15:00:01 <noonedeadpunk> #startmeeting openstack_ansible_meeting 15:00:01 <opendevmeet> Meeting started Tue Aug 8 15:00:01 2023 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:01 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:01 <opendevmeet> The meeting name has been set to 'openstack_ansible_meeting' 15:00:07 <noonedeadpunk> #topic rollcall 15:00:09 <noonedeadpunk> o/ 15:00:25 <damiandabrowski> hi! 15:03:36 * noonedeadpunk have some troubles with loading gerrit 15:03:37 <jamesdenton> o/ 15:04:17 <damiandabrowski> it doesn't work for me as well 15:04:39 <noonedeadpunk> Not sure if I'm happy or sad now... 15:04:54 <damiandabrowski> (it loaded after some time and now it looks ok) 15:05:00 <mgariepy> it works ok for me. 15:05:59 <noonedeadpunk> yup, jsut loaded 15:06:06 <noonedeadpunk> thanks, mgariepy, you've fixed it:) 15:06:25 <noonedeadpunk> should we go through the bugs? We had a couple during the last week 15:06:35 <noonedeadpunk> most were created by andrewbonney though :D 15:06:41 <andrewbonney> Sorry! 15:06:45 <mgariepy> lol 15:06:45 <noonedeadpunk> #topic bug triage 15:06:53 <noonedeadpunk> #link https://bugs.launchpad.net/openstack-ansible/+bug/2030748 15:07:17 <noonedeadpunk> I've already replied there though. 15:08:10 <NeilHanlon> o/ late :) 15:08:33 <noonedeadpunk> To be frank, I kinda don't know how to fix that either with fully re-writing clustercheck to python with async (like with asyncio) or with ProxySQL 15:08:49 <noonedeadpunk> *except these 2 15:09:31 <noonedeadpunk> Also dunno if anyone else is bothered enough to invest there 15:09:53 <noonedeadpunk> I guess, you should be able to just filter out `mariadbcheck` completely? 15:11:01 <noonedeadpunk> So I had no plans investing into fix soonish 15:11:28 <noonedeadpunk> #link https://bugs.launchpad.net/openstack-ansible/+bug/2029253 15:11:34 <noonedeadpunk> this one is fair one... 15:12:16 <noonedeadpunk> though I didn't know that non-root will work, as we don't have `become` anywhere in our playbooks 15:12:59 <andrewbonney> not sure if we have some extra config that might be enabling that 15:13:20 <andrewbonney> I've started a potential patch but about to go away for a week so won't get chance to do more for a bit 15:13:30 <noonedeadpunk> aha, I see 15:13:59 <noonedeadpunk> Can't promise I will look into that either :( 15:14:34 <noonedeadpunk> But would be great to get this patched sooner then later to be frank 15:14:42 <noonedeadpunk> As new deployments are indeed a concern 15:14:59 <noonedeadpunk> #link https://bugs.launchpad.net/openstack-ansible/+bug/2028946 15:15:06 <noonedeadpunk> I think proposed patch is really good 15:15:26 <noonedeadpunk> Seems we've jsut somehow overlooked that for ubuntu :( 15:16:13 <noonedeadpunk> #link https://bugs.launchpad.net/openstack-ansible/+bug/2029486 15:16:47 <noonedeadpunk> I think, we can add step for marking _member_ an implied in upgrade steps 15:16:57 <noonedeadpunk> And we should really do a clean-up anyway 15:16:59 <NeilHanlon> noonedeadpunk: i have an idea w.r.t. mariadb clustercheck... will throw in a DNM change 15:17:02 <noonedeadpunk> I can take a look there... 15:17:22 <noonedeadpunk> NeilHanlon: oh, that would be extermely sweet 15:17:33 <noonedeadpunk> I have an env where I'll be able to test that 15:17:59 <NeilHanlon> in theory, SyslogIdentifier can be set on the unit to force a name.. i think it even works on templated units 15:21:50 <noonedeadpunk> NeilHanlon: it doesn't look like it does from what I see 15:22:04 <NeilHanlon> :\ welp.. off to systemd internals :) 15:22:39 <noonedeadpunk> Unless I'm doing smth wrong https://paste.openstack.org/show/b6VJZX5etkWNilqGam05/ 15:23:41 <noonedeadpunk> #link https://bugs.launchpad.net/openstack-ansible/+bug/2028809 15:23:52 <noonedeadpunk> We've already discussed this one couple of weeks ago 15:24:24 <noonedeadpunk> I did not have any reply/discussion with keystone team, despite it seems that Dave has picked up the bug 15:24:29 <noonedeadpunk> Or well, at least saw it 15:24:36 <NeilHanlon> ah.. found the context. looks like it can be fairly trivially done w/ a simple python change to redirect the syslog target 15:24:44 <NeilHanlon> #link https://github.com/systemd/systemd/issues/8687 15:24:50 <NeilHanlon> #link https://github.com/systemd/systemd/issues/13566 15:26:24 <noonedeadpunk> Well, if a service was calling python daemon - there won't be such problem, as we could set access=no then ... 