13:00:04 <mnasiadka> #startmeeting kolla 13:00:04 <opendevmeet> Meeting started Wed Sep 18 13:00:04 2024 UTC and is due to finish in 60 minutes. The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:04 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:04 <opendevmeet> The meeting name has been set to 'kolla' 13:00:17 <mnasiadka> #topic rollcall 13:00:31 <SvenKieske> o/ 13:00:31 <jovial> o/ 13:00:46 <mnasiadka> o/ 13:01:00 <mhiner> o/ 13:01:18 <mattcrees> o/ 13:01:22 <kevko> \o/ 13:01:30 <mmalchuk> o/ 13:02:01 <mnasiadka> #topic agenda... (full message at <https://matrix.org/oftc/media/v1/media/download/AZ8DWFB-4WLSxI_dYsDe3rGB2f2-z4LGcJb4q5sdro8hxmDJkV27TceKI56yr1eU0DklwzJ4v-E7o_sAvwpVz_9CeSBvEpawAG1hdHJpeC5vcmcvQkpIWHFGeldrbFpGS2NuWW9WV29oQlN1>) 13:02:06 <mnasiadka> #topic CI status 13:02:11 <mnasiadka> Anything bleeding red? 13:02:46 <opendevreview> Michal Arbet proposed openstack/kolla-ansible master: [CI] Add mariadb backup into mariadb scenario https://review.opendev.org/c/openstack/kolla-ansible/+/929771 13:02:57 <mnasiadka> I guess not 13:03:03 <mnasiadka> #topic Release tasks 13:03:27 <mnasiadka> It's R-2 - release in two weeks 13:03:44 <mnasiadka> I'll try to post Highlights for Kolla, Kolla-Ansible and Kayobe so marketing team is happy 13:04:24 <jovial> Is feature freeze this week? 13:05:03 <mnasiadka> Yes, the docs say R-2 is Feature Freeze 13:05:18 <mnasiadka> Although freeze exception just needs two core agree on reviewing a patch 13:05:26 <mnasiadka> *two cores 13:06:56 <jovial> Got you - I will start marking kayobe patches for the the release as RP+1 13:07:18 <mnasiadka> Since we have more or less 3 months to release - we need to agree if we're still going to push in Ubuntu 24.04 LTS support 13:07:35 <mnasiadka> Since it's sort of in a good condition in Kolla and Kolla-Ansible - but I didn't start anything in Kayobe 13:08:41 <mnasiadka> We don't have to do it this cycle, but I guess it would be nice to get it in 13:08:50 <mnasiadka> Anybody any strong opinions? 13:09:21 <SvenKieske> not for kayobe unfortunately :) 13:09:25 <kevko> well, I've asked lastly about the k-a role patch vs kolla drop kolla-distro ...no feedback 13:09:43 <jovial> Nice - anything still in flight for 24.04 in kolla/kolla-ansible or has everything landed? 13:09:56 <mnasiadka> kevko: frickler seems to be a bit off, bbezak is on sick leave - let's wait another week? 13:10:12 <mnasiadka> jovial: https://review.opendev.org/q/topic:%22kolla-ubuntu-noble%22 13:10:52 <mnasiadka> jovial: Jakub should be working on it, I assume we would need to at least see if it's trivial to get it in Kayobe or not 13:11:10 <mnasiadka> So let's check and then decide next week 13:11:19 <kevko> mnasiadka: okay 13:11:35 <mnasiadka> We sort of already covered this release planning 13:11:45 <mnasiadka> #topic Additional agenda (from whiteboard) 13:12:02 <mnasiadka> jovial: Schedule data for kayobe PTG if not already discussed 13:12:19 <jovial> date* 13:12:51 <mnasiadka> jovial: As I mentioned last time, Kolla PTG slots will start from Tue 22nd (because I'm coming back from US on Monday) 13:13:17 <mnasiadka> So I think we could schedule some Tue/Wed slots for Kolla and a Thu slot for Kayobe? 13:13:32 <jovial> Yep, that works for me. Just wanted to get something in the diary 13:13:33 <mnasiadka> 22nd October of course 13:13:47 <mnasiadka> Ok, I'll hande reserving the bot slots and creating the Etherpad 13:13:52 <mnasiadka> And send a mail to openstack-discuss 13:14:17 <jovial> thanks 13:14:22 <mnasiadka> jovial: Just to announce that I have proposed to add Matt Crees as a Kayobe core on the mailing list if no one has seen the message 13:14:32 <mnasiadka> Yes, seen that, didn't have time to respond, but +1 from me 13:14:33 <opendevreview> Piotr proposed openstack/kolla-ansible master: Gnocchi Configuration file - Deprecated: Option "enable_proxy_headers_parsing" https://review.opendev.org/c/openstack/kolla-ansible/+/929477 13:15:25 <jovial> Just raising awareness... feel free to reply to the email on openstack-discuss with any comments 13:15:35 <mnasiadka> Thanks 13:15:46 <mnasiadka> That's all for additional agenda 13:15:49 <mnasiadka> #topic Open discussion 13:15:52 <mnasiadka> Anybody, anything? 