13:00:01 <mnasiadka> #startmeeting kolla
13:00:01 <opendevmeet> Meeting started Wed Apr 17 13:00:01 2024 UTC and is due to finish in 60 minutes.  The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:01 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:01 <opendevmeet> The meeting name has been set to 'kolla'
13:00:12 <mnasiadka> #topic rollcall
13:00:13 <mnasiadka> o/
13:00:43 <mhiner> o/
13:00:46 <ihalomi> \o
13:00:57 <SvenKieske> o/
13:01:56 <frickler> \o
13:02:53 <mnasiadka> #topic agenda
13:02:53 <mnasiadka> * CI status
13:02:53 <mnasiadka> * Release tasks
13:02:53 <mnasiadka> * Regular stable releases (first meeting in a month)
13:02:53 <mnasiadka> * Current cycle planning
13:02:55 <mnasiadka> * Additional agenda (from whiteboard)
13:02:55 <mnasiadka> * Open discussion
13:02:58 <mnasiadka> #topic CI status
13:03:21 <mnasiadka> Basically all green - apart the gnocchi issue that we fixed (and Kolla is now using Caracal sources for building)
13:03:39 <mnasiadka> And Octavia has problems tearing down an amphora LB - claims Nova did not free up the port
13:03:45 <mnasiadka> I'm having a slow look into that
13:03:59 <mnasiadka> #topic Release tasks
13:04:15 <mnasiadka> So basically - we should start preparing Kolla and Kolla-Ansible for branching
13:04:25 <mnasiadka> I'd like to squeeze ovn-bgp-agent in
13:04:51 <mnasiadka> There are patches for adding Scaphandre (energy consumption monitoring for processes/instances) which would be also nice to get in
13:05:24 <mnasiadka> So I'll be running around asking for reviews surely and forming up a queue of RP+1 things that need to get in
13:05:45 <mnasiadka> I'll also create a list on the Whiteboard - so we can track that properly
13:06:12 <SvenKieske> nice, I guess we would also happily review scaphandre, as we already use that downstream
13:06:17 <frickler> +1
13:06:33 <mnasiadka> Goodie
13:06:36 <mnasiadka> Let's move on
13:06:43 <mnasiadka> #topic Current cycle planning
13:07:07 <mnasiadka> vPTG was last week, I'll formulate some priorities for D cycle - but for now the focus is getting closer to releasing Caracal
13:07:41 <mnasiadka> From another front - our meeting is 13:00 UTC now - any volunteer to update the meetings page for summer daylight savings period?
13:09:28 <frickler> didn't we do that? I updated the channel topic at least, can also push another patch
13:10:07 <SvenKieske> I'm currently sorting out my local dev setup issues, else I would say I can do it, but I have various issues due to my fallback machine being windows based :(
13:10:24 <opendevreview> Michal Nasiadka proposed openstack/kolla master: WIP: kolla-toolbox: Use py311 on rpm distros  https://review.opendev.org/c/openstack/kolla/+/916102
13:10:29 <opendevreview> Dawud proposed openstack/kolla master: Change logic for plugins in grafana image build  https://review.opendev.org/c/openstack/kolla/+/913184
13:11:06 <mnasiadka> frickler: https://meetings.opendev.org/#Kolla_Team_Meeting says 14:00 UTC
13:11:18 <frickler> ok, will update
13:11:21 <mnasiadka> thanks!
13:11:30 <mnasiadka> #topic Additional agenda (from whiteboard)
13:11:35 <mnasiadka> Let's go through that, it's loooong
13:11:45 <mnasiadka> (SvenKieske): https://review.opendev.org/c/openstack/kolla-ansible/+/909912/comment/2da918f8_2a5d1097/ is a complete copy of "service-cert-copy" role ok for one task, or should it rather be extended instead?
13:12:01 <mnasiadka> SvenKieske: I think there's already separate patch for that?
