15:04:43 <noonedeadpunk> #startmeeting openstack_ansible_meeting
15:04:43 <opendevmeet> Meeting started Tue Jan 27 15:04:43 2026 UTC and is due to finish in 60 minutes.  The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:04:43 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:04:43 <opendevmeet> The meeting name has been set to 'openstack_ansible_meeting'
15:04:49 <noonedeadpunk> #topic rollcall
15:04:51 <noonedeadpunk> o/
15:04:55 <noonedeadpunk> sorry for being late
15:05:39 <damiandabrowski> hi!
15:10:24 <noonedeadpunk> #topic office hours
15:11:14 <noonedeadpunk> ok, I think one of big questions I have now - is what objections/comments do we have to the vault patch?
15:11:17 <noonedeadpunk> #link https://review.opendev.org/c/openstack/ansible-role-pki/+/948881
15:11:51 <jrosser> o/ hello
15:15:02 <noonedeadpunk> I think it had most of requested changes
15:15:46 <noonedeadpunk> so it would be nice if we could do another round of review and see where we are
15:16:35 <jrosser> its still difficult to review (i mean not because its bad, theres just complexity and vault is not necessarily obvious)
15:16:55 <noonedeadpunk> yeah, that's true
15:17:01 <jrosser> but if you are to be using it internally then thats a +ve vs something thats speculative
15:17:28 <jrosser> a bit like how we are running mcapi here so are able to maintain all that stuff better than if we were not
15:17:59 <jrosser> so i guess thats what i;m interested in, if you have this rolled out and working or will do that once it merges
15:18:11 <noonedeadpunk> so we already have the vault, we just workaround a lot, just because it needs changes to pretty much everything...
15:18:32 <noonedeadpunk> but I think damiandabrowski tested the full set in sandbox
15:19:15 <noonedeadpunk> so yes, we intended to use it once it becomes available, basically with upgrade to 2026.1
15:19:26 <jrosser> if you're able to test it as it merges (if we get on with it) then there will be some good confidence its ok
15:20:01 <damiandabrowski> that's right I tested it on my AIO environment some time ago, if it makes it any easier I to review, I can set up another AIO and give you an access
15:20:13 <jrosser> heh -ETIME
15:20:24 <noonedeadpunk> I think we will convert our dev envs to it once it merges, yes
15:20:34 <jrosser> i think it's just good confidence to know that you're using it and fixes will get made once/if nececessary
15:21:01 <jrosser> tbh i think that is the best way to move forward
15:21:17 <jrosser> take another pass at the review as far as is possible without running it
15:21:18 <noonedeadpunk> I have some level on confidence in molcule :D
15:21:25 <jrosser> and let it go in
15:21:39 <jrosser> yeah true
15:22:35 <noonedeadpunk> it feels like the most imporant things are covered in tests
15:23:16 <jrosser> did we decide if to have a top level test scenario that deploys vault
15:23:54 <jrosser> or perhaps we can be happy that if the molecule tests for PKI have good coverage, we don't need to
15:24:23 <noonedeadpunk> I feel like we are having too many jobs right now
15:24:39 <damiandabrowski> https://review.opendev.org/c/openstack/openstack-ansible/+/948889/comment/bf25225a_76df5807/
15:24:48 <noonedeadpunk> I'd really like to look through them and validate if we need all existing ones
15:25:30 <noonedeadpunk> but maybe we can add one vault job for ubuntu just in case
15:25:32 <damiandabrowski> there was a proposal here to run only a single vault job in our baseline, and run more vault scnearios for PKI role only
15:25:48 <noonedeadpunk> yeah
15:26:11 <noonedeadpunk> but this was not updated ;)
15:26:13 <noonedeadpunk> #link https://review.opendev.org/c/openstack/openstack-ansible/+/948889/2/zuul.d/project-templates.yaml
15:29:56 <noonedeadpunk> hm, I'd expect more patches to be pushed actually for https://review.opendev.org/q/topic:%22openstack/openstack-ansible-sync/sync-files%22
15:30:27 <noonedeadpunk> it feels there's somewhere an issue in role selection logic
15:30:33 <damiandabrowski> yeah, I'll fix it. Didn't focus on this because https://review.opendev.org/c/openstack/openstack-ansible/+/948889/13 doesn't run any jobs
15:30:35 <noonedeadpunk> *target repo
15:30:44 <damiandabrowski> but now I wonder, why?
