| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/2025.1: Set Rocky 9 molecule test as non-voting https://review.opendev.org/c/openstack/openstack-ansible/+/966258 | 07:59 |
|---|---|---|
| jrosser | we seem to be missing `liberasurecode` on rocky for swift role | 08:59 |
| noonedeadpunk | mornings | 09:19 |
| noonedeadpunk | do we have a related failure somewhere? | 09:19 |
| noonedeadpunk | til: https://github.com/openstack/liberasurecode | 09:21 |
| noonedeadpunk | so we were taking it from the rdo? | 09:22 |
| noonedeadpunk | as it does not seem to be ever in some more widespread repos | 09:24 |
| noonedeadpunk | damiandabrowski: btw. do you know if we have plans to add openbao driver to ssh_keypairs role? As that would kinda make sense for 2026.1? | 09:48 |
| damiandabrowski | didn't thought about it before, but it would make sense indeed | 09:49 |
| jrosser | noonedeadpunk: here on centos10 https://zuul.opendev.org/t/openstack/build/9cff0576602c426e958ceee85535df47/log/job-output.txt#14132 | 09:56 |
| jrosser | same for rocky10 https://zuul.opendev.org/t/openstack/build/f21ae5f9ead142ac904826f5777c8f02/log/job-output.txt#13281 | 09:56 |
| jrosser | i was just noticing that we have still a few of these to fix https://review.opendev.org/q/topic:%22osa-wsgi-module%22+status:open | 09:57 |
| noonedeadpunk | damn, I clean forgot about this topic | 09:58 |
| jrosser | i think mistral and aodh are probably good now | 09:59 |
| jrosser | the -1 might not be right | 10:00 |
| noonedeadpunk | yeah, I assume that it's wrong at this point - posted recheck already to get new logs | 10:05 |
| noonedeadpunk | but I indeed not sure what to do with liberasurecode except build c code from source, which is /o\ | 10:06 |
| noonedeadpunk | maybe worth checking RDO on master.... | 10:07 |
| noonedeadpunk | CI was very red there last time I checked | 10:07 |
| noonedeadpunk | and also RDO might not exist really soon. It's still unclear if there will be release for 2026.1 for it at all | 10:07 |
| jrosser | i wonder what swift does (maybe nothing?) | 10:13 |
| noonedeadpunk | it does in case of using EC | 10:19 |
| jrosser | i don't think there is any rhel-10-ish jobs though | 10:20 |
| jrosser | only centos-9-stream | 10:20 |
| noonedeadpunk | well there was smth? https://trunk.rdoproject.org/centos10-master/report.html | 10:21 |
| noonedeadpunk | not sure though | 10:21 |
| noonedeadpunk | https://trunk.rdoproject.org/centos10-master/consistent/ | 10:22 |
| noonedeadpunk | but probably not deps | 10:23 |
| noonedeadpunk | so it's here https://trunk.rdoproject.org/centos10-master/deps/latest/x86_64/ | 10:24 |
| noonedeadpunk | and it's present | 10:24 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts master: Install master RDO deps https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/966515 | 10:29 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_ironic master: Add apparmor rules for ironic inspector https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/951003 | 11:06 |
| noonedeadpunk | let's see if that's gonna help | 11:07 |
| opendevreview | Damian Dąbrowski proposed openstack/ansible-role-pki master: Add hashi_vault backend https://review.opendev.org/c/openstack/ansible-role-pki/+/948881 | 11:41 |
| jrosser | still not sure why we don't use signed_by instead of these paths in the user facing data structure for vault | 12:10 |
| jrosser | even if it ends up pointing to `pki_{{ blah.signed_by }}` for the actual path in vault | 12:10 |
| noonedeadpunk | so like https://paste.openstack.org/show/bMqPFpZ8M3AX6qoODG46/ ? | 12:11 |
| noonedeadpunk | or well, it should be name, I guess... | 12:12 |
| noonedeadpunk | pretty much as name is irrelevant is there... | 12:15 |
| noonedeadpunk | so to create intermediate, request need to go to `pki_root/root/sign-intermediate` | 12:15 |
| noonedeadpunk | and when you issue root it is `pki_root/root/generate/internal` | 12:16 |
| noonedeadpunk | and `pki/intermediate/generate/internal` for intermediate generation (first one was to sign) | 12:17 |
| noonedeadpunk | but indeed, we could rename vault_root_ca_path -> signed_by | 12:17 |
| noonedeadpunk | if that's what you meant | 12:18 |
| noonedeadpunk | damiandabrowski: ^ | 12:19 |
| damiandabrowski | we can rename vault_root_ca_path to signed_by | 12:27 |
| damiandabrowski | I didn't do this before because I wanted to avoid a situation where we have the same variable(signed_by) that accepts different values depending on the backend(cert name for standalone, vault path for hashi_vault). | 12:27 |
| damiandabrowski | but maybe you don't see it as a problem | 12:28 |
| damiandabrowski | and its name - vault_root_ca_path was more explicit that this variable is about the vault path | 12:28 |
| damiandabrowski | so that was the reason why I implemented it that way, but I don't have a strong opinion here. | 12:29 |
