14:00:01 <mnasiadka> #startmeeting kolla 14:00:01 <opendevmeet> Meeting started Wed Nov 8 14:00:01 2023 UTC and is due to finish in 60 minutes. The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:01 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:01 <opendevmeet> The meeting name has been set to 'kolla' 14:00:04 <mnasiadka> #topic rollcall 14:00:05 <mnasiadka> o/ 14:00:07 <mmalchuk> o/ 14:00:08 <jangutter> o/ 14:00:13 <jovial> \o 14:00:42 <mhiner> o/ 14:02:03 <frickler> \o 14:02:32 <bbezak> o/ 14:03:06 <mnasiadka> #topic agenda 14:03:06 <mnasiadka> * CI status 14:03:06 <mnasiadka> * Release tasks 14:03:06 <mnasiadka> * Regular stable releases (first meeting in a month) 14:03:06 <mnasiadka> * Current cycle planning 14:03:08 <mnasiadka> * Additional agenda (from whiteboard) 14:03:08 <mnasiadka> * Open discussion 14:03:10 <mnasiadka> #topic CI status 14:03:18 <mnasiadka> Looks good overall, no current weird breakages 14:03:44 <mnasiadka> Although if anybody is willing - rocky/debian ironic CI jobs have been failing long time 14:03:52 <mnasiadka> any volunteers to have a look? 14:04:48 <jangutter> I can spend 1 hour or so this week to take a look, I'm not familiar with them though. 14:05:15 <mnasiadka> ok then, thanks - if you need any help - shout :) 14:05:24 <mnasiadka> #topic Release tasks 14:05:53 <mnasiadka> So, in order to make a release - we need to merge Let's Encrypt followups - kevko do they have RP+1? 14:06:08 <mnasiadka> Basically - cores please add RP+1 to things we NEED to merge before branching 14:09:06 <mnasiadka> So I'm looking for reviewers on ephemeral disk usage in CI in Kolla-Ansible - https://review.opendev.org/c/openstack/kolla/+/899465 14:09:06 <mnasiadka> https://review.opendev.org/c/openstack/kolla-ansible/+/899466 14:09:19 <jovial> Dumb question... once the branch is cut are we not allowed to add features anymore? 14:09:49 <mnasiadka> basically that's the rule we've been adhering to - although sometimes we backport something that would not break existing users 14:10:22 <mnasiadka> if that improves user experience 14:10:33 <jovial> cool, thanks for the clarification 14:10:56 <mnasiadka> ok, kevko is not here, but let's try to find the LE followups 14:11:56 <mnasiadka> I started a list on the Whiteboard - https://etherpad.opendev.org/p/KollaWhiteBoard#L209 14:12:45 <mmalchuk> added one 14:12:58 <mnasiadka> Sven is not here, but I guess td-agent bump would be useful 14:13:37 <kevko> now 14:13:48 <mnasiadka> jangutter: etcd bump? 14:13:57 <kevko> sorry :/ 14:14:12 <mnasiadka> kevko: no problem ;) 14:14:24 <jangutter> yeah, I just queued up that one for a recheck. It's safe to go in, if folks are happy with it. 14:14:25 <jovial> For kayobe, I'm hoping to get the vGPU support in, but I need to address some review comments. Will try and do that today. 14:14:42 <mmalchuk> added 2nd review 14:15:07 <mmalchuk> #link https://review.opendev.org/c/openstack/kolla/+/899773 14:15:14 <mmalchuk> #link https://review.opendev.org/c/openstack/kolla-ansible/+/899895 14:15:36 <mmalchuk> LE reviews ^^^ 14:15:45 <kevko> thanks mmalchuk 14:15:46 <opendevreview> Michal Nasiadka proposed openstack/kolla master: CentOS/Rocky: use CentOS Cloud SIG repo instead of Delorean (Bobcat) https://review.opendev.org/c/openstack/kolla/+/899231 14:16:09 <mmalchuk> #link https://review.opendev.org/c/openstack/kolla/+/894948 14:16:10 <kevko> btw, we should also merge this ... https://review.opendev.org/c/openstack/kolla-ansible/+/900284 14:16:16 <mmalchuk> td in merge conflict 14:17:22 <mnasiadka> kevko: reviewed 14:17:34 <mnasiadka> so means td needs a lot of love 14:17:39 <kevko> ^^ it's fixing kolla_docker which is actually not working properly without container engine 14:17:40 <mnasiadka> it's not required to bump it this cycle 14:18:20 <kevko> and because we have common_options in group_vars I've also added common options to all kolla_docker occurences 14:19:06 <opendevreview> Mark Goddard proposed openstack/kayobe stable/2023.1: Stop merging config when symlink points to same file https://review.opendev.org/c/openstack/kayobe/+/900317 14:19:54 <mnasiadka> ok then, seems we have enough to review for the next couple of days 14:20:18 <mnasiadka> #topic Current cycle planning 14:20:27 <mnasiadka> I'll populate the whiteboard with priorities from the PTG 14:20:35 <mnasiadka> But for now nothing - let's focus on Bobcat for now 14:20:45 <mnasiadka> #topic Additional agenda (from whiteboard) 14:21:44 <mnasiadka> Anybody up for reviewing jsuazo's patches? 