14:00:06 <frickler> #startmeeting kolla 14:00:06 <opendevmeet> Meeting started Wed Aug 3 14:00:06 2022 UTC and is due to finish in 60 minutes. The chair is frickler. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:06 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:06 <opendevmeet> The meeting name has been set to 'kolla' 14:00:16 <frickler> #topic rollcall 14:00:47 <frickler> mnasiadka has excused themselves for the remainder of this week 14:00:51 <yoctozepto> o/ 14:01:05 <hrw> o\ 14:02:24 <frickler> \o/ not too crowded today it seems, so let's start 14:02:26 <frickler> #topic agenda 14:02:39 <frickler> * Roll-call 14:02:39 <frickler> * Agenda 14:02:39 <frickler> * Announcements 14:02:39 <frickler> * Review action items from the last meeting 14:02:39 <frickler> * CI status 14:02:41 <frickler> * Release tasks 14:02:44 <frickler> * Regular stable releases (first meeting in a month) 14:02:46 <frickler> * Current cycle planning 14:02:49 <frickler> * Additional agenda (from whiteboard) 14:02:51 <frickler> * Open discussion 14:03:02 <frickler> #topic Announcements 14:03:40 <frickler> as hrw has mentioned earlier, we are making good progress on creating a new cirros release, you may hold your breath 14:04:06 <frickler> anything else? 14:04:09 <yoctozepto> ok, the breath is being held 14:04:15 <yoctozepto> what can we expect from the new release? 14:04:21 <yoctozepto> will it run 10x faster? 14:04:23 <yoctozepto> :-) 14:04:36 <frickler> it will run the idle loop much faster, indeed 14:04:44 <yoctozepto> <3 it 14:04:49 <frickler> also better support for ipv6 and some other things 14:05:00 <frickler> maybe we should gather things and write a releasenote 14:05:15 <frickler> also about dropping some strange arches 14:05:38 <frickler> but maybe keep breathing for a bit, may take another day or so 14:05:43 <yoctozepto> ok 14:06:02 <frickler> #topic Review action items from the last meeting 14:06:13 <frickler> I didn't see any, something I missed? 14:06:26 <yoctozepto> likely none created 14:06:44 <frickler> #topic CI status 14:07:05 <frickler> I'm not aware of any issues, but I didn't look deeply 14:07:18 <frickler> I was happy to find the Jammy things merged when I returned 14:07:24 <yoctozepto> :-) 14:08:12 <frickler> the grafana dashboards are seemingly completely broken now, guess I'll just remove ours 14:08:21 <yoctozepto> I will not mind 14:08:35 <frickler> anything else on CI? 14:08:57 <yoctozepto> nada 14:09:09 <frickler> #topic Release tasks 14:09:21 <opendevreview> James Kirsch proposed openstack/kolla-ansible master: Add support for LetsEncrypt-managed certs https://review.opendev.org/c/openstack/kolla-ansible/+/741340 14:09:43 <frickler> I didn't prepare well for this, any known tasks? 14:10:02 <yoctozepto> what week is this 14:10:25 <yoctozepto> Aug 01 - Aug 05 14:10:25 <yoctozepto> R-9 14:10:50 <yoctozepto> R-8: Switch images to current release¶ 14:10:52 <yoctozepto> next week 14:11:02 <frickler> "R-8: Switch images to current release" is obsolete with source images only, right? 14:11:14 <frickler> because source was switched in R-17 14:11:19 <hrw> now quite 14:11:25 <yoctozepto> still relevant for deps 14:11:32 <hrw> R-8 means RDO/UCA for deps 14:11:42 <hrw> so may get newer libvirt/qemu/ceph etc 14:11:51 <yoctozepto> with jammy we are probably already there 14:11:59 <yoctozepto> and CS depends on CS9 support 14:12:06 <yoctozepto> so nothing extra to do I guess? 14:12:19 <frickler> jammy should have yoga pkgs natively, yes 14:12:25 <yoctozepto> ah 14:12:30 <yoctozepto> so UCA for Zed 14:12:52 <yoctozepto> CS9 to be seen 14:13:04 <frickler> but does canonical build that before the actual release? 