15:00:45 <noonedeadpunk> #startmeeting openstack_ansible_meeting 15:00:45 <opendevmeet> Meeting started Tue Jul 15 15:00:45 2025 UTC and is due to finish in 60 minutes. The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:45 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:45 <opendevmeet> The meeting name has been set to 'openstack_ansible_meeting' 15:00:50 <noonedeadpunk> #topic rollcall 15:00:55 <noonedeadpunk> o/ hey there 15:01:15 <DavidGomez> o/ 15:04:29 <damiandabrowski> hi! 15:04:34 <noonedeadpunk> #topic office hours 15:05:12 <noonedeadpunk> so, I think the first thing I wanted to raise is review state overall - we have quite some things waiting for approval, including some bug fixes 15:05:44 <damiandabrowski> ack, I'll have a look 15:06:47 <jrosser> o/ hello 15:07:27 <jrosser> we really do need more activity on reviews all the time 15:07:49 <jrosser> as noonedeadpunk and i seem to make a lot of patches and do a lot of reviews we can't merge each others stuff alone 15:08:05 <noonedeadpunk> right 15:08:27 <noonedeadpunk> unless we change the project rules 15:08:36 <noonedeadpunk> which would be better to avoid 15:08:56 <noonedeadpunk> but I'm getting tempted from time to time to be frank 15:09:06 <jrosser> andrew will be back in ~1 week 15:09:26 <noonedeadpunk> ┌(° ͜ʖ͡°)┘ 15:10:00 <noonedeadpunk> ok, then I'll manage my temptation :) 15:10:26 <noonedeadpunk> I see some work has started on adding Debian 13 job? 15:10:46 <noonedeadpunk> though we have an issue in hardening module 15:10:54 <noonedeadpunk> and python 13 compatability 15:12:25 <noonedeadpunk> 3.14=3 ofc, sorry 15:12:32 <noonedeadpunk> * 3.13 15:13:29 <jrosser> i have it in a vm here 15:13:42 <jrosser> there is a super strange pip error in the patch i pushed that i dont understqand 15:13:47 <jrosser> and i dont get locally 15:15:01 <noonedeadpunk> also, it seems that C10S at least nodesets are around 15:15:13 <noonedeadpunk> so probably we should check on adding CI for it as well 15:15:28 <noonedeadpunk> at least from what I see in kolla recent patches: https://review.opendev.org/c/openstack/kolla/+/950392/78/.zuul.d/centos.yaml 15:15:44 <noonedeadpunk> (I was searching for rocky10 though in gerrit) 15:16:19 <noonedeadpunk> I think that LXC was built for EPEL testing at least 15:16:55 <noonedeadpunk> https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2025-87f14463e7 15:17:35 <noonedeadpunk> interestingly, it seems that EL is gonna have more modern LXC then Noble does 15:17:47 <noonedeadpunk> jrosser: what lxc is shipped in debian - did you checked? 15:18:02 <noonedeadpunk> or can you if have VM handy :) 15:18:09 <jrosser> yeah one moment 15:18:15 <jrosser> carry on whist i look :) 15:18:24 <noonedeadpunk> sure 15:19:23 <noonedeadpunk> Do we have any updates or smth to discuss on PKI topic? 15:19:26 <noonedeadpunk> damiandabrowski: ? 15:20:06 <damiandabrowski> I'm working on the things we agreed on and testing them on my AIO 15:20:35 <jrosser> i have left the os_glance change as an example for where we might start the vault stuff 15:20:37 <damiandabrowski> so nothing to discuss, until I push updates to my patches :D 15:21:16 <damiandabrowski> https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/954269 15:21:19 <damiandabrowski> yeah I saw, thanks a lot 15:22:07 <noonedeadpunk> ok, cool, then no obvious blockers so far at least :) 15:22:41 <noonedeadpunk> and this is pki-side patch, I assume 15:23:40 <jrosser> debian 13 lxc `Installed: 1:6.0.4-4+b1` 15:23:42 <noonedeadpunk> btw, just today I was thinking if we should add some better documentation around the role with real-life examples, like generating certs for libvirt/nova, if to use it as a standalone role 15:23:58 <noonedeadpunk> ok, so it's the same with what C10S and EL10 gonna have 15:24:16 <noonedeadpunk> just noble seems to be still on 5 15:24:28 <noonedeadpunk> so I was a bit curious how badly 6 gonna break 15:24:43 <jrosser> as far as lxc goes it did just work (+/- a few bugfixes i pushed) on trixie 15:24:55 <jrosser> but those are our bugs in ansible code 15:25:24 <noonedeadpunk> right 15:26:30 <noonedeadpunk> Should we be mergning https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/954976 now? 15:27:16 <noonedeadpunk> As it's more of - get all patches out vs start merging things regardless 15:29:32 <noonedeadpunk> don't have anything against any approach, just checking how you wanna handle that 15:29:38 <jrosser> it might not hurt, we can merge that even if trixie is experimental/broken 15:30:22 <NeilHanlon> bleh, sorry.. distracted today 15:30:40 <noonedeadpunk> I kinda wonder, if we might actually do a list of OS as a variable.... 