| NeilHanlon | https://bugzilla.redhat.com/show_bug.cgi?id=2369112 | 02:44 |
|---|---|---|
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-lxc_container_create master: Remove duplicated default configuration parameters https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/951242 | 08:48 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-lxc_hosts master: Introduce common conf profile extension https://review.opendev.org/c/openstack/openstack-ansible-lxc_hosts/+/951257 | 09:39 |
| mnasiadka | NeilHanlon: I think now we're going to chase a lot of packages to be added to EPEL - see https://bugzilla.redhat.com/show_bug.cgi?id=2368765 ;-) | 09:50 |
| noonedeadpunk | it feels only the beginning, indeed | 10:04 |
| noonedeadpunk | mnasiadka: btw, I am right that EL10 docker containers also are failing to start properly on our infra? do you happen to know that? | 10:50 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_networkd master: Add EL10 support to the role https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/951137 | 10:51 |
| jrosser | noonedeadpunk: we could try a nested-virt node labels (which i think are coincidentally also the ones with the later generation instruction set) | 10:55 |
| mnasiadka | noonedeadpunk: if it’s x86 v3 stuff - just ise nested-virt node labels | 10:56 |
| mnasiadka | jrosser was faster ;) | 10:56 |
| noonedeadpunk | amazing, thanks folks! | 10:56 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Switch CAPI jobs to noble https://review.opendev.org/c/openstack/openstack-ansible-ops/+/951265 | 11:08 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_magnum master: Remove EL9 and Ubuntu Jammy from CI https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/951266 | 11:12 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add CentOS Stream 10 to molecule tests https://review.opendev.org/c/openstack/openstack-ansible/+/951132 | 11:24 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add CentOS Stream 10 to molecule tests https://review.opendev.org/c/openstack/openstack-ansible/+/951132 | 11:30 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Bump Ansible version to 11 for ELK https://review.opendev.org/c/openstack/openstack-ansible-ops/+/951267 | 11:35 |
| noonedeadpunk | Yeah, adding new Ubuntu or Debian version is sooooo trivial comparing to new EL version... | 11:37 |
| noonedeadpunk | mnasiadka: have you already implemented modular libvirt support in kolla? | 11:38 |
| mnasiadka | nope, not yet | 11:38 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add CentOS Stream 10 to molecule tests https://review.opendev.org/c/openstack/openstack-ansible/+/951132 | 11:42 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Remove Rocky 9 molecule testing https://review.opendev.org/c/openstack/openstack-ansible/+/951268 | 11:42 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Do not enable ebtables on modern kernels https://review.opendev.org/c/openstack/openstack-ansible/+/951160 | 11:42 |
| noonedeadpunk | nice, https://zuul.opendev.org/t/openstack/build/c6464f984a92453a91fcd7b573dcfcae is passing now | 11:43 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/2025.1: Bump SHAs for 2025.1 (Epoxy) RC2 https://review.opendev.org/c/openstack/openstack-ansible/+/951269 | 11:52 |
| noonedeadpunk | Let's land bumps, so I could propose releases at the same time :) https://review.opendev.org/q/topic:%22bump_osa%22+status:open | 11:53 |
| noonedeadpunk | unless, we wanna backport lxc changes to it... | 11:56 |
| jrosser | when is our release deadline? | 12:13 |
| noonedeadpunk | June 6 or smth | 13:02 |
| noonedeadpunk | so next week | 13:02 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Bump collection versions on master https://review.opendev.org/c/openstack/openstack-ansible/+/951287 | 13:33 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Bump global requirements https://review.opendev.org/c/openstack/openstack-ansible/+/951288 | 13:33 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_networkd master: Add EL10 support to the role https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/951137 | 14:10 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-hardening master: Add support for EL10 https://review.opendev.org/c/openstack/ansible-hardening/+/951296 | 14:25 |
