadka> please review https://review.opendev.org/c/openstack/kolla-ansible/+/923110
13:26:27 <mnasiadka> please review https://review.opendev.org/c/openstack/kolla-ansible/+/925444
13:26:33 <mnasiadka> mattcrees: makes sense
13:27:19 <mnasiadka> will have a look into those two
13:27:37 <r-krcek> Thank you
13:28:42 <mnasiadka> actually second one has my +2
13:28:57 <mnasiadka> bbezak, frickler, SvenKieske, kevko if you could have a look it would be nice
13:29:15 <mnasiadka> ok, whiteboard done
13:29:18 <mnasiadka> #topic Open discussion
13:29:31 <SvenKieske> https://review.opendev.org/c/openstack/kolla-ansible/+/925444 my first +2 ;)
13:29:40 <mmalchuk> congrats
13:29:55 <mmalchuk> Kayobe cores please review https://review.opendev.org/c/openstack/kayobe/+/921628
13:30:02 <mnasiadka> Yes, wanted to mention I've added Sven to kolla-core and kolla-ansible-core access lists
13:30:05 <mnasiadka> congrats SvenKieske
13:30:07 <kevko> mnasiadka: approved ..it's really simple patch just moving var
13:30:25 <SvenKieske> thanks all btw for enduring my stubbornness :)
13:30:29 <mattcrees> well deserved SvenKieske :)
13:30:57 <mmalchuk> one of the lost backports https://review.opendev.org/c/openstack/kolla-ansible/+/927665
13:30:58 <kevko> SvenKieske: cg
13:31:02 <mmalchuk> please merge
13:31:29 <mnasiadka> anybody have any additional topics?
13:33:00 <kevko> not from me probably ...
13:33:08 <bbezak> nice work SvenKieske
13:33:18 <SvenKieske> kevko: the proxysql needs some more work I guess, looking at CI: https://review.opendev.org/c/openstack/kolla-ansible/+/913724
13:33:47 <kevko> just want to mention that i was reviewing LE patch from Piotr (If i remember ) and reworking little bit ...
13:33:48 <SvenKieske> not sure where the exact error is yet, the recovery play does throw an error
13:34:31 <mnasiadka> kevko: yes, thanks for that - he has also some additional patch to support account binding
13:35:12 <kevko> SvenKieske: will check ...in few hours i will post working LE i hope ...than i will check proxysql
13:35:18 <kevko> mnasiadka: will also check
13:35:40 <mmalchuk> trivial backports: https://review.opendev.org/q/I7e30ee5754ca2ecf9d4a2d3519f9e0bdf3f0aeea
13:35:44 <mmalchuk> please merge
13:35:44 <mnasiadka> SvenKieske: https://ac1ddfca6aef635c4a17-a6bd106a0a856fe942ecaed6f7524924.ssl.cf1.rackcdn.com/913724/14/check/kolla-ansible-rocky9-mariadb/6bec05d/secondary1/logs/container_logs/mariadb_clustercheck.txt
13:36:19 <mnasiadka> are we sure clustercheck should be running? :)
13:36:48 <SvenKieske> the task has a negative condition on proxysql, afaik
13:36:58 <SvenKieske> maybe there's still something missing
13:37:23 <kevko> SvenKieske: i know what is going on
13:37:51 <mnasiadka> ok then, I guess that's it for today
13:37:53 <mnasiadka> thanks for coming!
13:37:57 <mnasiadka> #endmeeting