15:31:05 <noonedeadpunk> so that to be able to override in AIO/CI, but not allow in actual deployments 15:31:27 <NeilHanlon> regarding systemd-extras for el10: yes we should request it if not there. I'd asked about el9 specifically to the group and sorta assumed it was there for 10, too.. I can file that request quickly on my laptop 15:32:13 <NeilHanlon> regarding how the includepkgs: systemd-networkd from EPEL worked/works, it's because the package is "provided" by the systemd-extras, so it's addressable in it's own right. similar to how normally the `systemd` package provides all it's subcomponents 15:32:24 <NeilHanlon> also hi :) o/ 15:32:33 <noonedeadpunk> aha, ok, makes sense then 15:33:03 <noonedeadpunk> but I kinda wonder from standpoint of ecosystem... Why would epel need having it, if it's in SIG? 15:33:12 <noonedeadpunk> Or SIGs are only CentOS oriented? 15:33:17 <noonedeadpunk> and EPEL is also EL? 15:33:26 <NeilHanlon> well the Hyperscale SIG itself is only focused on centos stream 15:33:51 <NeilHanlon> so they wouldn't "allow" a build for systemd atop RHEL (and friends) in that SIG 15:33:58 <noonedeadpunk> I just kinda a little bit confused about versions provided by these 15:34:01 <NeilHanlon> so it'd have to be another SIG like cloud or nfv 15:34:08 <noonedeadpunk> aha, ok 15:34:08 <NeilHanlon> or, a rocky sig. or epel 15:34:30 <NeilHanlon> cause epel is against RHEL build roots, not Stream ones 15:34:36 <noonedeadpunk> so like epel shouldn't say smth like - go and use hyperscale sig now? 15:35:00 <NeilHanlon> no cause EPEL is maintaining compatibility with RHEL, not Stream 15:35:09 <noonedeadpunk> ok, then I think having a request to build networkd for el10 would be really nice 15:35:18 <NeilHanlon> 👍i'm gonna put in the request 15:35:29 <noonedeadpunk> amazing, thanks! 15:36:09 <NeilHanlon> looks like there's already one in: https://bugzilla.redhat.com/show_bug.cgi?id=2303892 15:36:15 <NeilHanlon> i'll poke rsc about it 15:36:53 <noonedeadpunk> yeah, it didn't go too far :( 15:38:23 <noonedeadpunk> lxc in it's turn should be around anytime I guess 15:39:25 <noonedeadpunk> mariadb has builders for both rocky and c10s, but they did not build packages yet... 15:39:44 <noonedeadpunk> and I wouldn't exepct any until next minor version is out 15:40:03 <NeilHanlon> yep i need to test the lxc packages and give them some karma so they will move thru faster 15:41:30 <noonedeadpunk> so next mariadb release is planned on July 24th if my eyes don't fail me 15:42:43 <NeilHanlon> i believe ya 15:43:22 <noonedeadpunk> so I think once we solve CI/LXC/MariaDB we can tell about EL10 support also on Epoxy 15:43:32 <noonedeadpunk> but will see :) 15:43:36 <noonedeadpunk> and systemd ofc 15:43:42 <noonedeadpunk> so 4 things 15:44:07 <noonedeadpunk> I will try to poke at CI sometime this week 15:45:30 <NeilHanlon> sweet :) 15:45:36 <NeilHanlon> i'm gonna work on ceph this week... promise 15:45:47 <noonedeadpunk> even lxc is probably not a blocker, as we getting it from your corp 15:45:50 <noonedeadpunk> *copr 15:46:44 <noonedeadpunk> ok, great then :) 15:47:06 <noonedeadpunk> there's also one topic about magnum and capi drivers which was raised yesterday 15:47:39 <noonedeadpunk> as there's really good chance to get compatability with azimuth driver quite easy 15:48:24 <noonedeadpunk> though some minor changes, except capo>=0.12, to collection which produces control cluster would be needed 15:49:27 <noonedeadpunk> will try to check on that during these weekeends 15:49:42 <noonedeadpunk> and then - we can move out capi parts from ops repo 15:49:57 <noonedeadpunk> as I think magnum was gonna drop heat driver this cycle 15:51:10 <noonedeadpunk> so we should be right in time :) 15:53:46 <NeilHanlon> wonderful :D 15:54:05 <noonedeadpunk> anything else we wanna discuss? 15:54:31 <NeilHanlon> nothing from me i don't think 15:56:18 <jrosser> no, i dont think so 15:56:33 <noonedeadpunk> ok, then thanks everyone for taking time and for your contributions :) 15:56:36 <noonedeadpunk> #endmeeting