| *** mhen_ is now known as mhen | 01:30 | |
| tobias-urdin | barbican's functional job is using devstack-tox-functional and i'm trying to derive a new job and set a new VAULT_PLUGIN_BACKEND env var in `tox_environment` passed to the job, but for some reason VAULT_PLUGIN_BACKEND env var is empty/unset when functions in `devstack/lib/barbican` is executed, this is the patch https://review.opendev.org/c/openstack/barbican/+/957206 | 08:40 |
|---|---|---|
| tobias-urdin | anybody got hints why the env var is not passed properly devstack-tox-functional -> tox role -> devstack, what am i missing? :( | 08:40 |
| opendevreview | Balazs Gibizer proposed openstack/devstack-plugin-ceph master: DNM: Revert "move ceph jobs to debian 12" https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/957417 | 11:30 |
| *** vhari_ is now known as vhari | 15:02 | |
| melwitt | gmaan: do you know if there is any notion of a tempest test requiring a tempest plugin in order to use a specific service client? for example for barbican service client https://opendev.org/openstack/barbican-tempest-plugin/src/branch/master/barbican_tempest_plugin/plugin.py#L70 | 16:46 |
| melwitt | or is there a different way recommended for using a service client outside of the standard ones in a tempest test? | 16:47 |
| gmaan | melwitt: yes, if test need any other service client than what we have in tempest in-tree then we need to install the plugin having that service client | 17:05 |
| melwitt | gmaan: cool thanks. is there a way to skip test if the plugin is missing or how is that recommended to handle? | 17:06 |
| gmaan | tempest only register the other service client when it detect the plugin and tempest register 1. plugin configs in tempest namespace 2. service clients in tempest service client registry 3. plugin tests in tempest discovry | 17:06 |
| gmaan | melwitt: yes, we can do it via service decorator on test | 17:06 |
| gmaan | let me find some example | 17:07 |
| melwitt | gmaan: ok. I have a skip for if the barbican service is available but I wasn't sure how that would relate to the barbican tempest plugin | 17:08 |
| gmaan | melwitt: ohk, I see. so we might have service but no plugin installed | 17:09 |
| gmaan | I am not how plugin detecting that but you can do it via attribute error when accessing service client. if plugin is missing then service client will not be registered | 17:10 |
| gmaan | that is one way | 17:10 |
| melwitt | gmaan: yes exactly. if there was a way to skip in that case I would add it but I was not sure if that's even possible. I didn't find examples | 17:10 |
| melwitt | ok cool | 17:10 |
| melwitt | that makes sense | 17:11 |
| gmaan | melwitt: there is list plugin command to know what all plugin available but skipping based on required service client availability is better | 17:13 |
| melwitt | gmaan: cool thank you for the guidance :) | 17:13 |
| gmaan | melwitt: I think for future I can add one decorator to detect the plugin is available or not (this is good use case of mixed tets) but for now checking service client can be way. | 17:14 |
| melwitt | ++ | 17:15 |
| sean-k-mooney | elodilles: gmaan by the way i notice we stil have nto dropped ubuntu 22.04 testing form os-vif | 17:20 |
| sean-k-mooney | it timed out on my linux brdige removal patch so im going ot create a patch ahead fo it to remove ubuntu 22.04 assuming there is no objections | 17:21 |
| sean-k-mooney | we kept it last cycle for upgrades | 17:21 |
| sean-k-mooney | https://github.com/openstack/governance/blob/master/reference/runtimes/2025.1.rst#additional-testing-for-a-smooth-upgrade | 17:21 |
| sean-k-mooney | but its not required this cycle https://github.com/openstack/governance/blob/master/reference/runtimes/2025.2.rst | 17:22 |
| sean-k-mooney | althogu readign the testing version im not sure if there is a bug in the runtime defintion its listed for py3.10 testing | 17:22 |
| sean-k-mooney | is tox coverage enouch | 17:24 |
| sean-k-mooney | the job isnt actully fialing it just timed out during the execution of the senario test | 17:26 |
| sean-k-mooney | opendev.org%2Fopendev%2Fzuul-providers/rax-dfw-main ok so that the old xen based rax cloud | 17:27 |
| sean-k-mooney | maybe ill just recheck it and we can see if we want to keep it or not later | 17:27 |
| sean-k-mooney | https://zuul.opendev.org/t/openstack/builds?job_name=os-vif-ovn-ubuntu-jammy&project=openstack/os-vif yat that looks pretty stabel to its proably not actully a problem | 17:29 |
| opendevreview | Douglas Mendizábal proposed openstack/devstack master: Fix iniset to escape backslash characters https://review.opendev.org/c/openstack/devstack/+/956914 | 18:32 |
| opendevreview | Yosi Ben Shimon proposed openstack/tempest master: Add new test case for volume retype https://review.opendev.org/c/openstack/tempest/+/956225 | 18:39 |
| gmaan | sean-k-mooney: sure, there is no harm for extra testing but yes we can drop those | 23:13 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!