| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible stable/2023.1: Ensure "parallel" package is installed for CI log collection https://review.opendev.org/c/openstack/openstack-ansible/+/929405 | 07:52 |
|---|---|---|
| jrosser | good morning | 08:05 |
| kleini | good morning :) | 08:08 |
| noonedeadpunk | пщщв ьщктштп | 09:24 |
| noonedeadpunk | crap, mornings :D | 09:28 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Use mariadb client instead of mysql for healthcheck https://review.opendev.org/c/openstack/openstack-ansible/+/922839 | 09:29 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-haproxy_server master: Respect defined interface for external VIP with LE https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/924333 | 09:29 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-galera_server master: Improve handling of galera_install_method changing https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/927901 | 10:01 |
| opendevreview | Merged openstack/openstack-ansible-rabbitmq_server master: Manage apt repositores and keys using deb822_repository module https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/907833 | 11:24 |
| opendevreview | Merged openstack/openstack-ansible stable/2023.1: Ensure "parallel" package is installed for CI log collection https://review.opendev.org/c/openstack/openstack-ansible/+/929405 | 13:05 |
| noonedeadpunk | huh, weird bug https://bugs.launchpad.net/openstack-ansible/+bug/2080765 - and I think we doo see the same in our CI actually | 14:29 |
| noonedeadpunk | So I kinda wonder if we should jsut install selinux bindings only for utility container or everywhere just to be on the safe side | 14:29 |
| noonedeadpunk | so add python3-libselinux here https://opendev.org/openstack/openstack-ansible/src/branch/master/inventory/group_vars/utility_all/defaults.yml#L31 or here: https://opendev.org/openstack/openstack-ansible-openstack_hosts/src/branch/master/vars/redhat-9.yml#L48 | 14:31 |
| noonedeadpunk | I'm thinking about second actually | 14:32 |
| noonedeadpunk | but that feels really as CentOS bug | 14:35 |
| noonedeadpunk | NeilHanlon: maybe you have some idea of wtf CentOS is doing with selinux to require bindings when selinux is disabled? | 14:36 |
| jrosser | i wonder if we should mark support for centos as experimental now | 14:53 |
| jrosser | becasue we basically make the CI nv and just ignore it | 14:53 |
| noonedeadpunk | given that other openstack projects dropped CentOS CI overall... | 15:12 |
| noonedeadpunk | and it's not even in PTI | 15:12 |
| noonedeadpunk | though we got 2 quite close to each other reports in a wekk | 15:13 |
| noonedeadpunk | which means there's some interest for it | 15:13 |
| noonedeadpunk | frankly -I've spotted failure a while ago, though was assuming it will just go away as it used to with CentOS | 15:18 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts master: Ensure python libselinux bindings for containers https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/929540 | 15:18 |
| NeilHanlon | mornin' -- will take a look | 15:19 |
| noonedeadpunk | to have that said - rocky is unaffected | 15:19 |
| noonedeadpunk | but if I understand how things are done - if nothing to do with CentOS - it will get promoted to RHEL and thus to Rocky? | 15:20 |
| noonedeadpunk | but it's really weird, I'm not even completely sure where things go wrong - Ansible, some collection or CentOS. | 15:22 |
| noonedeadpunk | It;s jsut I'm inclined to blame CentOS as Rocky is absolutely fine. | 15:22 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts master: Rename libselinux python package bindings https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/929542 | 15:23 |
| NeilHanlon | very well could be something that we'll see in Rocky/RHEL in November | 15:44 |
| NeilHanlon | i can think of at least four other times CentOS Stream broke in a similar way | 15:45 |
| NeilHanlon | so, not a crazy thought lol | 15:45 |
| noonedeadpunk | yeah.... but this time it kinda feels more ansible tbh, as I kinda don't feel anything too weird running shell commands inside containers | 15:46 |
| noonedeadpunk | bu again - could be a lot of things... | 15:46 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron stable/2024.1: Disable uWSGI usage by default https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/929549 | 16:10 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron stable/2023.2: Disable uWSGI usage by default https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/929550 | 16:10 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron stable/2023.1: Disable uWSGI usage by default https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/929551 | 16:11 |
