15:01:27 <noonedeadpunk> #startmeeting openstack_ansible_meeting
15:01:27 <opendevmeet> Meeting started Tue May 26 15:01:27 2026 UTC and is due to finish in 60 minutes.  The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:27 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:27 <opendevmeet> The meeting name has been set to 'openstack_ansible_meeting'
15:01:33 <noonedeadpunk> #topic rollcall
15:01:36 <noonedeadpunk> o/ hey there
15:02:16 <anfimovir> hello
15:03:06 <damiandabrowski> hi!
15:03:10 <chubinidzedr> o/
15:07:28 <jrosser> o/ hello
15:07:28 <noonedeadpunk> # topic office hours
15:07:37 <noonedeadpunk> sorry, got a bit distracted
15:07:46 <noonedeadpunk> So, we had a good progreess in multiple areas
15:08:25 <noonedeadpunk> except of manila and Zun, we have landed all hashi vault patches
15:08:30 <noonedeadpunk> #link https://review.opendev.org/q/topic:%22osa_hashi_vault%22+status:open
15:09:02 <damiandabrowski> I'm currently trying to fix CI for manila, potentially other roles can be affected as well.
15:09:03 <noonedeadpunk> manila reamins with broken CI for no known reason
15:09:17 <noonedeadpunk> oh
15:09:21 <damiandabrowski> https://zuul.opendev.org/t/openstack/build/66f6e1849f4c4e57bd809e715b86329e
15:09:48 <noonedeadpunk> Ok, so that I know what is
15:09:58 <damiandabrowski> there is a conflict between ceph apt repo deployed by ceph-ansible(one-line format) and the one deployed by ceph_client(deb822) format
15:10:00 <noonedeadpunk> there should have been a patch proposed to ceph_client role
15:10:15 <noonedeadpunk> https://review.opendev.org/c/openstack/openstack-ansible-ceph_client/+/989497
15:10:25 <noonedeadpunk> it should have been fixing it for general client...
15:10:47 <noonedeadpunk> but other then that -
15:10:50 <noonedeadpunk> #link https://github.com/ceph/ceph-ansible/pull/7701
15:10:54 <noonedeadpunk> this is the proper fix
15:11:47 <noonedeadpunk> I talked with folks who have an access to ceph-ansible about reviving CI for it
15:12:09 <noonedeadpunk> and I wanna step in to adopt molecule testing there instead of existing broken CI
15:12:20 <noonedeadpunk> as Ceph has removed resource quota from ceph-ansible for CI
15:12:50 <jrosser> i think that the site.yml is not allowing tentacle yet
15:12:57 <damiandabrowski> yeah, it seems that the above patch works for cinder, but not for manila. I don't know why and that's where I am now :D
15:13:14 <noonedeadpunk> there's also https://github.com/ceph/ceph-ansible/pull/7692
15:13:16 <jrosser> OSA uses only the roles so there is also patches needed to ceph-ansible site.yml
15:14:43 <noonedeadpunk> I will try to find some after-hours time to work on CI
15:14:53 <noonedeadpunk> and then I think it would be fine to merge these patches
15:17:16 <noonedeadpunk> magnum capi patches are mainly landed as well
15:17:18 <noonedeadpunk> #link https://review.opendev.org/q/topic:%22osa/k8s-mcapi%22+status:open
15:17:32 <noonedeadpunk> there is one nit about debian support and backport serpo
15:17:46 <noonedeadpunk> I will try to come up with patch for that asap
15:18:02 <noonedeadpunk> and then I also realized that azimuth driver actually does not do any testing
15:18:22 <noonedeadpunk> so the only thing we're testing, that play is passing, not that k8s cluster can be created with magnum
15:18:55 <noonedeadpunk> But I would suggest to proceed with branching is such state
15:19:02 <jrosser> i would hope we can recycle eventually the sonobouy code for that?
15:19:16 <noonedeadpunk> yes, I am pretty sure we  should be
15:19:33 <noonedeadpunk> it's just a different image we need to pull and different tag to provide to cluster template
15:19:47 <noonedeadpunk> I mentioned all that in docs
15:19:53 <noonedeadpunk> #link https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/989853
15:22:56 <noonedeadpunk> ubuntu 26.04 support also had some progress
15:23:00 <noonedeadpunk> #link https://review.opendev.org/q/topic:%22osa/resolute%22+status:open
15:23:16 <noonedeadpunk> thye nice part - metal job is passing
15:23:18 <noonedeadpunk> #link https://zuul.opendev.org/t/openstack/build/c2c2bcce98f0481f872190db4fdbace0
15:23:29 <noonedeadpunk> but LXC is failing on rally
15:23:36 <noonedeadpunk> I was able to reproduce that actually
15:23:54 <noonedeadpunk> wonder though why I haven't catched that earlier
15:25:38 <noonedeadpunk> one potentially controversial thing is https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/989979
15:26:05 <noonedeadpunk> which aims to filter out upper-constraints on repo host preparation, and drop global-requirements from there
15:26:58 <noonedeadpunk> and it seems to work, as otherwise Nova would fail to install as well
15:28:15 <noonedeadpunk> I did not have time to work on observability collection and help out with that
15:28:30 <noonedeadpunk> I'd guess we should branch first at this point anyway...
15:28:46 <noonedeadpunk> Talking about branching, we need to do that by end of the week regardless
15:28:59 <noonedeadpunk> so we have just couple of days to merge whatever needed kinda
15:29:11 <noonedeadpunk> as we're not gonna have much time for backports either
15:29:59 <noonedeadpunk> and I didn't work on ironic-inspector removal...
15:30:51 <noonedeadpunk> but I'd say we're in quite a decent shape now, and it should be good enough release
15:40:01 <noonedeadpunk> do we have anything else to cover?
15:42:34 <damiandabrowski> not from my side ;)
15:49:55 <noonedeadpunk> #endmeeting