14:13:05 <yoctozepto> also soon 14:13:06 <yoctozepto> R-5: Cycle highlights deadline¶ 14:13:14 <frickler> I can try to check until next week 14:13:16 <yoctozepto> need to think about these in the upcoming meetings 14:13:30 <yoctozepto> frickler: I think yes, since we added this bullet there 14:13:35 <yoctozepto> and only excluded debian 14:13:40 <hrw> frickler: they build binary packages during cycle 14:13:49 <yoctozepto> they brave then 14:14:12 <frickler> #action frickler check availability of zed UCA for Jammy 14:14:53 <hrw> frickler: http://ubuntu-cloud.archive.canonical.com/ubuntu/dists/jammy-updates/zed/ 14:15:20 <yoctozepto> well, that was quick 14:15:44 <frickler> thx hrw 14:16:03 <frickler> anything else release-wise? 14:16:45 <frickler> #topic Regular stable releases (first meeting in a month) 14:16:56 <frickler> yoctozepto: did you tackle that yet? 14:17:04 <yoctozepto> no, did not touch it 14:17:22 <frickler> do you want to? otherwise I'll do it after the meeting 14:17:40 <yoctozepto> ok, I'll try the instructions :-) 14:18:00 <frickler> ping me if it doesn't work 14:18:23 <frickler> #topic Current cycle planning 14:19:13 <frickler> any updates here? 14:20:08 <frickler> or should we rather move through the additional items? 14:20:08 <yoctozepto> proxysql got farther 14:20:29 <yoctozepto> yeah, let's go over those others unless someone speaks up 14:21:06 <frickler> #topic Additional agenda (from whiteboard) 14:21:23 <frickler> (yoctozepto) Remove nova_legacy endpoint 14:21:42 <frickler> I'm +1 on that, discussed that internally recentish 14:22:02 <yoctozepto> any objections? 14:22:15 <yoctozepto> as far as I am concerned, we can make it optional and default to not enabled 14:22:21 <frickler> maybe have a flag for someone who still needs it? 14:22:24 <yoctozepto> so that new deployments do not get it 14:22:26 <frickler> yes 14:22:31 <yoctozepto> by default at least 14:22:39 <yoctozepto> and old deployments can remove it 14:22:40 <yoctozepto> themselves 14:23:03 <frickler> did we do the same for admin endpoints? 14:23:18 <yoctozepto> no, we have just scrubbed them 14:23:25 <yoctozepto> only for keystone it is done 14:23:31 <yoctozepto> still in review 14:23:48 <yoctozepto> (someone start reviewing my patches, they are gold) 14:23:58 <frickler> the question is whether we want to delete on upgrade or not 14:23:59 <yoctozepto> (yes, I know I'm humble) 14:24:09 <frickler> the former would ensure consistent setups 14:24:16 <yoctozepto> we default to being lazy 14:24:22 <frickler> the latter would be nicer for legacy envs 14:24:22 <yoctozepto> and letting users delete themselves 14:24:55 <frickler> ok, then that should be fine for nova_legacy, too 14:25:06 <yoctozepto> btw, keystone is in https://review.opendev.org/c/openstack/kolla-ansible/+/843730 14:25:32 <frickler> yes, I'll take a look at that. soon. ish. 14:25:41 <yoctozepto> roger that 14:25:59 <frickler> (yoctozepto) The status of the dynamic OVS role assignment https://review.opendev.org/c/openstack/kolla-ansible/+/840895 14:26:13 <yoctozepto> here, we would need further input from mgoddard 14:26:18 <yoctozepto> not sure if he's around 14:26:29 <yoctozepto> I think the change is sane 14:27:04 <yoctozepto> it is a different approach to the current ones but aligns better with what we need this role for 14:27:09 <frickler> I must admit that I'm not confident approving this without some local testing, would prefer if mgoddard could re-review 14:27:51 <yoctozepto> the best test is test on a production environment, you know; weeds out the bugs fastest :-) 14:28:06 <frickler> so I guess we'll keep this topic for next time 14:28:12 <yoctozepto> indeed 14:28:25 <frickler> (frickler) Virtual PTG planning 14:28:41 <frickler> not much we can decide with again low attendance I guess 14:28:52 <yoctozepto> agreed 14:29:04 <yoctozepto> but let's do it soonish 14:29:08 <frickler> (headphoneJames) please review let's encrypt patch: https://review.opendev.org/c/openstack/kolla-ansible/+/741340 14:29:26 <yoctozepto> James has recently updated the patch 14:29:40 <yoctozepto> it's an ever-standing wish to get this merged 14:29:50 <yoctozepto> I guess it's just the right thing to look at after proxysql 14:29:57 <yoctozepto> just right 14:30:27 <frickler> oh, that's a new version now that still waits for CI results 14:30:49 <yoctozepto> yup 14:31:05 <frickler> mnasiadka: mgoddard: ^^ you set RP+1 on this one 14:31:23 <frickler> (i.halomi) Podman and systemd support status 14:31:31 <yoctozepto> the agenda point stays as a good reminder that we have low review throughput for bigger changes 14:31:39 <frickler> review of systemd with docker a current status on podman version : https://review.opendev.org/c/openstack/kolla-ansible/+/816724 14:31:52 <yoctozepto> I mean, I can easily squeeze in smaller stuff but biggur things need to wait and stand in a queue 14:32:16 <yoctozepto> yesh, this is the 3rd biggie after the proxysql and le 14:32:37 <frickler> maybe we want to create a more explicit queue like nova does? 14:32:55 <yoctozepto> what do they do? 14:33:03 <opendevreview> Mark Goddard proposed openstack/kayobe master: Support configuration of swap https://review.opendev.org/c/openstack/kayobe/+/851195 14:33:34 <frickler> I have to look up the details, but something like two slots for stacks under priority review 14:33:45 <frickler> and a queue for thinks to enter these slots 14:33:47 <frickler> things 14:34:23 <frickler> and the expectation that items in those slots get reviewed and merged within like two weeks 14:34:25 <yoctozepto> hmm, I have on slot for myself and it was occupied by proxysql thus far :D 14:34:35 <yoctozepto> ah, that surely differs 14:35:01 <frickler> so a bit more coordination between cores maybe 14:35:16 <frickler> not sure that would work for us with our diverging interests though 14:36:11 <yoctozepto> me neither 14:36:19 <frickler> anyway I guess these two patches should get some priority if possible 14:36:34 <frickler> anything to discuss about the contents yet? 14:37:14 <yoctozepto> I don't see the relevant folks here so I doubt 14:37:27 <frickler> else we'd have 14:37:30 <frickler> (frickler) Discuss Horizon/JS/xstatic situation 14:37:48 <frickler> https://lists.openstack.org/pipermail/openstack-discuss/2022-August/029825.html 14:38:14 <headphoneJames> The let's encrypt patch should be ready as far as I know 14:38:31 <hrw> frickler: this one feels tricky 14:38:35 <frickler> IMO the essence of this is that we may want to look into replacing xstatic pkgs with native JS from npm or whatever 14:39:05 <hrw> frickler: I have a feeling that we need to do 'pip install' as we have now, then track all those JS, drop them and fetch from upstreams 14:39:06 <frickler> but no idea how to actually do that 14:39:13 <frickler> or we wait for osa to implement the same and just copy what they do 14:39:14 <yoctozepto> yup, but I suppose we need guidance from the Horizon team if we don't want to spend time reverse-engineering 14:39:37 <hrw> best would be to see horizon having code for it ;d 14:39:48 <frickler> I think horizon team is mostly non existent, I wouldn't bet on them 14:40:05 <hrw> 'pip install horizon-python-stuff' + 'npm install horizon-js-stuff' 14:40:53 <yoctozepto> yeah, I'm afraid frickler is right 14:41:01 <frickler> hrw: are you interested in investigating this further? 14:41:13 <yoctozepto> tbh, the most time/cost-effective would be to get rid of horizon in the first place ;p 14:41:21 <hrw> never dealt with npm stuff etc 14:41:25 <yoctozepto> not sure how skyline fares though 14:41:30 <yoctozepto> and what pandora boxes it opens 14:42:13 <frickler> also the situation seems to have been the same over the last 10 years or so, so not sure how urgent this actually is 14:42:25 <yoctozepto> ++ 14:42:56 <frickler> maybe plan for some joint session (with distros and other deployers) at the PTG or something might be fast enough 14:43:53 <yoctozepto> agreed 14:44:28 <frickler> o.k., so I mainly wanted others be aware of the situation, which I think I achieved 14:44:44 <frickler> #topic Open discussion 14:44:56 <frickler> anything else? 14:45:08 <yoctozepto> nope 14:46:04 <yoctozepto> btw, I'm going on vacations the following 2 weeks so my availability will vary 14:46:10 <yoctozepto> and I will not be there for the meetings 14:46:28 <frickler> enjoy your time off 14:46:38 <yoctozepto> thanks 14:46:39 <hrw> yoctozepto: have fun 14:46:43 <yoctozepto> thanks 14:47:26 <frickler> o.k., then also enjoy the 12 minutes I'm returning you now, everyone ;) 14:47:32 <frickler> #endmeeting