| scoopex | johnsom: I did some research work and made a proposal to adress that topic: https://input.scs.community/JgMcHJ8oQTKk0cC9UzgUjw | 06:48 |
|---|---|---|
| scoopex | johnsom: Can we add this to the agenda on the next octavia meeting on Wednesday? | 06:49 |
| scoopex | johnsom: i will be available on meeting time. | 07:47 |
| danfai | scoopex: would something like this cover your use-case already? Just saw it while going through some MRs https://review.opendev.org/c/openstack/octavia/+/923939 | 07:56 |
| scoopex | Probably, i will have a look | 08:01 |
| skraynev | @johnsom : hi. did you find a chance to propose an alternative fix for https://bugs.launchpad.net/octavia/+bug/2052682 ? I want to help with review it. | 08:52 |
| scoopex | danfai: A quick view into it gives me the assumption that at least it is at least dependend. I am not sure if it implements placement logic (see octavia/controller/worker/v2/controller_worker.py, comment line 393 and 509 and octavia/controller/worker/v2/tasks/compute_tasks.py, line 190) | 11:08 |
| danfai | scoopex: I haven't checked the MR in detail either, just conceptually it would use explicitly defined different availability zones to spawn amphoras in those, instead of server groups. The configuration for AZs in nova would then be the one to make sure they are placed differently, so Octavia does not need to know about placement | 11:12 |
| danfai | In general it might make sense to have the proposed server group scope in nova, that you were proposing though. And if present probably also makes sense to have this option in Octavia then. (personal opinion) | 11:15 |
| opendevreview | Tobias Urdin proposed openstack/octavia stable/2025.1: Support rotating server_certs_key_passphrase key https://review.opendev.org/c/openstack/octavia/+/949503 | 13:13 |
| tobias-urdin | johnsom: ^ i have a vague memory from PTG that you said backporting above would be ok | 13:13 |
| frickler | johnsom: gthiemonge: please expedite https://review.opendev.org/c/openstack/python-octaviaclient/+/948981 if possible | 15:07 |
| johnsom | frickler Yes, I have already +2'd that | 15:08 |
| johnsom | skraynev Yes, I started on this, but had to switch to addressing some stable branch and gate issues. I am still working on it | 15:10 |
| priteau | Hello Octavia developers. We discovered a bug in the OVN provider after upgrading to Caracal (I think it affects Bobcat too), I found an existing patch and updated it. Could you please review? https://review.opendev.org/c/openstack/ovn-octavia-provider/+/925893 | 15:11 |
| johnsom | priteau The ovn-octavia-provider is a neutron project (not octavia), you will need to request this in the #openstack-neutron channel | 15:12 |
| frickler | johnsom: yes, maybe you could single-approve if no other reviewers are around, as this is affecting kolla-ansible CI and possibly others? | 15:13 |
| frickler | we'll likely also need a 2025.2 release with it once merged | 15:13 |
| johnsom | frickler Yes, I will give the other cores a few hours and push it. (some are travelling) | 15:13 |
| priteau | johnsom: Thanks, I guess I should open a bug in neutron with the ovn-octavia-provider tag too? | 15:15 |
| johnsom | priteau We can just move that bug over to the neutron project | 15:16 |
| johnsom | priteau Oh, that bug is already closed, so yes, neutron may want a bug for that patch | 15:16 |
| priteau | I am opening a new one now | 15:18 |
| priteau | Done: https://bugs.launchpad.net/neutron/+bug/2110488 | 15:27 |
| johnsom | Thanks! | 15:27 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!