| opendevreview | Merged openstack/neutron master: bgp: Fix pep8 errors https://review.opendev.org/c/openstack/neutron/+/980912 | 00:11 |
|---|---|---|
| ykarel | thx haleyb we can do same in neutron-lib, /me pushes that | 03:11 |
| opendevreview | yatin proposed openstack/neutron-lib stable/2026.1: [2026.1 only] Switch to 2026.1 neutron-tempest-plugin jobs https://review.opendev.org/c/openstack/neutron-lib/+/981203 | 03:18 |
| ykarel | ohh realized we can't do that yet as some projects have not branched stable/2026.1 so need to wait for them else we would be missing coverage in the CI | 03:54 |
| ykarel | like devstack | 03:54 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == DNM == OVN L3 scheduler test https://review.opendev.org/c/openstack/neutron/+/978861 | 07:05 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Fix OVN L3 scheduler (3) https://review.opendev.org/c/openstack/neutron/+/981211 | 07:44 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == DNM == OVN L3 scheduler test (3) https://review.opendev.org/c/openstack/neutron/+/981212 | 07:44 |
| *** mikal0 is now known as mikal | 08:20 | |
| opendevreview | Harald Jensås proposed openstack/neutron master: Add docstrings for filter_dynamic parameter https://review.opendev.org/c/openstack/neutron/+/981214 | 08:32 |
| opendevreview | Eduardo Olivares proposed openstack/neutron master: bgp: Fix one more pep8 error from functional tests https://review.opendev.org/c/openstack/neutron/+/981215 | 08:36 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Fix OVN L3 scheduler (3) https://review.opendev.org/c/openstack/neutron/+/981211 | 09:12 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == DNM == OVN L3 scheduler test (3) https://review.opendev.org/c/openstack/neutron/+/981212 | 09:13 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Mark ``test_gateway_chassis_balanced_scheduling_multiple_gw_networks`` unstable https://review.opendev.org/c/openstack/neutron/+/981223 | 09:15 |
| ralonsoh | slaweq, lajoskatona ^ please approve this patch until I have a solution for it | 09:15 |
| ralonsoh | thanks! | 09:16 |
| slaweq | ralonsoh +2 | 09:17 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Fix OVN L3 scheduler (3) https://review.opendev.org/c/openstack/neutron/+/981211 | 09:55 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == DNM == OVN L3 scheduler test (3) https://review.opendev.org/c/openstack/neutron/+/981212 | 09:55 |
| opendevreview | Merged openstack/neutron master: Mark ``test_gateway_chassis_balanced_scheduling_multiple_gw_networks`` unstable https://review.opendev.org/c/openstack/neutron/+/981223 | 10:49 |
| ralonsoh | sean-k-mooney, hello! About https://bugs.launchpad.net/neutron/+bug/2144312. The main issue here is always the host_id in the virtual port. We would need this information but adding `ml2_port_bindings.host` to the port breaks many things, including the mech driver itself, that tries to bind the port | 10:49 |
| ralonsoh | so, if we need to have this info but not to update `ml2_port_bindings.host` (that, btw, is not done in any other mech driver... my bad), we can add this info in the `vif_details` | 10:50 |
| ralonsoh | this is a dictionary, mainly used by Nova | 10:51 |
| ralonsoh | but we can add extra parameters that will be discarded by Nova if not recognized | 10:51 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Fix OVN L3 scheduler (3) https://review.opendev.org/c/openstack/neutron/+/981211 | 12:10 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == DNM == OVN L3 scheduler test (3) https://review.opendev.org/c/openstack/neutron/+/981212 | 12:10 |
| sean-k-mooney | ralonsoh: so nova consider it a security hardening operturingyt if a tenatn can ever see the hostname | 12:28 |
| sean-k-mooney | ralonsoh: we provcide a hash of it for normal users that is a hash of the hsot value and there project id so that they cant actully see the host name but can tell if 1 vms are on the same host just not wich host they are on | 12:29 |
| sean-k-mooney | ralonsoh: so my intiall reaction is you can put it in the binding_details but it proably should not be visable to a normal user | 12:30 |
| sean-k-mooney | i knwo the bidning host is i think for legacy reasons | 12:31 |
| sean-k-mooney | but it woudl be preferable if the vif_details was required the service or admin roles to view long term | 12:31 |
| sean-k-mooney | so that octavia can see it but not end users | 12:32 |
| opendevreview | Eduardo Olivares proposed openstack/neutron master: bgp: Fix one more pep8 error from functional tests https://review.opendev.org/c/openstack/neutron/+/981215 | 12:49 |
| sean-k-mooney | ralonsoh: byasiclly to summerise form an api contract point fo view the vif_details is the location owned by neutron (via the mech driver) to store any info that may be consumed but external service that are binding the port | 13:03 |
| sean-k-mooney | so it is entilry propper to use it to provide info to octavia if needed. | 13:04 |
| sean-k-mooney | howver i still think virtual port shoudl be a first clase api resouce like trunk prots in the long term | 13:04 |
| sean-k-mooney | virtual prot shoudl never be attached to a vm so it wont impact nova and yes nova will discard once it does not recognise | 13:05 |
| sean-k-mooney | so that shoudl be releivily safe, ideally you just would not populate the new fiels for non virutal ports | 13:05 |