13:12:03 <SvenKieske> that was actually addressed I believe, there is a patchset
13:12:06 <SvenKieske> yes
13:12:16 <SvenKieske> but it could get some reviews :)
13:12:26 <mnasiadka> (SvenKieske): cores: please review the following:
13:12:26 <mnasiadka> ovn-exporter https://review.opendev.org/c/openstack/kolla-ansible/+/855498
13:12:27 <mnasiadka> skyline SSO https://review.opendev.org/c/openstack/kolla-ansible/+/905860
13:12:27 <mnasiadka> neutron service limit https://review.opendev.org/c/openstack/kolla-ansible/+/877776
13:12:27 <mnasiadka> add new sysctl role: https://review.opendev.org/c/openstack/kolla-ansible/+/912351
13:12:27 <mnasiadka> cell0 VIP change fix https://review.opendev.org/c/openstack/kolla-ansible/+/910924
13:12:29 <mnasiadka> oh boy
13:13:17 <frickler> and that's already only half of what was there on Monday ;)
13:13:24 <SvenKieske> skyline, sysctl and afaik neutron service limit are rather uncontroversial imho :)
13:14:29 <SvenKieske> especially the skyline stuff would be nice for downstream users using that, also it would be nice for the new contributors seeing there patches being merged sometime :D
13:14:40 <SvenKieske> the sysctl is nice to have i guess, not that important
13:14:52 <mnasiadka> ok, I think all of them except ovn-exporter have my +2
13:15:06 <SvenKieske> the cell0 stuff depends on how much we expect kolla vip to change
13:15:12 <mnasiadka> ovn-exporter is failing all over the place
13:15:29 <SvenKieske> yeah I need to triple check something about ovn, but my local test env has currently issues :( will report back
13:15:35 <mnasiadka> usually we don't, but there's no sense in breaking people doing that, if there's a contributor willing to fix it
13:16:06 <mnasiadka> ok then, next list
13:16:09 <mnasiadka> (mhiner) status of adding python3-docker to ospbo repository
13:16:09 <mnasiadka> please review: https://review.opendev.org/c/openstack/kolla-ansible/+/911417
13:16:09 <mnasiadka> and its following change: https://review.opendev.org/c/openstack/kolla-ansible/+/911505
13:16:09 <mnasiadka> plus the podman fix: https://review.opendev.org/c/openstack/kolla-ansible/+/912521
13:16:09 <mnasiadka> this one is also ready for reviews, it fails only because of docker--py 5.0.3 version on Debian: https://review.opendev.org/c/openstack/kolla-ansible/+/908295
13:16:37 <mhiner> Podman fix is merged, thanks for that
13:17:00 <opendevreview> Matúš Jenča proposed openstack/kolla-ansible master: Add backend TLS between MariaDB and ProxySQL  https://review.opendev.org/c/openstack/kolla-ansible/+/909912
13:17:27 <opendevreview> Matt Crees proposed openstack/kayobe master: Docs: add a link to the gerrit review dashboard  https://review.opendev.org/c/openstack/kayobe/+/916110
13:18:26 <ihalomi> few weeks ago we had discussion about adding python3-docker to ospbo repository or changing the way it is installed on the newest debian
13:18:31 <ihalomi> any news about it?
13:18:36 <mnasiadka> ihalomi: can you reply to the comments in 911417?
13:18:52 <mnasiadka> and resolve those that are answered ;
13:18:53 <mnasiadka> ;)
13:18:57 <SvenKieske> also https://review.opendev.org/c/openstack/kolla-ansible/+/911417 needs a manual rebase, just checked
13:19:49 <ihalomi> mnasiadka: i did, i was waiting for SvenKieske to close it or add more comments, since they are his suggestions
13:19:54 <mnasiadka> So what about 908295 ?
13:20:56 <SvenKieske> I did an initial review of that, I'll do another round today.
13:21:05 <mnasiadka> It's failing
13:21:06 <SvenKieske> it's quite some amount of code
13:21:27 <ihalomi> the tests on 908295 is failing only because on debian there is no way to install newer version of docker
13:21:50 <ihalomi> we talked about this few weeks ago that on debian we dont use pip to install docker for some reason
13:21:53 <SvenKieske> ah it would've been nice to note that on the change, if you know the reason.
13:22:39 <mnasiadka> well, we don't use venvs for some reason
13:22:44 <SvenKieske> that being said I faintly recall the discussion, wasn't there some agreement to make this the same for all distros? not sure what the outcome was, I remember some experimental patch in that regard at least.
13:23:22 <mnasiadka> anyway, it seems like this is not going to make it to C, unless somebody has a lot of time to make the CI work in venvs (because Debian without venv will install python3-docker deb package)
13:23:43 <ihalomi> there was 2 suggestions, one to make it same for all distros and other was to add python3-docker to repository we use for podman
13:24:16 <opendevreview> Roman Krček proposed openstack/kolla-ansible master: Optimise config  https://review.opendev.org/c/openstack/kolla-ansible/+/745164
13:24:17 <opendevreview> Roman Krček proposed openstack/kolla-ansible master: Refactor services' check-containers and optimise  https://review.opendev.org/c/openstack/kolla-ansible/+/773243
13:24:33 <mnasiadka> the same for all distros == use venv for CI, but then we would need to force users to use venv
13:24:48 <mnasiadka> Debian does not support using pip to install package globally
13:25:19 <ihalomi> look at approach for podman https://github.com/openstack/ansible-collection-kolla/blob/master/roles/podman_sdk/tasks/main.yml#L4 it adds this osbpo repository and install podman from there for debian
13:25:20 <frickler> that's the future for other distros, too
13:26:21 <ihalomi> then we would have control over which version of docker or podman we use also for non pip installs
13:26:49 <SvenKieske> afaik kevko said he can fix osbpo, no problem? (for the short term) but he's not here today it seems
13:27:19 <opendevreview> Michal Nasiadka proposed openstack/kolla master: WIP: kolla-toolbox: Use py311 on rpm distros  https://review.opendev.org/c/openstack/kolla/+/916102
13:27:38 <opendevreview> Roman Krček proposed openstack/kolla-ansible master: Refactor services' check-containers and optimise  https://review.opendev.org/c/openstack/kolla-ansible/+/773243
13:28:31 <mnasiadka> well, then you need to chase kevko
13:28:35 <mnasiadka> let's move on
13:28:39 <mnasiadka> (r-krcek)
13:28:39 <mnasiadka> thoughts on https://review.opendev.org/c/openstack/kolla-ansible/+/914997 ?