| damiandabrowski | If you think it's better to rename it to signed_by, I'm okay with this | 12:29 |
| noonedeadpunk | we'd need another vote on the bump: https://review.opendev.org/c/openstack/openstack-ansible/+/966496 | 13:14 |
| opendevreview | Merged openstack/openstack-ansible stable/2025.1: Set Rocky 9 molecule test as non-voting https://review.opendev.org/c/openstack/openstack-ansible/+/966258 | 13:20 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tacker master: Fix usage of modules via FQCN https://review.opendev.org/c/openstack/openstack-ansible-os_tacker/+/966553 | 13:35 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-pki master: Add hashi_vault backend https://review.opendev.org/c/openstack/ansible-role-pki/+/948881 | 14:20 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible-ops master: DNM: Test updating mcapi versions to latest https://review.opendev.org/c/openstack/openstack-ansible-ops/+/966563 | 14:22 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible-os_magnum master: DNM: Test mcapi CI https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/966564 | 14:24 |
| opendevreview | Ivan Anfimov proposed openstack/openstack-ansible-os_tacker master: Remove deprecated run_tests/vagrantfile https://review.opendev.org/c/openstack/openstack-ansible-os_tacker/+/960901 | 14:35 |
| opendevreview | Ivan Anfimov proposed openstack/openstack-ansible-os_tacker master: Remove functional test environments from tox.ini https://review.opendev.org/c/openstack/openstack-ansible-os_tacker/+/960903 | 14:35 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible-ops master: DNM: Test updating mcapi versions to latest https://review.opendev.org/c/openstack/openstack-ansible-ops/+/966563 | 14:53 |
| mnaser | andrewbonney: hey! just saw your pr for the facts, is there an ansible-lint option we can toggle to static scan this/ | 14:54 |
| andrewbonney | I'm not familiar enough I'm afraid, but jrosser or noonedeadpunk may know (this is re: inject facts) | 14:55 |
| mnaser | context: https://github.com/vexxhost/ansible-collection-containers/pull/72 :) | 14:55 |
| jrosser | ahha i was looking for that just earlier | 14:55 |
| noonedeadpunk | I'm not sure about the ansible-ling, but there is for molecule... | 14:56 |
| jrosser | like not really according to this https://github.com/ansible/ansible-lint/discussions/3457 | 14:56 |
| mnaser | ah we use molecule to run all these so we can flip that | 14:56 |
| noonedeadpunk | that's what we have: https://opendev.org/openstack/ansible-role-pki/src/branch/master/molecule/default/molecule.yml#L30-L32 | 14:57 |
| noonedeadpunk | so doing same for your molecule config should do the trick | 14:57 |
| mnaser | ok cool cool | 14:58 |
| mnaser | btw what we've been doing is using delegated inventory in molecule so we feed it the zuul inventory.yaml and we're able to have molecule exec natively directly into it | 14:58 |
| mnaser | essentially molecule.yml is as small as this: https://github.com/vexxhost/ansible-collection-containers/blob/main/extensions/molecule/docker/molecule.yml -- and the nwe copy the zuul inventory file in the task before the one that runs molecule | 14:59 |
| mnaser | and so we can just test against any platform against a native zuul vm | 14:59 |
| noonedeadpunk | oh, that's neat actually | 15:00 |
| jrosser | what do you do for local hacking? | 15:00 |
| mnaser | that is the part that remains not user friendly lol | 15:00 |
| mnaser | but i've come to terms with a big red warning in prepare.yml that says | 15:01 |
| mnaser | this is going to blow up your system because it is going to run against localhost | 15:01 |
| mnaser | way too many quirks with docker based testing in my experience that gets you not a real experience :\ | 15:01 |
| jrosser | oh yes we have also a bad experience with this | 15:02 |
| mnaser | also the nice thing with this approach is its entirely ssh based, so any of those weird things that dont get uncovered bc it tries to use local or system python.. | 15:02 |
| mnaser | they are all eliminated because its pure ssh | 15:02 |
| noonedeadpunk | I kinda can recall some networking challanges if you'd want to have your hosts connected in a "special" way in zuul? | 15:02 |
| mnaser | so the build node usually has an ssh key that can ssh to itself | 15:02 |
| mnaser | and so that.. just works | 15:02 |
| mnaser | https://zuul.oss.vexxhost.dev/build/b0cda8ed24ee4b9e91060c8e39b78925/console | 15:03 |
| noonedeadpunk | well, I'm more thinking about multi-node things | 15:03 |
| mnaser | yeah if you need to expose a network across multiple nodes or something like lets say keepalived, then yeah that will be tricky | 15:03 |
| mnaser | but you can always use the multinode roles to get the vxlan/etc bridge wired and then use that in your role downstream | 15:04 |
| noonedeadpunk | oh, yes, true | 15:04 |
