13:00:33 #startmeeting kolla 13:00:33 Meeting started Wed May 31 13:00:33 2023 UTC and is due to finish in 60 minutes. The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:33 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:33 The meeting name has been set to 'kolla' 13:00:37 #topic rollcall 13:00:38 o/ 13:01:05 \o 13:01:06 o/ 13:01:17 q/ 13:01:17 o/ 13:01:17 o/ 13:01:17 o/ 13:01:40 o/ 13:01:56 o/ 13:03:20 #topic agenda 13:03:20 * Announcements 13:03:20 * CI status 13:03:20 * Release tasks 13:03:20 * Regular stable releases (first meeting in a month) 13:03:21 * Current cycle planning 13:03:21 * Additional agenda (from whiteboard) 13:03:23 * Open discussion 13:03:26 #topic Announcements 13:03:44 First of all, kolla/kolla-ansible/ansible-collection-kolla and kayobe have had their antelope (2023.1) rc1 13:04:24 second - I'm off for the next week, and then in two weeks from now I'm in Vancouver for OpenStack Summit 13:05:01 bbezak has agreed to host the weekly meetings, unless you agree those are not required - then we can cancel them 13:05:46 #topic CI status 13:06:17 not really required, but reviews postponed for three weeks (at least in kayobe) :( 13:06:21 I think it's green-ish, not counting some cephadm-upgrade jobs that I'm trying to fix 13:07:02 #topic Release tasks 13:07:31 not. CI zed and yoga broken due baremetal deploy problems in rl9 13:07:53 related fix: https://review.opendev.org/c/openstack/kayobe/+/883727/2 13:08:49 kayobe only 13:09:02 why oh why this patch didn't have review priority 13:09:11 and why it's not written on the kolla whiteboard 13:09:14 well, added RP+2 13:09:24 ok then 13:09:42 after 2023.1 rmq pin fix merges - I think we should push for rc2 and final 13:10:03 and work out rmq and ansible mariadb/rmq rework in master and backport that to 2023.1 after final release 13:10:14 frickler, bbezak, mgoddard, kevko: what do you think? 13:10:48 +1 13:11:37 ok then, everyone else is asleep I guess 13:11:52 when https://review.opendev.org/c/openstack/kolla/+/884815 merges - I'll post rc2 and then final 13:11:53 +1 too 13:12:17 #topic Current cycle planning 13:12:28 I was not asked; but I hope the mariadb changes don't turn out to be too complex :) 13:12:55 SvenKieske: there's a patch raised by mgoddard https://review.opendev.org/c/openstack/kolla-ansible/+/884182 13:14:03 but we'd need first to make cephadm-upgrade jobs better in https://review.opendev.org/c/openstack/kolla-ansible/+/884488 13:14:05 yeah I know, I'm on the reviewer list :) so I hope that patch works out well, as I said, it's fairly complex/large to backport imho 13:14:50 it shouldn't be much work to backport to 2023.1 13:15:01 backporting further I would not support 13:15:11 yes, we just need to backport to 2023.1 13:15:48 ok, for current cycle - I'll clean up the whiteboard this week, add priorities for that cycle 13:15:59 obviously podman and let's encrypt are getting moved to Bobcat 13:16:09 hopefully we can find somebody interested in getting those in 13:16:38 #topic Additional agenda (from whiteboard) 13:16:54 (ihalomi): do we really need shared /run mount among containers? 13:17:00 ihalomi: didn't we discuss it already? 13:17:55 I think we did 13:18:06 I guess that topic just never got deleted from the whiteboard? but I can't honestly remember the outcome. 13:18:13 Yeah, I deleted it now 13:18:22 yes we did but outcome was that we need to look more into it 13:18:37 I guess it was agreed to add the needed /run subdirs? 13:18:57 Did you try to test doing the mount like tripleo does? I think /run is not shared in tripleo deployments? 13:19:02 about the priorities: I'm blocked by https://review.opendev.org/c/openstack/kayobe/+/881679 to continue work on repos.yaml 13:19:09 need help 13:19:15 yea I tried it but we need to create these subdirs before each start of container 13:19:36 mmalchuk: repos.yaml is in kolla, this is only to support repos.yaml in kayobe - what further work are we talking about? something in kayobe? 13:19:49 so we would have to add checklist for it into service files and systemd then runs creation before each start so it wont fail 13:19:59 ihalomi: could some tmpfiles config help? 13:20:12 ihalomi: or just create that dir in kolla_container module? 13:20:22 I use kayobe for deploy 13:21:14 but if you create it in module it gets deleted after each reboot of host 13:21:34 fricklers idea could solve it 13:21:35 can't switch to redhat repos until this solved. this means full redeploy 13:21:47 i can look into it 13:22:17 Merged openstack/kolla stable/2023.1: rabbitmq: pin to 3.11.16 https://review.opendev.org/c/openstack/kolla/+/884815 13:22:20 I think you have no other choice :) 13:22:29 yay, time to raise rc2 13:22:47 second topic from SvenKieske 13:22:49 fix for https://bugs.launchpad.net/kolla/+bug/2019557 seems to be incomplete, help in debugging would be appreciated 13:23:21 I think the best fix is to pin td-agent to 4.4.2 with older version of elasticsearch plugin 13:23:30 instead of doing that dance 13:23:42 +1 13:24:02 okay, will iterate on the patch 13:24:34 actually my bad that I didn't do that earlier 13:24:38 SvenKieske: thanks 13:24:49 #topic Open discussion 13:25:29 I want to ask about mariadb issue I have in migration patch 13:26:20 and the issue is? 13:26:24 On multinode setup, deployment fails when there is already existing mariadb volume saying I should run mariadb_recovery pb 13:26:53 Is there a way to bypass this or should I just run that playbook? 13:27:25 when you stop all mariadb containers - there is no other way than to run recovery playbook 13:27:38 well, maybe on single node we survive 13:27:48 but multinode - you would need to go host by host 13:28:03 yes, single nodes are allowed to continue 13:28:09 thanks for clarification 13:29:38 ok, anyone else? 13:31:34 I guess not 13:31:37 thanks for coming! 13:31:39 #endmeeting