| noonedeadpunk | So setup-hosts should be done now | 14:26 |
| noonedeadpunk | except missing LXC package in EPEL :) | 14:26 |
| jrosser | how far does a metal aio get? | 14:27 |
| noonedeadpunk | to gluster... | 14:27 |
| jrosser | aaaah ok | 14:27 |
| noonedeadpunk | adding in Neils repo now | 14:27 |
| noonedeadpunk | and tbh I'd expect once this is done, it should go all way to nova... | 14:28 |
| noonedeadpunk | but will see | 14:28 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Bump Ansible version to 11 for ELK https://review.opendev.org/c/openstack/openstack-ansible-ops/+/951267 | 14:32 |
| noonedeadpunk | capi test look bad. seems to be requiring specific privsep version or smth like that | 14:33 |
| noonedeadpunk | but fails not on magnum. but on coe template creation: https://zuul.opendev.org/t/openstack/build/e9958964200a4929a494c189f6a95d5d | 14:34 |
| noonedeadpunk | and heat magnum jobs also look quite bad.... | 14:36 |
| noonedeadpunk | anyway, not focusing on this now | 14:36 |
| jrosser | no, for later | 14:36 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_ironic master: Add apparmor rules for ironic inspector https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/951003 | 14:38 |
| jrosser | i wonder if the latest oslo.privsep is too new for the mcapi version we run, thats been pinned back now for a very long time | 14:53 |
| jrosser | i spend some time looking at how to get the rust stuff sorted out for the newer versions, which seems do-able | 14:54 |
| noonedeadpunk | `The ''oslo.privsep>=2.7.0'' distribution was not found and is required by magnum-cluster-api` | 14:55 |
| noonedeadpunk | so it more like... if magnum dropped oslo.privsep from it's requirements? | 14:55 |
| noonedeadpunk | I am frankly more concerned about absent modern k8s in the driver | 14:55 |
| noonedeadpunk | and that things never get merged there | 14:55 |
| noonedeadpunk | then rust | 14:55 |
| noonedeadpunk | but I think it's overall development priorities which are concerning | 14:56 |
| jrosser | it does seem to install oslo.privsep 3.7.0 which is a bit confusing it doesnt find it | 14:58 |
| noonedeadpunk | Epoxy had 3.6.0 | 14:58 |
| jrosser | though actually having said this, we actually should be concerned about the utility container as thats where the error is happening | 14:58 |
| jrosser | doh no | 14:59 |
| jrosser | its in the response from magnum isnt it | 14:59 |
| * jrosser tired | 14:59 | |
| noonedeadpunk | https://opendev.org/openstack/oslo.privsep/commit/cc5981a43631a049448f9104e4da0d9025e64b7b was the only functional change in between | 14:59 |
| jrosser | ah yes but we do try to change to noble | 15:00 |
| jrosser | so that is a whole new python blah blah | 15:00 |
| noonedeadpunk | noble used to be passing: https://zuul.opendev.org/t/openstack/build/6c94b5e413bc44f28dd2ba5ad4be6372 | 15:00 |
| noonedeadpunk | it could be change to magnum easily... | 15:02 |
| jrosser | true | 15:06 |
| noonedeadpunk | which also result in non-fuynctional heat (unless they've dropped heat completely) | 15:07 |
| noonedeadpunk | dunno, I'd really love to spend some time on another capi driver. As reportedly, control cluster does not matter much, so how we deploy it might be suitable | 15:08 |
| noonedeadpunk | and then it's only the image and settings... | 15:08 |
| noonedeadpunk | will see if I could get time for that | 15:08 |
| noonedeadpunk | but migration between is not possible for sure | 15:09 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-plugins master: Add support for EL10 https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/951314 | 15:10 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-plugins master: Add support for EL10 to GlusterFS role https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/951314 | 15:11 |
| noonedeadpunk | well. there's no MariaDB for EL10 | 15:13 |
| noonedeadpunk | wrote to their zulip: https://mariadb.zulipchat.com/#narrow/channel/236699-Buildbot/topic/el10-and-centos-stream10-packages-availability/with/521116806 | 15:27 |
| noonedeadpunk | It somehow feels that nobody really cares about EL anymore. | 15:32 |