13:16:02 <chembervint> hi, can we discuss my patch https://review.opendev.org/c/openstack/kolla-ansible/+/920377 13:17:20 <SvenKieske> I prepared some OIDC related small patches, if someone could look at those: https://review.opendev.org/c/openstack/kolla-ansible/+/928940 https://review.opendev.org/c/openstack/kolla-ansible/+/928949 13:18:13 <chembervint> It seams that the last question according 920377 is a documention - should we add something or not 13:18:27 <mnasiadka> chembervint: sure - from my perspective the case is clear - either we're moving the octavia interface hacky feature written for CI to production status and we rework the docs for it to sound like that - or we're keeping that for Test/Dev like it is now - and then nobody should be doing big scale in Test/Dev :) 13:20:11 <mmalchuk> please review and merge backports: https://review.opendev.org/q/(project:openstack/kolla+or+project:openstack/kolla-ansible)+(branch:unmaintained/zed+or+branch:unmaintained/yoga)+status:open 13:20:17 <mmalchuk> yoga CI fixed 13:20:19 <mnasiadka> SvenKieske, kevko any opinions on chembervint 's change? 13:20:26 <chembervint> mnasiadka: our customers (our customers are operators;) ) using this feature in the prod environment with a scale up to 100computes. and actually, I think it should be prod ready feature, because not all DC are ready to spread L2 between racks 13:20:45 <kevko> mnasiadka: just checking ... 13:21:19 <SvenKieske> well LGTM, it's a minor detail if it's not working in all cases, we use this in prod already, documented or not :) 13:21:43 <SvenKieske> sure would be nice to have more clear docs, as always, but not a deal breaker to me. 13:21:48 <kevko> mnasiadka: my opinion is that octavia stuff in kolla-ansible - i mean that part with interface etc ...is not for production ... 13:22:00 <SvenKieske> sorry, can't write that much, swamped with upgrade work 13:22:36 <SvenKieske> so leave the docs as is, merge the fix don't advertise it for prod but we still use it in prod? fine with me :) 13:22:39 <mnasiadka> That's also my problem - we're injecting a systemd unit for octavia-interface, which is if I remember correctly a bash script outside network manager or any other interface management subsystem 13:23:12 <SvenKieske> but we do this already, this is not new, is it? 13:23:27 <mnasiadka> this was meant for CI and development environments 13:23:29 <kevko> on the other side, if everywhere it will be specified that it's not for production ..why not ... 13:23:54 <mnasiadka> ok then, I can live with big scale under test/dev 13:23:56 <mnasiadka> let me merge it 13:23:59 * frickler sneaks in and reads backlog 13:24:54 <mnasiadka> done 13:25:12 <chembervint> as for me - it could be used also for production, if we understand how to tune it (exactly what I've added to the docs) 13:25:51 <kevko> Then it's the operator's responsibility. The problem is that as soon as we start investing time into similar patches, people will adopt them, and that's a path to the hell... So once again, I don't have a problem with it, but the pitfalls of this solution must be clearly described everywhere... and that it's NOTexactly ideal to use in production. 