| jrosser | noonedeadpunk: i see we don't actually test on centos here https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/929540/1 | 17:39 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: DNM - Trest CentOS https://review.opendev.org/c/openstack/openstack-ansible/+/929562 | 17:49 |
| noonedeadpunk | quick hack ^ | 17:49 |
| noonedeadpunk | wow | 17:49 |
| noonedeadpunk | zuul obviously got updated | 17:49 |
| noonedeadpunk | I can't say it become more convenient though - I'd say otherwise | 17:50 |
| noonedeadpunk | takes at least 3 more clicks to see jobs for the tas | 17:51 |
| WireLost | hello, very nice to see OSA adding Ubuntu 24.04 support | 20:33 |
| WireLost | quick note: to deploy AIO with 'install_method: distro', I had to add a line to file '/etc/ansible/roles/uwsgi/vars/debian.yml', below '_uwsgi_distro_libpython:' block containing: ' noble: libpython3.12' | 20:35 |
| WireLost | otherwise it would fail to setup openstack | 20:36 |
| jrosser | WireLost: be sure to understand what is written here about 24.04 https://governance.openstack.org/tc/reference/runtimes/2024.2.html | 20:56 |
| jrosser | we also classify install_method: distro as experimental https://docs.openstack.org/openstack-ansible/latest/admin/upgrades/compatibility-matrix.html | 21:00 |
| jrosser | and finally, we do no testing whatsoever of ubuntu noble for install_method: distro | 21:00 |
| WireLost | jrosser, ok, thanks for the info! | 21:55 |
| jrosser | WireLost: if you would like to submit a patch for the uwsgi role for the libpython fix, please do.... else i can take a look at it tomorrow | 21:58 |
| WireLost | sounds fun! but I don't have any contributor account yet... No rush though! :-P | 22:00 |
| WireLost | I also want to share that I just ran 'openstack-ansible' and it passed 'ceph-install.yml' without any issues! With OSA AIO master branch on 24.04, I have Ceph 19 up and running in a laboratory, looks promising! | 22:28 |
| WireLost | I had to set some variables to `user_variables.yml` after `bootstrap-aio.sh`, so that Ceph playbooks passed... | 22:29 |
| WireLost | Added to '/etc/openstack_deploy/user_variables.yml': 'install_method: distro', 'apply_security_hardening: false', 'ceph_pkg_source: distro' | 22:30 |
| WireLost | Added to '/etc/openstack_deploy/user_variables_ceph.yml': 'ceph_origin: distro', 'ceph_stable_release: squid', 'docker: false, 'containerized_deployment: false', and 'dashboard_enabled: false'. | 22:31 |
| jrosser | well like I say we don’t test any of that :) | 22:47 |
| jrosser | and there is some value having your deployment exactly replicated in what does pass our CI | 22:47 |
| jrosser | support for installing from distro packages has mostly been added by enthusiastic people associated with those distros | 22:48 |
| jrosser | the source install stuff on the other hand is mostly coming from cloud operators | 22:49 |
| jrosser | and the secret weapon of the source installes is being able to trivially apply patched code to your deployment for whatever bugs or other issues you encounter | 22:50 |
| WireLost | got it, but Ceph didn't work initially because Ceph upstream doesn't provide 'noble' APT repo yet, so 'distro' for Ceph to pass was the only way I got it working on 24.04... And gee, no docker please lol | 22:52 |
| jrosser | we also do not recommend using the built in ceph integration for production | 22:55 |
| jrosser | as that is there primarily as a test harness | 22:55 |
| WireLost | Right, I also got isolated Ceph 19 being deployed only with 'ceph-ansible' 'stable-8.0' branch, on 24.04, all good, no OSA. But I think it's kind of neat to use 'openstack-ansible' to deploy and maintain Ceph while leveraging its nice inventory under '/etc/openstack_deploy/'... | 22:57 |
| WireLost | what could possibly go wrong? :-D | 22:58 |
| jrosser | well indeed | 23:00 |
| jrosser | there’s a fairly sparse grid of all these choices that are actually a good idea | 23:00 |
| WireLost | yep, understood, looks like it's tricky to find the perfect (tested enough) 'combo' =P | 23:03 |
| WireLost | I knew that 24.04 uses Python 3.12, but I didn't realize it's not tested with OpenStack (thanks for the info, BTW). How does Canonical plan to provide support for it in an LTS release? | 23:07 |
| WireLost | I also had to change '/etc/apparmor.d/local/abstractions/libvirt-qemu' to '/etc/apparmor.d/abstractions/libvirt-qemu', new path in 24.04, it seems (TASK: 'ceph_client : Write AppArmor configuration for ceph immutable object caching') | 23:36 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!