| mnasiadka | noonedeadpunk: there is in CS10 AppStream for now ;) | 15:33 |
| noonedeadpunk | yeah... | 15:34 |
| noonedeadpunk | mnasiadka: btw, it looks like it's coming: https://jira.mariadb.org/browse/MDBF-847 | 15:34 |
| noonedeadpunk | but I'd expect it to be already there, as task was closed more then a month ago | 15:35 |
| noonedeadpunk | they seem to build smth at the moment https://buildbot.mariadb.org/#/builders/872 | 15:41 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Add support for EL10 https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/951315 | 15:44 |
| noonedeadpunk | And no rabbitmq: https://cloudsmith.io/~rabbitmq/repos/rabbitmq-server/packages/?q=distribution%3Ael+AND+distribution%3A10 | 15:46 |
| noonedeadpunk | doh | 15:47 |
| noonedeadpunk | seems really literally nobody cares about it | 15:48 |
| noonedeadpunk | it's almost half a year passed since C10S release | 15:48 |
| noonedeadpunk | and I guess I should not mention, that there're packages for Ubuntu 24.10 | 15:51 |
| NeilHanlon | i can poke at cloudsmith/rabbit/erlang... | 15:52 |
| mnasiadka | Oh boy, just don't mention OpenStack - they get extremely angry then :-) | 15:55 |
| noonedeadpunk | ++ | 15:55 |
| noonedeadpunk | they're VMware folks, so yeah | 15:55 |
| noonedeadpunk | but we had a really sweet discussion the other day, without mentioning openstack :) | 15:55 |
| noonedeadpunk | NeilHanlon: I actually isntalled from EL9 repo without issues | 15:56 |
| noonedeadpunk | as I saw your comment, as el9 contains packages marked as el8 | 15:56 |
| noonedeadpunk | so now, on CentOS 10 Stream, I got that: rabbitmq-server-4.0.9-1.el8.noarch | 15:56 |
| noonedeadpunk | ( ͡° ʖ̯ ͡°) | 15:57 |
| NeilHanlon | so cursed | 15:58 |
| NeilHanlon | just like java packaging now for some reason | 15:58 |
| NeilHanlon | some portable jdk gets built and then the "real" jdk just unpacks what that is.. so that it "looks" right from the outside but java's only built once even if you release the same build with four different skins | 15:59 |
| noonedeadpunk | was really stream announcement _that_ devastating for the ecosystem | 15:59 |
| noonedeadpunk | omfg | 15:59 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Add EL10 support https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/951316 | 16:01 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Do not verify MariaDB version when installed from distro repos https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/951317 | 16:06 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Add support for EL10 https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/951315 | 16:07 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-repo_server master: Fix role idempotency https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/951319 | 16:15 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-repo_server master: Clean-up nginx-related task after 2025.1 release https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/951320 | 16:19 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-python_venv_build master: Add support for EL10 https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/951322 | 16:23 |
| noonedeadpunk | so, setup-infrastructure is passing. The only big trade-off is mariadb version, which is coming from distro | 16:23 |
| noonedeadpunk | but keystone setup failed right away on attempt to build systemd-python wheels | 16:28 |
| noonedeadpunk | and that's coming from glibc, as now it compares float version with string.... | 16:31 |
| noonedeadpunk | And I'd to guess, that the root cause could be in updated systemd from hyperscale | 16:31 |
| noonedeadpunk | https://paste.openstack.org/show/bmKFuo1teehgoc6Dv8gE/ | 16:32 |
| noonedeadpunk | oh, ok, that's our issue apparently | 16:37 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Remove LIBSYSTEMD_VERSION override for AIO https://review.opendev.org/c/openstack/openstack-ansible/+/951324 | 16:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-uwsgi master: Add support for EL10 https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/951325 | 16:53 |
| noonedeadpunk | we're having 2 generic nits as of today related to UX more or less | 16:54 |
| noonedeadpunk | first - /usr/local/bin is not in $PATH for EL. | 16:55 |
