15:01:20 #startmeeting kolla 15:01:20 Meeting started Wed Jan 5 15:01:20 2022 UTC and is due to finish in 60 minutes. The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:20 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:20 The meeting name has been set to 'kolla' 15:01:28 :-) 15:01:29 o/ 15:01:42 o/ 15:01:47 Radosław Piliszek proposed openstack/kolla-ansible master: [CI] Drop unused nodeset https://review.opendev.org/c/openstack/kolla-ansible/+/823548 15:02:03 Radosław Piliszek proposed openstack/kolla-ansible stable/xena: [CI] Drop unused nodeset https://review.opendev.org/c/openstack/kolla-ansible/+/823466 15:02:06 #topic rollcall 15:02:07 o/ 15:02:22 Radosław Piliszek proposed openstack/kolla-ansible stable/wallaby: [CI] Drop unused nodeset https://review.opendev.org/c/openstack/kolla-ansible/+/823467 15:02:40 \o 15:04:35 o/ 15:05:40 #topic agenda 15:06:12 * Roll-call 15:06:12 * Agenda 15:06:12 * Announcements 15:06:12 * Review action items from the last meeting 15:06:12 * CI status 15:06:13 * Release tasks 15:06:13 * Current cycle planning 15:06:15 * Additional agenda (from whiteboard) 15:06:15 * Open discussion 15:06:20 #topic Announcements 15:06:44 Point releases for Kolla deliverables (stable branches) are out 15:06:59 #topic Review action items from the last meeting 15:07:37 mnasiadka to triage security bugs and update them with resolution plan (if needed) 15:07:37 mnasiadka post a patch for docs - standard topics that should be discussed over PTG and then revisited in mid-cycle 15:07:37 yoctozepto remove CentOS 8 based CI jobs 15:07:37 mgoddard raise changes for stable point releases 15:07:44 I've started to triage the sec bugs 15:07:55 mgoddard did his 15:08:01 Merged openstack/kolla-ansible stable/xena: Add ironic-inspector policy configuration https://review.opendev.org/c/openstack/kolla-ansible/+/823209 15:08:10 yoctozepto: yours is in progress, right? 15:08:15 yes 15:08:24 but you can drop the action point 15:08:33 ok 15:08:42 #action mnasiadka to triage security bugs and update them with resolution plan (if needed) 15:08:49 #action mnasiadka post a patch for docs - standard topics that should be discussed over PTG and then revisited in mid-cycle 15:09:03 #topic CI Status 15:09:09 How's the CI? 15:09:23 arm64 debian builds are repaired 15:10:30 mgoddard, priteau: what about the kayobe-tox-ansible job failing on setuptools? 15:11:28 When did this happen? 15:12:02 https://zuul.openstack.org/builds?job_name=kayobe-tox-ansible 15:12:04 https://etherpad.opendev.org/p/KollaWhiteBoard - L148 15:12:43 I was not aware. Is it resolved 15:12:48 failed around christmas time 15:12:53 but now it seems it's good 15:13:12 holidays to blame 15:13:17 https://zuul.openstack.org/builds?job_name=kayobe-tox-ansible&branch=stable%2Fwallaby 15:13:25 Last failure 2021-12-28 14:48:44 15:13:41 so let's mark as GREEN and revisit if needed. 15:13:56 Same for victoria 15:14:19 updated on the whiteboard 15:14:20 let's move on 15:14:33 #topic Release tasks 15:15:10 Today is R-12 week 15:15:33 next task is at R-8 (beginning of Feb) 15:15:43 #topic Current cycle planning 15:16:39 Anybody wants to discuss any of their feature changes? 15:17:54 We have a lot of prio items on the Whiteboard without any owner - maybe somebody wants to sign up to something? 15:18:03 Line 312 15:19:15 I can't commit more time 15:21:37 yoctozepto: understandable 15:24:14 ok then, let's move on 15:24:39 #topic CentOS 8 testing removal on Kayobe pre-Victoria means removing all Kayobe testing - per https://review.opendev.org/c/openstack/kayobe/+/823266 15:25:20 Does that mean we should EOL Kayobe pre-Victoria? 15:26:57 I know of a company using Kayobe which still has customers on older releases 15:27:11 We all know that company 15:27:38 So should we keep ussuri and train open for commits with minimal testing only? 15:27:50 Or should that company switch to use a fork 15:27:55 minimal testing means unit testing and ansible lint at most ;-) 15:28:02 Better than nothing 15:28:14 Question what would break, if we would run the same testing on c8s in pre-Victoria 15:28:30 Merged openstack/kayobe stable/wallaby: ipa: Use openstack_branch instead of master https://review.opendev.org/c/openstack/kayobe/+/823354 15:28:38 if it's ok for you, I will proceed with removal 15:28:39 Merged openstack/kayobe stable/wallaby: Build neutron-sriov-agent image when enabled https://review.opendev.org/c/openstack/kayobe/+/823460 15:29:03 mgoddard: opinions? 