| ralonsoh | sean-k-mooney, if we remove the host assignation from the virtual ports, it won't be needed to create the new "virtual" type | 13:32 |
| ralonsoh | right now, this is what you have with admin or normal user: | 13:34 |
| ralonsoh | https://paste.opendev.org/show/bO9qetZlMhzKiCrZVGQ7/ | 13:34 |
| sean-k-mooney | ok that is fine | 13:35 |
| ralonsoh | so any binding detail (host, vif details, etc) are not shown | 13:35 |
| sean-k-mooney | normaly user don see the binding_host_id | 13:35 |
| sean-k-mooney | ya if that already filter via policy then no hardening is required | 13:35 |
| sean-k-mooney | you can just have the host be in plain text | 13:35 |
| ralonsoh | so I'm going to propose this, to use vif_details to store the parent port host ID in the virtual port | 13:36 |
| sean-k-mooney | ack | 13:40 |
| haleyb | bcafarel: if you're around can you look at https://review.opendev.org/c/openstack/neutron/+/977907 ? tia | 14:38 |
| opendevreview | Jakub Libosvar proposed openstack/neutron master: bgp: Fix more pep8 errors https://review.opendev.org/c/openstack/neutron/+/981215 | 15:13 |
| opendevreview | Jakub Libosvar proposed openstack/neutron master: bgp: Fix more pep8 errors https://review.opendev.org/c/openstack/neutron/+/981215 | 15:20 |
| opendevreview | Dmitrii Shcherbakov proposed openstack/neutron master: [OVN] Add post-commit external_mac assessment https://review.opendev.org/c/openstack/neutron/+/950757 | 15:34 |
| bcafarel | haleyb: finally out of meetings, let me take a look | 15:38 |
| haleyb | bcafarel: +1 | 15:42 |
| opendevreview | Merged openstack/neutron master: Add docstrings for filter_dynamic parameter https://review.opendev.org/c/openstack/neutron/+/981214 | 16:26 |
| opendevreview | Merged openstack/neutron master: Include dynamic segments in NetworkContext https://review.opendev.org/c/openstack/neutron/+/980637 | 17:40 |
| haleyb | jlibosva: so it seems pep8 for tests might have been disabled with the move to pyproject.toml, i'll see if i can fix it | 18:01 |
| jlibosva | haleyb: I can have a look as well. funny thing we found it just now :) | 18:02 |
| haleyb | i know i fixed it and added a bunch of excludes to .pylintrc | 18:03 |
| jlibosva | aha, so that's why I was confused. I remember we used to exclude all tests modules from the pep8 checks - and I was after why the tests are not skipped in our downstream CI and found out the .pylintrc lists only some specific files or modules | 18:04 |
| jlibosva | so nowadays we do run pep8 on most of the tests files? | 18:04 |
| haleyb | jlibosva: well i thought we did but you seem to have found some that got missed - was that in downstream CI only? | 18:07 |
| haleyb | the last two lines in pyproject.toml should have triggered warning, only ignore "S" style things | 18:08 |
| jlibosva | haleyb: yes, I was able to spot it downstream only - but that is Antelope based and I don't think we do anything special there - just tox -epep8, really | 18:10 |
| haleyb | it might be we only run pylint and not ruff on test tree, need to look at logs | 18:14 |
| opendevreview | Merged openstack/neutron master: bgp: Move lrps_to_chassis_routers under commands.py https://review.opendev.org/c/openstack/neutron/+/980962 | 18:15 |
| opendevreview | Jakub Libosvar proposed openstack/neutron master: bgp: Fix more pep8 errors https://review.opendev.org/c/openstack/neutron/+/981215 | 19:17 |
| opendevreview | Jakub Libosvar proposed openstack/neutron master: bgp: Fix more pep8 errors https://review.opendev.org/c/openstack/neutron/+/981215 | 19:22 |
| *** elvira1 is now known as elvira | 19:51 | |
| haleyb | jlibosva: so i can't reproduce this pep8 failure locally, do you see this running 'tox -e pep' ? I can push a patch to be more selective in what ruff complains about, but have seen it not complain about some issues with spaces. looking in .pre-commit-config.yaml we run autopep8 on everything | 19:55 |
| haleyb | s/pep/pep8 | 19:55 |
| haleyb | if you figure it out please push a patch | 20:00 |
| opendevreview | Rodolfo Alonso proposed x/whitebox-neutron-tempest-plugin master: DNM == Testing https://review.opendev.org/c/x/whitebox-neutron-tempest-plugin/+/980425 | 20:09 |
| opendevreview | Elvira García Ruiz proposed openstack/neutron-lib master: Add PVLAN api extension https://review.opendev.org/c/openstack/neutron-lib/+/975931 | 20:36 |
| jlibosva | haleyb: you tried to reproduce it on master? | 21:53 |
| haleyb | jlibosva: yes, on master | 22:17 |
| jlibosva | haleyb: I'll have a look. I found the issues on the Antelope based version | 22:29 |
| haleyb | thanks. i mean some of those are in non-test code too | 22:32 |
| opendevreview | Merged openstack/neutron-tempest-plugin master: Enable logging API in the fwaas-openvswitch scenario job https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/977780 | 22:53 |
| opendevreview | Brian Haley proposed openstack/neutron master: Change ruff lint checks to be more file-specific https://review.opendev.org/c/openstack/neutron/+/981417 | 23:00 |
| haleyb | jlibosva: ^^ that's my update and kind-of proves there's something running against the test tree | 23:00 |
| opendevreview | Brian Haley proposed openstack/neutron master: Remove ruff E741 ignore directive https://review.opendev.org/c/openstack/neutron/+/981421 | 23:37 |
| opendevreview | Merged openstack/neutron master: Open the 2026.2 (Hibiscus) DB branch https://review.opendev.org/c/openstack/neutron/+/980303 | 23:54 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!