16:00:23 <gthiemonge> #startmeeting Octavia
16:00:23 <opendevmeet> Meeting started Wed Feb  7 16:00:23 2024 UTC and is due to finish in 60 minutes.  The chair is gthiemonge. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:23 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:23 <opendevmeet> The meeting name has been set to 'octavia'
16:00:26 <gthiemonge> hi
16:01:08 <tweining> o/
16:01:36 <johnsom> o/
16:02:52 <gthiemonge> #topic Announcements
16:02:58 <gthiemonge> * EOL/Unmaintained releases
16:03:22 <gthiemonge> FYI we are transitioning Yoga to unmaintained, a final release of octavia for yoga will be released soon (10.1.1)
16:03:53 <gthiemonge> and Train and Ussuri releases are now EOL (End of Life)
16:04:33 <gthiemonge> * 2024.1 Caracal Release Schedule
16:05:00 <gthiemonge> a reminder, the final release for non-client libraries (octavia-lib) is in 2 weeks
16:05:26 <gthiemonge> and Caracal-3 milestone (Feature freeze, final release for client libs) is in 3 weeks!
16:06:37 <gthiemonge> anything to add guys?
16:07:34 <johnsom> Nope, just gasping for air given the short window we have
16:08:28 <gthiemonge> #topic CI Status
16:09:50 <gthiemonge> Now that Yoga is unmaintained for many projects, we need to remove the stable-yoga jobs from octaiva-tempest-plugin gates
16:10:00 <gthiemonge> tkajinam has proposed a change
16:10:05 <gthiemonge> https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/907959
16:10:19 <gthiemonge> but it's currently blocked because of random failures in the CI
16:10:38 <gthiemonge> it seems that haproxy consumes too much memory during some tests and kills itself
16:10:46 <gthiemonge> we are working on a fix in
16:10:50 <gthiemonge> https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/861743
16:12:12 <gthiemonge> #topic Brief progress reports / bugs needing review
16:12:29 <gthiemonge> Just a reminder we have 2 patches for important bugs/patches for the neutron auth settings
16:12:34 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/905794
16:12:37 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/907426
16:13:35 <johnsom> I have posted another patch in the SR-IOV chain enabling nftables on the VIP.
16:13:54 <tweining> https://review.opendev.org/q/topic:%22add_tls_terminated_api_tests-patch16%22+-status:abandoned I've worked on THTTPS and HSTS tempest test patches and I think I fixed most if not all open issues
16:14:02 <pyjou> Hello guys sorry I´m late. I respond to you @johnsom about the resize.
16:14:03 <johnsom> This patch is semi-related and could use reviews for merge now: https://review.opendev.org/c/openstack/octavia/+/898863/1
16:14:13 <johnsom> It removes a bunch of dead code
16:14:42 <johnsom> I posted it in October and it hasn't had any reviews
16:16:53 <johnsom> pyjou Thank you, I think it was mostly about documentation and some guard rails on the resize flavor
16:18:30 <pyjou> I´m agree with you. I think we can add a check before resize that prevents resizing from one topology to another.
16:18:51 <johnsom> Yes, I think that is at least a good short term solution
16:19:24 <pyjou> Perfect, so I'll integrate it into my code
16:22:47 <tweining> because I haven't brought that topic up for some time:
16:23:08 <tweining> https://review.opendev.org/q/topic:%22mypy%22+repo:openstack/octavia please don't forget about the typing/mypy patches
16:23:40 <tweining> https://review.opendev.org/q/topic:%22pyupgrade%22+repo:openstack/octavia ... and also the pyupgrade patches. I've split them into small parts so it's easier to review.
16:24:06 <gthiemonge> ack
16:25:45 <gthiemonge> #topic Open Discussion
16:27:37 <QG> o/
16:27:54 <QG> do you want to talk about https://bugs.launchpad.net/octavia/+bug/2052455 ?
16:28:28 <gthiemonge> QG yep
16:29:30 <gthiemonge> my opinion on it: in case of multivip, the user has to explicitely pass all the subnets they want to plug (with vip_subnet_id or additional_vips[].subnet_id)
16:30:21 <gthiemonge> if a port is passed with vip_port_id, if it has only one subnet, the subnet is the vip_subnet, if it has multiple subnets and the user didn't specify them with the API, it's a validation error
16:30:35 <gthiemonge> like this
16:30:38 <gthiemonge> https://opendev.org/openstack/octavia/src/commit/bbb4713464eee61add6ce32427621e36fa4086e9/octavia/api/v2/controllers/load_balancer.py#L224-L226
16:31:10 <gthiemonge> IMHO the only thing that is missing is that the multivip feature is not validated when a port is passed to POST /loadbalancer
16:33:16 <QG> Totally agree with this
16:34:55 <gthiemonge> I can work on a fix
16:35:44 <gthiemonge> and I think the multivip feature is poorly (or not) documented, we need to improve it (with use cases, limitations, etc..)
16:39:22 <johnsom> I would agree with that
16:41:28 <gthiemonge> cool
16:41:42 <gthiemonge> QG: thanks for the bug report BTW!
16:41:58 <johnsom> +1
16:43:09 <gthiemonge> any other topics?
16:45:04 <johnsom> Nothing from me this week
16:46:00 <gthiemonge> ok
16:46:17 <gthiemonge> thanks for the discussion guys
16:46:19 <gthiemonge> bye
16:46:24 <gthiemonge> #endmeeting