15:29:52 It might be good to be able to merge fixes for other Kayobe users too. There might be people still on Ussuri and Train. 15:30:26 But we should add warnings that it is not actively tested / maintained 15:30:29 I'd suggest we keep these branches open with whatever testing is available 15:30:56 we'll just need to take a bit more care when merging 15:31:38 it is what I wanted to know 15:32:06 Would we be able to run some jobs on c8s without massive changes? 15:32:18 Seems reasonable, just wanted a discussion. 15:32:25 if you want to go that route, then I'm not doing it ;-) 15:32:50 Well, the change in Victoria needs to be merged, that's logical. 15:33:54 I can invest little time in trying to run Ussuri on Stream, if that works - we can survive with that - if not - we just remove the CI depending on CentOS Linux 8. 15:35:13 So - let's merge the Victoria change and hold off before the next meeting with removing those jobs pre-Victoria - I'll do a dumb try to run Ussuri on Stream. 15:35:57 * yoctozepto likes this 15:36:34 I'm not sure how much we'd gain 15:37:09 unless we're going to do it properly and support ussuri on stream 15:37:11 Merged openstack/kayobe stable/xena: Build neutron-sriov-agent image when enabled https://review.opendev.org/c/openstack/kayobe/+/823459 15:37:24 might be issues with RDO? 15:37:27 Ussuri and beyond are EM, so we wouldn't be able to post point releases - but I'm not a fan of keeping something alive with basically no testing 15:38:21 mgoddard: I thought we would keep using c8 containers, using the latest that are published on docker hub, but use c8s host from opendev 15:38:36 Yes, true - for sure some packaging issues - if we use RDO in Kayobe (and I thought we don't) 15:38:59 ok, c8 containers would be easier 15:41:48 Ironically... those operators that use C7 and Train are in a better position than C8 ones :) 15:41:58 indeed 15:43:20 all the more reasons to go debian-only :-) 15:43:20 not really - they are basically stuck there 15:43:29 and even more :-) 15:44:37 so, c8 containers on c8s, I'm willing to try if that works in CI - if it doesn't we just keep them in bare testing regime (them = Train + Ussuri) 15:44:42 agreed? 15:44:45 +1 15:45:08 +1 15:45:45 goodie, then I'll try (on Monday/Tuesday) - I'm off for the next two days (like probably most of PL people) 15:45:51 would at least provide a route from train to ussuri 15:46:07 true, for those poor people 15:46:27 Ok, last topic 15:46:31 #topic Open discussion 15:47:58 Anyone? Anything? 15:48:04 not a thing 15:48:13 fluentd cleanup and unpin stack should be ready for reviews 15:48:15 https://review.opendev.org/c/openstack/kolla/+/821695 15:48:41 well, except for shameless plugs it seems :D 15:49:28 Downgrade ES gems, oh boy 15:49:43 The move from Elasticsearch to OpenSearch has become more important due to the log4j issues, because we are locked on old packages 15:49:47 So we upgrade td-agent, but downgrade ES plugins? Is it even worth it? ;-) 15:51:09 would opensearch have their own gems? 15:51:16 it would be good to unpin td-agent 15:51:19 that's one thing 15:51:30 elasticsearch is pita atm 15:52:23 let's roll out rsyslog-based solution lol 15:55:05 Or replace Elasticsearch with something else ;) 15:56:24 frickler: https://github.com/opensearch-project/opensearch-ruby - there is this, don't know if that can be used with td-agent 15:56:25 well, rsyslog is that thing (o_o)' 15:56:49 https://github.com/uken/fluent-plugin-elasticsearch/issues/915 15:56:58 mnasiadka, frickler: it would need support in the plugin 15:57:00 or another plugin 15:57:01 yeah 15:57:04 I feel OpenSearch is a bit of a hell hole today 15:57:16 How about - Venus (Log management service) 15:57:51 spatel: we don't know yet what it really offers; it seems to integrate with elasticsearch so it leaves it around 15:58:16 mnasiadka: yeah, that is one fugly situation 15:58:36 also, venus does not exist on earlier branches 15:58:43 it's still in review on master heh 15:59:34 ehh, maybe we need to start thinking what to do with this crap 15:59:35 ES is pretty solid solution for logging, I am new here so not sure what is the difficulty to run it with kolla :) 15:59:52 time to finish the meeting 15:59:54 thanks for coming! 15:59:56 #endmeeting