15:01:11 #startmeeting kolla 15:01:11 Meeting started Wed Sep 22 15:01:11 2021 UTC and is due to finish in 60 minutes. The chair is mnasiadka. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:11 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:01:11 The meeting name has been set to 'kolla' 15:01:46 #topic Roll-call 15:01:56 o/. 15:02:02 o/ 15:02:03 .\o 15:02:04 \o 15:04:35 #topic Agenda 15:04:44 * Announcements 15:04:44 * Review action items from the last meeting 15:04:44 * CI status 15:04:44 * Discuss "Transition Keystone admin user to system scope" https://review.opendev.org/c/openstack/kolla-ansible/+/692179/ 15:04:44 * Should we switch CentOS to upstream MariaDB as well? And should we try to sync version for all in Xena? And which one? https://review.opendev.org/c/openstack/kolla/+/810161 15:04:46 * Should we support Rocky in images? 15:04:46 * Release tasks 15:04:48 * Xena cycle planning 15:04:48 * Yoga PTG planning 15:04:50 * Open discussion 15:05:09 [°][o] 15:05:56 #topic Announcements 15:06:04 I have none, let's move on. 15:06:11 #topic Review action items from the last meeting 15:06:31 mnasiadka to update docs encouraging to build your own containers and run your own registry 15:06:31 mnasiadka to create pull-retag-push blueprint based on kayobe RFE: https://storyboard.openstack.org/#!/story/2007731 15:06:31 yoctozepto to fix multinode docs after defaulting to secure registry 15:06:31 mnasiadka to send a mail to ML about Kolla PTG 15:06:36 haven't done mine - yoctozepto ? 15:07:32 mnasiadka: fixed in the change 15:07:43 need to merge it 15:08:00 ok 15:08:06 fair enough :) 15:08:17 #topic CI status 15:08:49 Green, but I see a note about Train failing? 15:10:23 someone cares about train then 15:10:27 (not me) 15:10:48 Well, let's see if somebody has the will to fix it then. 15:10:59 #topic Discuss "Transition Keystone admin user to system scope" https://review.opendev.org/c/openstack/kolla-ansible/+/692179/ 15:11:14 headphoneJames - so what's the current status? 15:12:00 https://www.irccloud.com/pastebin/HZG1PxBZ/ 15:12:29 I scaled the patch back to the proposed scope 15:13:01 I saw that, seems CI hasn't broken - I think we're waiting for mgoddard to review next week. 15:13:15 it got an initial review from mgoddard - it should now reflect all requested changes 15:13:51 and I agree - now it needs mgoddard stamp of approval 15:14:53 of course open to any other comments/feedback 15:15:30 Ok, are there any other chunks we want to try to squeeze in for Xena? 15:15:49 I assume not if the door is closing 15:16:03 Ok, situation is clear. 15:16:16 Next topic 15:16:19 #topic Should we switch CentOS to upstream MariaDB as well? And should we try to sync version for all in Xena? And which one? https://review.opendev.org/c/openstack/kolla/+/810161 15:16:56 yoctozepto: from my perspective - if all other distros use upstream MariaDB - it would be a bit weird if we would keep CentOS using distro MariaDB rpms? 15:17:40 mnasiadka: same here 15:17:47 now 15:17:51 switch only on master 15:17:52 or everywhere 15:17:56 ? 15:18:12 Do we have any bugs opened/issues around MariaDB in stable branches? 15:18:23 yes, for debian 15:18:25 victoria and below 15:18:29 yes 15:18:38 ok, so we need to sync up Debian with Ubuntu, right? 15:18:47 yes 15:18:56 that's what I'm doing already 15:19:07 want to continue that down to affected ussuri 15:19:12 then do the other switches 15:19:19 let's plan the other ones 15:20:02 So we have no reason to switch CentOS anywhere else than master - so probably let's not do this - unless I'm mistaken. 15:20:12 definitely works for me 15:20:53 so #agreed ? 