13:29:12 <opendevreview> Matúš Jenča proposed openstack/kolla-ansible master: Patch service-cert-copy role to be used w/o HAProxy  https://review.opendev.org/c/openstack/kolla-ansible/+/915901
13:30:05 <mnasiadka> It's failing for now, I rechecked that - should be probably better
13:30:16 <mnasiadka> and the last one is
13:30:17 <mnasiadka> (mgoddard): Masakari CI is passing on all branches back to stable/zed. Can we de-deprecate it?
13:30:25 <SvenKieske> well tbh I went - quickly - over the linked markdown document and the performance benefit really doesn't seem to stand out. could you maybe provide some more numbers there? afaik even some tests regressed in performance?
13:31:04 <SvenKieske> I don't doubt there might be performance benefits, but I think it should be possible to show them in a more convincing fashion :)
13:31:17 <mnasiadka> yeah, I asked in the review
13:31:21 <mnasiadka> so, Masakari
13:31:27 <mnasiadka> I think it's about this: https://review.opendev.org/c/openstack/kolla-ansible/+/914877
13:31:50 <mnasiadka> And then reverting this: https://review.opendev.org/c/openstack/kolla-ansible/+/901193
13:32:58 <frickler> so shpc want's to have that downstream and support it?
13:37:14 <mnasiadka> One of our customers is using that extensively and we have stepped up to make sure that project upstream is viable
13:37:25 <opendevreview> Merged openstack/kolla stable/2023.2: Fix OpenSearch file ownership issues  https://review.opendev.org/c/openstack/kolla/+/916081
13:39:01 <frickler> ok, I don't mind doing that, will add to my review list
13:39:45 <mnasiadka> Ok then, lists of patches are done
13:39:48 <opendevreview> Verification of a change to openstack/kolla stable/2023.1 failed: Fix OpenSearch file ownership issues  https://review.opendev.org/c/openstack/kolla/+/916082
13:39:48 <mnasiadka> #topic Open discussion
13:39:50 <mnasiadka> anybody?
13:40:29 <r-krcek> Hi. I was here at the wrong time and missed my spot for the performace upgrade regarding the use or filters.
13:41:08 <r-krcek> Any thoughts on this https://review.opendev.org/c/openstack/kolla-ansible/+/914997
13:41:33 <mnasiadka> we discussed that shortly, first of all it needs to pass CI - second of all - we'd like some marketing information how much faster is that
13:41:53 <mnasiadka> (in Kolla-Ansible case, not some general case)
13:42:27 <SvenKieske> sure, as I said above - and also commented on the patchset - maybe run some more stats to show the real benefit there. also please account for caching et cetera. I'm missing some basic description describing the benchmark environment (virtual machine, software stack with versions et cetera). good benchmarking is very very hard.
13:43:08 <r-krcek> Okay, I will try to come up with some "real-world" scenario and put the specific number into the comments of the patchset.
13:43:51 <SvenKieske> doesn't even need to be real world, but should - in the best case - be reproducible, not just some fluctuation in local dev env :)
13:45:21 <SvenKieske> I usually refer to this when it comes to benchmarking, it has many do's and dont's and many pitfalls covered: https://www.brendangregg.com/activebenchmarking.html
13:45:52 <SvenKieske> and you don't need to do all that is advised there, but it can give you some ideas what to look for :)
13:46:22 <r-krcek> Thanks for the tip :)
13:47:10 <SvenKieske> thanks, looking forward to a v2 of that performance document :) better performance is surely greatly appreciated by all :)
13:48:17 <SvenKieske> you can ping me also most of the time here if you have a question or general feedback in this regard. I did quite some benchmarking in the past and did many of the common errors myself :D
13:49:39 <r-krcek> Thank you, that would be great. I'll be sure to reach out, once I have some more questions/updates. :)
13:50:51 <mnasiadka> good
13:50:58 <mnasiadka> ok then, let's finish for today
13:51:02 <mnasiadka> Thanks for coming!
13:51:04 <mnasiadka> #endmeeting