13:26:01 <SvenKieske> maybe frickler has to add something as I think he implemented the octavia interface stuff on our side :) 13:26:32 <mnasiadka> From my perspective that octavia-interface thing is more geared towards Kayobe than Kolla-Ansible 13:26:51 <mnasiadka> I mean it's easier to implement it there, maybe in NetworkManager and with static IPs 13:26:56 <frickler> well it is a hacky solution, somehow abusing the neutron mechanism I think 13:26:59 <mnasiadka> and then it might be something we could recommend in production 13:27:15 <mnasiadka> But for now, it's hacky :) 13:27:27 <chembervint> NM is a good idea, but not all linux distros support it in the same way ... ubuntu - netplan. centos - NM .... and so on 13:27:52 <frickler> yes, switching to static configuration is on our todo list for sure 13:28:36 <chembervint> yes, I agree, it's hacky. but as it has been discuses in the patch comments - probably we have to fix the current problem. and next think how to rework it in the master? 13:29:52 <frickler> I also see that I had +2d the patch already, so seems to be fine 13:30:25 <mnasiadka> Yes, I silenced my Asperger 13:30:45 <kevko> is it already time to dicsuss patches then ? :D 13:30:46 <mnasiadka> ok, unless frickler has anything else - I'd be closing the meeting :) 13:30:57 <mnasiadka> ah 13:30:59 <kevko> I have something on frickler :D 13:31:08 <chembervint> thank you! 13:31:28 <mnasiadka> frickler: did you have time to look into kevko 's hacky lib/python linking approach vs my service role? :) 13:31:39 <chembervint> and one more short question - could you please push the button on the backports for merged patch? https://review.opendev.org/c/openstack/kolla-ansible/+/928077 13:31:58 <kevko> mnasiadka: i will still argue that's not hacky ... symlink to directory is normal in linux 13:32:14 <kevko> example :) 13:32:16 <kevko> (nova-compute)[root@compute0 /]# ls -la /lib 13:32:16 <kevko> lrwxrwxrwx 1 root root 7 Apr 27 02:02 /lib -> usr/lib 13:32:30 <kevko> var/lib/kolla/venv/lib64 -> Points to: /var/lib/kolla/venv/lib 13:32:42 <kevko> var/spool/mail -> Points to: /var/mail 13:32:52 <kevko> /etc/xdg/systemd/user -> Points to: /etc/systemd/user 13:32:54 <kevko> etc etc etc 13:33:12 <frickler> mnasiadka: kevko: yes, I looked at it, but didn't make up my mind yet. I think kevko's solution is easier and I tend to prefer that 13:33:12 <kevko> /bin -> Points to: /usr/bin 13:33:50 <mnasiadka> kevko: easy, not kilometer long monologue - we've been there already, everybody knows you want that merged 13:34:32 <mnasiadka> let frickler and others make up their mind, it's not a race 13:34:48 <kevko> but i didn't want to discuss this ... i wanted to just ask to another review for https://review.opendev.org/c/openstack/kolla/+/829295 https://review.opendev.org/c/openstack/kolla/+/915440 https://review.opendev.org/c/openstack/kolla/+/929516 13:34:52 <frickler> I place a vote after the meeting 13:35:19 <frickler> mnasiadka: I also would like to see 24.04 support for 2024.2, do you think you could use help for that? 13:35:43 <frickler> (though I'll also note that I'll be offline for some days starting tomorrow) 13:36:19 <kevko> well, that symlink also helps me to have one patch for every distro as i need to patch for several customers and it's nice to has one header in patch file ... for this patch https://review.opendev.org/c/openstack/kolla/+/829295 13:37:19 <mnasiadka> geez, 211 files in one patch 13:37:39 <kevko> mnasiadka: i squashed two commits into one ... 13:37:46 <kevko> mnasiadka: but i can unsquash it 13:38:10 <kevko> mnasiadka: all 208 files are same 13:38:24 <mnasiadka> frickler: basically kolla/kolla-ansible parts are ready - there's a bug in python3-docker in Ubuntu that is getting fixed (so we'll use noble-proposed repo for some time) - I don't think I need any more help in that, and Kayobe is rather on SHPC side 13:38:39 <mnasiadka> kevko: no need, just love going through 208 files to check if they are really the same 13:38:59 <kevko> mnasiadka: don't click expand all :D 13:40:10 <mnasiadka> anyway 13:40:37 <mnasiadka> I guess we can wrap up 13:40:52 <mnasiadka> Thanks for coming - see you next week! 13:40:53 <mnasiadka> #endmeeting