15:00:28 <noonedeadpunk> #startmeeting openstack_ansible_meeting
15:00:28 <opendevmeet> Meeting started Tue Jan 25 15:00:28 2022 UTC and is due to finish in 60 minutes.  The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:28 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:28 <opendevmeet> The meeting name has been set to 'openstack_ansible_meeting'
15:00:37 <noonedeadpunk> #topic rollcall
15:00:41 <noonedeadpunk> \o/
15:00:53 <damiandabrowski[m]> hey!
15:02:10 <jrosser> hello
15:02:13 <spatel> \o/
15:04:13 <noonedeadpunk> #topic bug triage
15:04:22 <noonedeadpunk> we have several new bugs lately
15:04:36 <noonedeadpunk> they seem pretty simple though
15:05:31 <noonedeadpunk> admin16: sorry for stupid question, but aren't you admin0 for co-incedence ?:)
15:06:06 <noonedeadpunk> as regarding https://bugs.launchpad.net/openstack-ansible/+bug/1955676 would be great to have a descision if it's go or no go
15:07:05 <noonedeadpunk> Other then that, there's other interesting bug
15:07:08 <noonedeadpunk> #link https://bugs.launchpad.net/openstack-ansible/+bug/1958645
15:07:40 <noonedeadpunk> I'm not sure I see why comma should be added there?
15:08:03 <noonedeadpunk> as we put it inside horizon_local_settings, which is regular python code
15:09:05 <spatel> seems typo
15:09:08 <noonedeadpunk> I'm not sure it does correct thing this way though https://opendev.org/openstack/openstack-ansible-os_horizon/src/branch/master/templates/horizon_local_settings.py.j2#L92-L93
15:11:19 <jrosser> huh https://opendev.org/openstack/horizon/src/branch/master/doc/source/configuration/settings.rst#openstack_keystone_domain_choices
15:11:32 <jrosser> what is that all about
15:11:34 <noonedeadpunk> I haven't used horizon for quite a while, but are choices even a required thing?
15:12:15 <noonedeadpunk> I mean - that could be even avoided by default, or...?
15:12:34 <jrosser> same thing appears to be the case for AVAILABLE_THEMES and another few things in there
15:12:42 <noonedeadpunk> `option can be used`
15:13:05 <noonedeadpunk> And if it's not defined - would all domains appear by default....
15:13:50 <noonedeadpunk> I guess not, as default setting would prevent it I believe...
15:13:58 <jrosser> well - seems our code does not match the horizon docs either way
15:14:57 <noonedeadpunk> well that fully worked for me on T though...
15:15:05 <noonedeadpunk> But yes, agree, we need to change that
15:15:09 <jrosser> https://opendev.org/openstack/kolla-ansible/src/branch/master/ansible/roles/horizon/templates/local_settings.j2#L102
15:16:05 <noonedeadpunk> yeah, fair
15:17:05 <noonedeadpunk> ANother trivial bug
15:17:08 <noonedeadpunk> #link https://bugs.launchpad.net/openstack-ansible/+bug/1958434
15:17:38 <noonedeadpunk> I think that's also fair solution suggested?
15:19:07 <spatel> how does CI pass then? because we don't have that role for CI job?
15:19:43 <jrosser> probably becasue we do a metal deploy and it picks it up off the host
15:19:57 <noonedeadpunk> Yeah
15:19:58 <jrosser> theres quite a few lxc<>metal bugs occur that way
15:22:12 <opendevreview> Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_horizon master: Fix default multidomain choises  https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/826310
15:22:17 <noonedeadpunk> #topic office hours
15:23:12 <noonedeadpunk> Oh, backport candidate label has just been merged
15:24:07 <opendevreview> Jonathan Rosser proposed openstack/openstack-ansible-os_aodh master: Ensure libxml2 is installed on debian systems  https://review.opendev.org/c/openstack/openstack-ansible-os_aodh/+/826311
15:24:56 <jrosser> do we abandon these https://review.opendev.org/q/topic:reno-ocata ?
15:25:07 <noonedeadpunk> yep
15:25:37 <noonedeadpunk> it was really nasty side effect of EOLing integrated repo...
15:25:59 <jrosser> andrewbonney: we need this for centos-8 removal https://review.opendev.org/c/openstack/openstack-ansible/+/826135
15:25:59 <noonedeadpunk> So back in days integrated and role repos were defined in single file in releases
15:26:40 <noonedeadpunk> So roles were eoled, and I decided to eol integrated as follow-up, once a-r-r will be switched to eol tag as well
15:27:07 <noonedeadpunk> but when adjusting list, CD decided to create branch on every repo, and then delete it....