15:21:13 #agreed to switch CentOS and Debian to MariaDB upstream repo, and only backport Debian change to stable branches to sync up with Ubuntu and close existing bug. 15:21:24 great 15:21:25 If I haven't mixed up anything, I guess this is what we agreed to :) 15:21:25 there are two more questions there 15:21:30 yeah 15:21:36 the next q is 15:21:44 "And should we try to sync version for all in Xena?" 15:21:54 now we have debian 10.5 and others 10.3 15:22:00 we can't go down obviously 15:22:04 the current stable is 10.6 15:22:22 either do not sync or sync to 10.5 or 10.6 15:23:06 seems OSA reverted their jump to 10.6 a month ago 15:23:11 due to some upgrade problems 15:23:14 wallaby was 10.5 in Debian at least so what about sync all to 10.5 now and higher in Y? 15:23:28 https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/805822 15:24:05 I think since it's close to feature freeze and other fantastic events in X release - we should just bump everything to 10.5 15:24:51 makes sense to me 15:25:08 #agreed to sync MariaDB on all distros to 10.5 and pursue 10.6 in next cycle 15:25:28 any more questions for that topic? 15:25:32 mnasiadka: re osa - they rely on deb to upgrade so... you know ;-) 15:25:45 not from me 15:25:59 galera_repo_host: "downloads.mariadb.com" in their role defaults/ 15:26:15 ah, you meant something else 15:26:21 mnasiadka: I mean, they upgrade it with apt/dpkg 15:26:24 yeah 15:26:29 now I get it :) 15:26:32 :D 15:26:32 #topic Should we support Rocky in images? 15:26:52 I'm not sure there are clear benefits 15:26:58 what's wrong with stream images on rocky? 15:27:16 drivers go on host anyway 15:27:29 I don't know if I would feel safe in running that combination, but maybe it would work. 15:27:40 then run debian (-: 15:27:43 Wasn't Rocky EOL already? 15:27:52 OpenStack Rocky? Maybe :) 15:28:06 khekhe 15:29:55 it seems we are unsure on the decision to make 15:29:59 hmm 15:30:29 i think we should also have mgoddard for the discussion, and hrw seems to have no opinion ;-) 15:30:45 I think that it wwe should rather concentrate on Rocky Linux as host in k-a 15:30:52 if RL at all 15:31:05 then hrw sounds like me 15:31:25 note that for images you need also to convince infra to mirror RockyLinux repos 15:31:44 otherwise you eat lot of bandwidth on CI 15:31:55 True, I was not thinking of merging it this cycle, easy. 15:32:12 Let's just maybe add this as a PTG topic, maybe we'll have more opinions then. 15:32:15 ok, then we can postpone this discussion 15:32:17 yeah 15:32:17 it does not matter which cycle 15:32:25 For now DIB/bindep/etc don't support Rocky Linux 15:32:28 lot of work for nearly no gain 15:32:37 as you want source images only 15:32:53 hrw sounds like a harsher version of me 15:33:00 but true that 15:33:18 we have like 2-4 developers, 3 distros, 2 build types... 15:33:30 precisely 15:33:43 I would rather drop binary images than add another distro 15:33:54 I was thinking about that the other day ;-) 15:34:02 I know kevko might be unhappy about that 15:34:06 need to persuade him 15:34:11 Ok, let's postpone this discussion to the PTG. 15:34:18 I'm all in to drop everything and leave one distro and source only ;p 15:34:25 he always can apply for core membership. 15:34:29 I mean, for images 15:34:46 Rocky/Alma are good candidates for host distro 15:34:56 yeah, host is fine 15:35:03 as long as it has decent kernel 15:35:05 and runs docker 15:35:12 we should be good 15:35:22 and that's k-a job. where it will be mostly s/centos/any-rhel-clone/ 15:35:31 yup 15:35:45 That's not complicated, question what happens with CentOS Stream 9 and Yoga - will we be able to run Stream9 images on RHEL 8 compatible operating systems. 