14:21:56 <mnasiadka> jsuazo: These proposals should be ready for merge / +2 uness reviewers have further comments: 14:21:56 <mnasiadka> 875983: Enable the Fluentd Plugin Systemd | https://review.opendev.org/c/openstack/kolla-ansible/+/875983 (ready for cores) 14:21:56 <mnasiadka> 885417: Configures the tap-as-a-service neutron plugin | https://review.opendev.org/c/openstack/kolla-ansible/+/885417 (missing a single +2) 14:21:56 <mnasiadka> 844614: Deploy Glance and Cinder Backup with S3 backend support | https://review.opendev.org/c/openstack/kolla-ansible/+/844614 (ready for cores) 14:22:12 <mnasiadka> Well, any core reviewer 14:22:30 <mnasiadka> kevko, bbezak, frickler, mgoddard: would be nice to get some reviews on those soon :) 14:22:36 <mnasiadka> let's go for the next one 14:22:47 <mnasiadka> mhiner - Generalize Docker specific names | https://review.opendev.org/c/openstack/kolla-ansible/+/900412 14:22:51 <mmalchuk> we need more cores) 14:23:00 <mnasiadka> I sort of agree that we could do that (rename docker to container) 14:23:17 <mnasiadka> But that will make backports a bit complicated 14:23:59 <mnasiadka> So probably we could pass for now 14:24:03 <mnasiadka> any other comments? 14:24:14 <mhiner> not really 14:24:16 <kevko> probably not 14:24:34 <kevko> actually i left a comment on mhiner patch 14:24:49 <mnasiadka> ok then 14:24:52 <mnasiadka> #topic Open discussion 14:25:01 <kevko> it's good idea ..but what about users who overriden defaults ? 14:25:25 <mmalchuk> its a bad idea 14:25:31 <mnasiadka> well, we should have at least a cycle with supporting both variables 14:25:54 <mnasiadka> to ask users to move 14:26:03 <kevko> even worse idea 14:26:46 <mnasiadka> basically we could get away with changing module name 14:26:56 <mnasiadka> but this is changing too much 14:27:46 <mhiner> okay, i will revert it to just change module name and we'll see ow that looks 14:27:51 <mnasiadka> thanks 14:28:02 <mnasiadka> anything else for discussion? 14:28:26 <mmalchuk> this will bring probles for users with overrides, for developers to backport stuff, for almost all peoples 14:28:38 <mmalchuk> *problems 14:28:58 <kevko> agree 14:29:28 <kevko> i am not saying that we shouldn't do it in some point ...but it's massive change for now i think 14:30:59 <jsuazo> hello, had to step away for a minute so i don't know if you already discussed it, by I left some patches on the board that are ready for +2's (had them but were wiped by an update) 14:31:28 <Vii> 1. Will multi-ceph be available in Bobcat? 14:31:56 <Vii> 2. can you straighten the permissions for the folders located in /var/log/kolla/ 14:31:56 <Vii> currently, for example, adding an additional filter to fluentd (td-aget) to collect logs from /var/log/kolla/openvswitch/ does not work because the "fluentd" user cannot read the files 14:32:14 <mnasiadka> Vii: 1. if a patch is merged today in master, it will be available in Bobcat 14:32:36 <mnasiadka> I think we merged a patch to fix openvswitch dir permissions, but maybe not backported 14:32:38 <mnasiadka> let me see 14:33:39 <kevko> Vii: Glad to hear that someone using my multi ceph patch :P 14:33:45 <mnasiadka> gerrit barfed on me 14:33:52 <mnasiadka> so no searching now 14:34:19 <jovial> What do people think of using capabilities to allow fluentd to read log files where we can't set the ownership easily? See: https://docs.fluentd.org/deployment/linux-capability#using-cap_dac_read_search 14:34:21 <kevko> mnasiadka: we have zed deployed somewhere if i remember ..and zed has this bug ..i am sure 14:34:24 <Vii> kevko: I'm very glad you did it :) 14:34:50 <mnasiadka> kevko: not saying that it doesn't, but it seems Gerrit and Github are not liking me - so can't find if we really patched it :) 14:34:54 <kevko> (we as Ultimum) .. 14:35:04 <kevko> mnasiadka: maybe didn't :D 14:35:04 <mnasiadka> oh, opendev works 14:35:24 <jovial> One example where you can't easily set permissions is the qemu instance logs in the libvirt container (we don't currently collect these) 14:36:37 <mnasiadka> well, libvirt container should get some revamp and we should support virtlogd 14:37:01 <mnasiadka> we could switch from monolithic to the per service approach 14:37:16 <mnasiadka> #link https://libvirt.org/daemons.html#monolithic-vs-modular-daemons 14:37:22 <jovial> interesting - I hadn't heard of virtlogd before 14:38:24 <jovial> looks promising - thanks for the link 14:39:27 <mnasiadka> ok then 14:39:41 <mnasiadka> jsuazo: I pasted your patches asking for reviews, so yes - we covered that 14:40:51 <mnasiadka> seems no more topics 14:40:58 <WJeffs35> Hey all, I was wondering if anyone is running AMD Genoa that I could pick their brains on something to compare outside of the meeting. 14:42:10 <mnasiadka> So let's finish the meeting 14:42:12 <mnasiadka> #endmeeting