15:26:39 <NeilHanlon> right, clustercheck is bash.. nvm 15:26:52 <noonedeadpunk> was worth trying though... 15:28:03 <noonedeadpunk> So, for the 2028809 I still think it's worth to change the default to use scrypt rather then bcrypt which is default in keystone 15:28:25 <noonedeadpunk> As default limitation to passwords with length 54 is not fun in 2023 15:28:37 <NeilHanlon> yeah agreed 15:28:37 <noonedeadpunk> *to password length 15:29:19 <noonedeadpunk> given old password will resume working (which are not longer then 54 ofc) - we at least prevent everyone from further frustration 15:29:46 <noonedeadpunk> And we likely should land the fix, as it's blocking backports to 2023.1 which we have quite a few by now 15:30:13 <noonedeadpunk> #link https://bugs.launchpad.net/openstack-ansible/+bug/2029209 15:30:24 <noonedeadpunk> damiandabrowski: will you be able to have a look at this one? 15:31:05 <noonedeadpunk> as it sounds quite fair 15:31:33 <damiandabrowski> i think the fix is already proposed: https://review.opendev.org/c/openstack/openstack-ansible/+/889023 15:31:36 <noonedeadpunk> aha 15:31:51 <damiandabrowski> but since stable/2023.1 branch is blocked we couldn't merge it :D 15:32:00 <noonedeadpunk> ok, gotcha 15:32:03 <noonedeadpunk> andrewbonney: ^ 15:32:20 <andrewbonney> Ta, hadn't spotted that 15:33:04 <noonedeadpunk> #topic office hours 15:33:33 <noonedeadpunk> We have couple of topics that are ongoing. 15:33:54 <noonedeadpunk> First - simple one is https://review.opendev.org/q/topic:osa%252Fdb_port 15:34:15 <noonedeadpunk> I was thinking to also backport that at least to 2023.1 or maybe even zed, as it's pretty much trivial 15:34:19 <noonedeadpunk> But kinda nasty 15:37:11 <damiandabrowski> +1 15:37:27 <noonedeadpunk> next are linters for core-2.15 15:37:29 <noonedeadpunk> #link https://review.opendev.org/q/topic:osa%252Fcore-2.15 15:37:56 <noonedeadpunk> These get to merge conflicts quick enough 15:38:47 <damiandabrowski> i'll review them during the evening 15:39:11 <andrewbonney> noonedeadpunk: unless I missed it I think there was one more bug too: https://bugs.launchpad.net/openstack-ansible/+bug/2028944 15:39:34 <noonedeadpunk> Ah, sorry, I have missed this one indeed 15:40:33 <noonedeadpunk> yes, this is actually a good one 15:41:03 <damiandabrowski> yeah, nice catch andrewbonney 15:41:10 <noonedeadpunk> I kinda thought that haproxy will fail due to conflict to bind 2 fronts the same way to be frank 15:42:14 <noonedeadpunk> but yeah we should fix this 15:43:24 <andrewbonney> Considering the amount of change to haproxy setup it worked really well in general 15:44:30 <noonedeadpunk> ok, that's nice to hear:) 15:44:56 <noonedeadpunk> I bet horizon wasn't spotted as we jsut don't use it hehe 15:45:24 <noonedeadpunk> Ok, last but not least topic is quorum queues 15:45:26 <noonedeadpunk> #link Saying that, I'm not sure - have we covered vars scope good enough in release notes? That stuff defiend for haproxy_all should be moved else 15:45:32 <noonedeadpunk> #link https://review.opendev.org/q/topic:osa%252Fquorum_queues 15:46:27 <noonedeadpunk> This is smth that would be great to get in for this release 15:50:19 <opendevreview> Merged openstack/openstack-ansible-os_nova master: Deprecate nova_ram_weight_multiplier https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/888177 15:52:05 <damiandabrowski> not sure if I'll be able to review it this week as i need to learn more about quorum queues before, but I added it to TODO :D 15:52:49 <noonedeadpunk> Sure, we have some time for that 15:53:16 <damiandabrowski> also, feel free to assign me to #2028944 if you want 15:53:24 <noonedeadpunk> And yeah, let's try to unblock 2023.1 soonish, along with fixed bugs we likely can do 27.1.0 15:54:42 <noonedeadpunk> damiandabrowski: can you kindly assign yourself? As I don't have proper "a" on my keyboard to search :D 15:55:15 <damiandabrowski> ą - you can copy that xD 15:55:18 <damiandabrowski> just kidding, done 15:55:24 <noonedeadpunk> lol 15:55:41 <noonedeadpunk> yeah, will paste it somewhere safe :D 15:56:15 <noonedeadpunk> I'm also working on some article to superuser regarding scaling clusters with OSA 15:56:35 <noonedeadpunk> now for rabbit/galera clusters 15:56:57 <noonedeadpunk> Wanna move smth from there to our docs as well 15:57:21 <noonedeadpunk> (once it will be ready) 15:58:21 <noonedeadpunk> #endmeeting