15:31:09 <noonedeadpunk> I think there was smth with dependencies?
15:31:21 <noonedeadpunk> like one was abandoned or smth?
15:31:41 <damiandabrowski> it complains that several changes "are needed" but I'm not sure what exactly does it mean
15:31:56 <damiandabrowski> hmm, you may be right...I'll make sure there's no dependency to abandoned patch
15:32:34 <noonedeadpunk> let's maybe cross the bridge when we come to it?
15:32:44 <noonedeadpunk> I think it will be easier to untangle when we land things
15:32:55 <damiandabrowski> okay, makes sense
15:33:58 <noonedeadpunk> there is one thing, which brakes upgrades:
15:34:04 <noonedeadpunk> #link https://review.opendev.org/c/openstack/openstack-ansible/+/972469
15:34:25 <noonedeadpunk> it would be actually nice to have that fix in, as 2025.2 is detected as 2025.1 tag
15:34:30 <noonedeadpunk> so venvs are not rebuilt
15:34:49 <noonedeadpunk> or well, they kinda are, but also their naming is not mathing the tag
15:35:32 <noonedeadpunk> and then pretty much we'd need to agree on what's needed for the 32.1.0 tag
15:37:40 <noonedeadpunk> there are actually plenty of stuff lying around all over repos...
15:37:57 <jrosser> yes theres a lot of outstanding patches to deal with
15:38:02 <jrosser> some of which i'm unsure about
15:39:01 <noonedeadpunk> feel free to bring them in for a chat once you see them?
15:39:48 <noonedeadpunk> or just -1 :D
15:39:56 <noonedeadpunk> (-2)
15:41:03 <jrosser> sure, i need to ask about more of them here really
15:43:37 <opendevreview> Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Move from setup.cfg to pyproject.toml  https://review.opendev.org/c/openstack/openstack-ansible/+/972481
15:44:19 <noonedeadpunk> I have not yet started looking at freezer/watcher roles
15:44:38 <noonedeadpunk> though, it was discovered that we have not properly re-retired monasca ones
15:44:53 <noonedeadpunk> #link https://review.opendev.org/c/openstack/project-config/+/974773
15:45:19 <noonedeadpunk> I mentally burried it, but on paper it was "revived"
15:46:02 <noonedeadpunk> I am trying to "fix" some freezer parts first, before moving to role part
15:46:36 <jrosser> were you going for scheduled cinder backups?
15:46:42 <noonedeadpunk> yup
15:46:48 <jrosser> excellent i would be interested in that
15:47:06 <jrosser> it is so bad from a user POV that cinder backup currently cant schedule
15:47:09 <noonedeadpunk> that's pretty much the goal we have
15:47:56 <noonedeadpunk> there's been quite some patches merged in last 2 weeks to have that said.
15:47:58 <noonedeadpunk> but yeah
15:48:41 <jrosser> kind of way off topic but did manila virtio-fs ever get to be real?
15:48:57 <noonedeadpunk> I think it should be in now?
15:49:06 <noonedeadpunk> though you'd need hugepages enabled I think
15:49:10 <jrosser> hmmm i totally took my eye off of that
15:49:13 <jrosser> ewwww
15:50:40 <noonedeadpunk> and also no hotplug
15:51:05 <noonedeadpunk> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/WEB3XJYWWSW672N4UAX3VFJSQEE6V7RA/
15:52:19 <damiandabrowski> when it comes to Manila, it looks like client-side support is still missing
15:52:20 <damiandabrowski> https://review.opendev.org/q/topic:%22bp/manila_shares_attachments_v3%22
15:52:42 <damiandabrowski> and it looks quite abandoned :/
15:53:06 <noonedeadpunk> well...
15:53:41 <noonedeadpunk> I'd say client side overall looks abandoned today
15:57:32 <noonedeadpunk> on this positive note....
15:57:35 <noonedeadpunk> #endmeeting