| noonedeadpunk | so then it's more tradeoff of local testing vs CI testing kinda | 15:05 |
| jrosser | i think i am still having nightmares about trying to test the connection plugin with molecule | 15:06 |
| jrosser | it so really does not want to do that | 15:06 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible-ops master: DNM: Test updating mcapi versions to latest https://review.opendev.org/c/openstack/openstack-ansible-ops/+/966563 | 15:07 |
| mnaser | anyways its an interesting approach that you can keep at the back of your mind if it comes in handy :) certainly made our testing easier and more "realistic" | 15:07 |
| noonedeadpunk | but then for other chunk of stuff it's really neat to spend 5 mins on local test to ensure you didn't broke anything | 15:07 |
| noonedeadpunk | it's really nice one, and I didn't realize it's actually possible | 15:07 |
| noonedeadpunk | so thanks for sharing it! | 15:08 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible-ops master: DNM: Test updating mcapi versions to latest https://review.opendev.org/c/openstack/openstack-ansible-ops/+/966563 | 15:08 |
| mnaser | im gonna get the ci fixed since its trying to deploy python 3.14rc right now, and get andrew change landed, and then enable that ci option so we dont break again | 15:08 |
| andrewbonney | thanks! | 15:08 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-python_venv_build master: Define trusted host for wheels build https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/966404 | 15:23 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible-ops master: DNM: Test updating mcapi versions to latest https://review.opendev.org/c/openstack/openstack-ansible-ops/+/966563 | 15:27 |
| mnaser | andrewbonney: looks like testing caught a few other instances, ill try an addres sthem in my pr | 15:28 |
| mnaser | affecting rockylinux-9 | 15:28 |
| andrewbonney | ah ok, I was only testing on ubuntu so that makes sense | 15:28 |
| mnaser | https://github.com/vexxhost/ansible-collection-containers/pull/74 if curious | 15:29 |
| noonedeadpunk | it's so annoying that there's no nice way to test collections still... | 15:39 |
| noonedeadpunk | but it's a smart layout :) | 15:39 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible-ops master: DNM: Test updating mcapi versions to latest https://review.opendev.org/c/openstack/openstack-ansible-ops/+/966563 | 15:59 |
| mnaser | do you guys point to commits of the collection or releases? | 16:05 |
| mnaser | aka would you like a release for the collection or not necessary? | 16:05 |
| noonedeadpunk | I think we should be able to install just with SHA, so release is highly optional, imo | 16:05 |
| jrosser | hmmm `osa_toolkit.generate.MultipleIpForHostError: Host aio1 has both 162.253.55.186 and 162.253.55.70 assigned` | 16:36 |
| jrosser | wonder what that is | 16:36 |
| jrosser | argh doh https://github.com/openstack/openstack-ansible-ops/blob/master/mcapi_vexxhost/playbooks/files/openstack_deploy/conf.d/k8s.yml | 16:56 |
| jrosser | that really needs to template in these two things https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/roles/bootstrap-host/defaults/main.yml#L119-L120 | 16:56 |
| jrosser | /o\ so much complexity here with the driver playbooks being out of tree | 16:56 |
| jrosser | not sure what to do about this tbh as it's just a `copy:` module currently to drop the config file into place | 17:04 |
| jrosser | noonedeadpunk: if you have any good idea here ^ | 17:16 |
| noonedeadpunk[e] | I need to finally figure out missing part for azimuth and move that all together... | 17:20 |
| noonedeadpunk[e] | There was just some extra helm charts which needed to be present in k8s for it to work.... | 17:21 |
| jrosser | i think we kind of broke things with this https://github.com/openstack/openstack-ansible/commit/31f12f54eaccd3ff5a458cb8593897a46b2d923a | 17:25 |
| jrosser | we dont have access to the `bootstrap_*` variabled except in the bootstrap-host role | 17:26 |
| opendevreview | Merged openstack/openstack-ansible-plugins stable/2025.1: Revert "Set Ansible config directory to /tmp for molecule" https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/966248 | 17:42 |
| opendevreview | Merged openstack/openstack-ansible-os_tacker master: Fix usage of modules via FQCN https://review.opendev.org/c/openstack/openstack-ansible-os_tacker/+/960920 | 20:29 |
| opendevreview | Merged openstack/openstack-ansible-os_tacker master: tox: Remove ineffective ignore_basepython_conflict and bump minimum version https://review.opendev.org/c/openstack/openstack-ansible-os_tacker/+/960902 | 20:29 |
| opendevreview | Merged openstack/openstack-ansible-os_tacker master: Remove deprecated run_tests/vagrantfile https://review.opendev.org/c/openstack/openstack-ansible-os_tacker/+/960901 | 20:29 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!