| noonedeadpunk | second - bash-completition may be either not installed or not properly configured. this also affects ubuntu 24.04 | 16:55 |
| noonedeadpunk | So I wonder if we wanna apply some bush fixes for environment for that | 16:56 |
| noonedeadpunk | and if we wanna ensure that these things work | 16:56 |
| noonedeadpunk | as I can see how bash-completition can be mis-used in a malicious way... | 16:56 |
| noonedeadpunk | but from other side, it's a great UX tool... | 16:57 |
| noonedeadpunk | EL10 now passes to Nova and fails with modular libvirt. | 17:04 |
| noonedeadpunk | oh wait a sec | 17:11 |
| noonedeadpunk | it's not failing on libvirt | 17:11 |
| noonedeadpunk | it's not even modular there.... | 17:12 |
| noonedeadpunk | or well. it still seems to be able to work in a monolythionc mode | 17:14 |
| noonedeadpunk | so it's kinda up to what services you start I guess.... | 17:29 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Add EL10 support https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/951328 | 17:34 |
| noonedeadpunk | jrosser: so. EL10 is fully passing a simple metal aio right now with https://review.opendev.org/q/topic:%22osa/el10%22 | 17:40 |
| noonedeadpunk | also, it somehow feels _much_ faster to execute roles.... | 17:40 |
| noonedeadpunk | so might be x86_64-v3 requirement was not pointelss after all | 17:41 |
| jrosser | that should come out in the job stats when we get some nodes | 17:52 |
| noonedeadpunk | like second run of setup-openstack goes from start to tempest in 10 mins | 17:52 |
| jrosser | huh | 17:53 |
| noonedeadpunk | in CI we would be also having more modern hardaware | 17:53 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Add EL10 support https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/951332 | 18:13 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest master: Drop EL8 support https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/951333 | 18:13 |
| noonedeadpunk | seems noble has issues with mirrors... | 18:15 |
| noonedeadpunk | https://zuul.opendev.org/t/openstack/build/2d019fc00eee48698c7fb17b417391ee | 18:15 |
| noonedeadpunk | but I wonder why these are not infra mirrors, actually... | 18:15 |
| noonedeadpunk | quite happy about the progress tbh. somehow was under impression, that it's gonna be much-much worse overall | 18:21 |
| noonedeadpunk | but ability to run old monolythinc libvirt helped a lot | 18:22 |
| jrosser | we could check if the failing Ubuntu mirrors thing is using ipv6 | 18:50 |
| jrosser | like CI node is dual stacked | 18:51 |
| noonedeadpunk | and also it fails during lxc cache preparation. as we indeed don't use infra mirrors there iirc | 18:52 |
| jrosser | we can add an apt config to force ipv4 | 18:52 |
| noonedeadpunk | true | 18:53 |
| jrosser | (see #opendev from earlier) | 18:53 |
| noonedeadpunk | I see now | 18:53 |
| jrosser | idk is thats the same issue but certainly worth a check | 18:54 |
| noonedeadpunk | But I was also thinking why not to override lxc_apt_mirror | 18:54 |
| jrosser | do we just forget to use the infra mirror for noble? | 18:55 |
| noonedeadpunk | it was failing here: https://opendev.org/openstack/openstack-ansible-lxc_hosts/src/branch/master/vars/debian.yml#L16-L21 | 18:55 |
| noonedeadpunk | and we don't set mirrors for it | 18:55 |
| noonedeadpunk | btw we are also supposedly using an old/barely supported method to fetch CI info | 18:57 |
| jrosser | hmmm yes I do remember that coming up before | 18:57 |
| noonedeadpunk | this thing: https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/roles/bootstrap-host/tasks/gather_nodepool_vars.yml#L26-L28 | 19:00 |
| jrosser | do you know what the proper way is? | 19:04 |
| jrosser | zuul.blah in a pre playbook perhaps? | 19:04 |
| noonedeadpunk | I don't remeber :( | 19:04 |
| noonedeadpunk | yes, there was smth abouyt zuul vars | 19:05 |
| noonedeadpunk | and can't find that fdast | 19:06 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Use infra mirror for LXC debootstrap https://review.opendev.org/c/openstack/openstack-ansible/+/951338 | 19:11 |
| noonedeadpunk | but this should address current issue ^ | 19:11 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!