15:35:48 with small "shit, rocky" or "meh, centos stream" 15:36:02 mnasiadka: test with debian images only 15:36:26 what's the need for stream in images? ;d 15:36:43 I would ask who needs Stream at all :) 15:36:46 I would say we should change a bit our scope 15:36:50 mnasiadka: that too! lol 15:37:25 *if* we do binary images then we need Stream 15:37:41 yeah, I assume source-only for any further discussion 15:37:47 unless you want to drop rhel family 15:38:15 good topic for PTG then: Should we drop binary images? 15:38:17 I'll add a couple of related topics to the PTG etherpad, like dropping binary, using one distro as base, etc 15:38:21 mnasiadka: #action? 15:38:24 ok 15:38:37 #action mnasiadka to add topics to PTG around dropping binary, using one distro as base, etc 15:38:47 and let's close this pandora box 15:39:07 #topic Release tasks 15:39:26 hrw: I saw a change around adding binary packages to Debian in Xena? 15:39:44 correct 15:40:12 Ok, I guess it would be nice to get that merged as part of release tasks. 15:40:26 yep 15:40:27 also correct 15:40:34 good 15:40:42 #topic Xena cycle planning 15:40:56 I need to update the last of Ansible bump up changes 15:41:25 kevko seems to have updated some of his changes, I'll try to review tomorrow 15:41:49 yoctozepto: what about the Swift role remodeling? Can we stick to one change? Whiteboard mentions two :) 15:42:15 mnasiadka: not following that one atm 15:42:34 ok, so unlikely to get merged I guess 15:42:52 we can always try 15:42:59 it does not change the interface for the users 15:43:05 and shouldn't change experience so 15:43:18 for me it's a bugfix really 15:43:32 fixing the ugliness of the role (-: 15:43:38 yoctozepto: https://review.opendev.org/c/openstack/kolla-ansible/+/776662 seems to be waiting for you 15:43:50 oh boy 15:44:01 why am I not in "reviewers" then? :D 15:44:51 leaving it in a tab to review at some point 15:44:57 ok, good 15:45:15 seems we've got covered most of the Xena items in kolla/kolla-ansible 15:45:28 the rest is Kayobe, and we've got no representation from them today... 15:45:40 #topic Yoga PTG planning 15:45:55 There's an action on me to send email, populate etherpad with some ideas 15:46:08 #link https://etherpad.opendev.org/p/kolla-yoga-ptg 15:46:17 If anybody has any ideas for topics - please add them as well 15:46:21 is mnasiadka not a representation of kayobe? 15:46:48 I am :) 15:47:02 then mnasiadka is a little liar :-) 15:47:14 But I'll update the status of the changes after the meeting and chase reviewers using a different medium ;-) 15:47:23 ok! 15:48:26 ok then, last but not least 15:48:31 #topic Open discussion 15:48:36 12 minutes left in the meeting, incredible. 15:48:42 :O 15:48:53 perhaps we skipped some important topic and did not notice 15:49:15 do we have any potential candidates for core? 15:49:30 afaik kevko is trying 15:50:39 good 15:50:50 I think it's good he's trying, but he needs to at least double his review numbers... 15:51:11 also outside of his interests 15:51:17 correct 15:51:28 ++ 15:53:44 Ok, seems no other topics 15:54:37 Or maybe I'll start one - regarding Monasca CI - I don't want to be rude, but I feel we're trying to keep it alive, but one of the components is failing - one that calculates thresholds and alerts on those thresholds (like cpu busy). 15:54:53 just propose to drop the job 15:54:56 I'll accept that 15:55:02 I mean +2 15:55:38 I'm not into deprecating it now, but let's drop the job and ensure there's a bug raised claiming Monasca does not work in Y or even W (maybe even reflect that in Support Matrix) 15:55:43 Ok, I'll do that. 15:55:55 good thinking 15:55:57 ++ 15:55:59 +2 15:56:06 I'm also working on that 15:56:15 "ugly projects" 15:56:28 to document and hide behind a var barrier 15:56:39 will propose later this week or next one 15:56:48 ok 15:57:02 so let's end this fine meeting 15:57:05 #endmeeting kolla