15:27:12 <noonedeadpunk> spawning tons of rubbish
15:28:24 <noonedeadpunk> btw returning back to horizon, there's one patch that is interesting I guess https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/820928
15:30:37 <noonedeadpunk> I tested it in aio both for distro and source with some extra panel (like heat?). Source worked nicely, but for distro iirc panel wasn't added for $reason. It actually wasn't added even with current state of role, so might be issue with panel itself that is packaged...
15:30:47 <noonedeadpunk> but not sure
15:31:47 <noonedeadpunk> And I wasn't able to take a look on CI stuff and how to re-use that efficiently during previous week
15:32:04 <jrosser> no worries
15:32:14 <jrosser> i have been out for a few days too
15:32:24 <jrosser> just managed to pick up the keypairs stuff again today
15:32:30 <noonedeadpunk> Oh, would be awesome to merge https://review.opendev.org/q/topic:integrated_lintersbtw
15:32:53 <jrosser> andrewbonney: looks like more for you ^
15:32:58 <noonedeadpunk> Yeah, I saw you pushed changes, will hopefully review them tomorrow morning
15:32:58 <andrewbonney> :)
15:33:22 <noonedeadpunk> as well as damiandabrowski[m] to tempest
15:33:26 <jrosser> so on the keypairs stuff - i *think* it will work for nova on a new deployment, though i've only tested AIO
15:33:54 <noonedeadpunk> will it for upgraded one ?:)
15:33:58 <jrosser> there is code to undo all the old config, but i need to add handling for existing sshd config which may conflict with what this new role wants to do
15:34:13 <jrosser> i did add stuff to remove the old ssk keys, so it will probably work
15:34:29 <jrosser> but similar to SSL stuff you probably need to have run the play against all the hosts
15:36:27 <opendevreview> Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Define variables for tempest plugins  https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/826121
15:37:47 <noonedeadpunk> I think it's quite fair to distribute across all computes.
15:38:06 <noonedeadpunk> Thankfully this should be done one time and could be run with limit afterwards?
15:38:36 <jrosser> yes, i think with this method there is no coupling between the compute hosts at all
15:38:56 <jrosser> where are we with the centos-8 removal
15:39:41 <opendevreview> Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Define variables for tempest plugins  https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/826121
15:39:47 <jrosser> xena is broken on tempest vars?
15:39:58 <jrosser> vars/requirements
15:40:01 <noonedeadpunk> X is broken with neutron plugin version
15:40:22 <noonedeadpunk> Eventually what I realized is that starting from X tempest is not in u-c anymore
15:41:22 <noonedeadpunk> So I got fed with that and decided that it's worth fixing plugins versions as well
15:43:46 <noonedeadpunk> With removal as far as I got, chain looks like - we merge https://review.opendev.org/c/openstack/openstack-ansible/+/826135, then we should able to merge https://review.opendev.org/c/openstack/openstack-ansible/+/824567 and https://review.opendev.org/c/openstack/openstack-ansible/+/824570 likely go last
15:47:11 <jrosser> i need to look again at the centos-9 patches
15:47:31 <opendevreview> Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest stable/xena: Define variables for tempest plugins  https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/826060
15:48:17 <opendevreview> Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Convert xinetd clustercheck to systemd socket service  https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/824042
15:48:37 <jrosser> ^ this is important if the proxysql stuff might take a while
15:51:02 <noonedeadpunk> It's in merge conflict now
15:51:22 <jrosser> ok no worries
15:51:31 <jrosser> i think i need to test it locally on a fresh AIO
15:51:35 <jrosser> something wierd is not working
15:52:03 <noonedeadpunk> Main issue with proxysql, that I'm a bit struggling with, is that it requires users to be created for itself as well https://proxysql.com/documentation/users-configuration/ and all this layering of config that is applied...
15:52:57 <noonedeadpunk> So it's a bit fishy and would require big changes and some migration...
15:53:12 <noonedeadpunk> But honestly time is main thing that I'm struglling for real :)
15:59:17 <noonedeadpunk> And now I can finally work on updating dashboard to include BC
16:01:37 <spatel> noonedeadpunk I am working on W - X upgrade in my lab
16:01:50 <noonedeadpunk> #endmeeting