16:00:12 <gthiemonge> #startmeeting Octavia
16:00:12 <opendevmeet> Meeting started Wed Feb 21 16:00:12 2024 UTC and is due to finish in 60 minutes.  The chair is gthiemonge. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:12 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:12 <opendevmeet> The meeting name has been set to 'octavia'
16:00:15 <gthiemonge> hello
16:00:18 <johnsom> o/
16:00:24 <oschwart> o/
16:00:33 <tweining> o/
16:01:14 <gthiemonge> #topic Announcements
16:01:25 <johnsom> The final caracal release of octavia-lib is due this week.
16:01:27 <gthiemonge> * 2024.1 Caracal Release Schedule: Caracal-3 milestone
16:01:32 <gthiemonge> right
16:02:23 <gthiemonge> I haven't yet approved the release
16:02:25 <tweining> can we get https://review.opendev.org/c/openstack/octavia-lib/+/882127 in before?
16:02:49 <tweining> Takashi updated it today
16:03:07 <johnsom> That is for dev, so not really critical to get into the release
16:03:25 <tweining> right
16:03:45 <johnsom> What about this repr patch?
16:04:30 <johnsom> #link https://review.opendev.org/c/openstack/octavia-lib/+/881701/
16:04:39 <gthiemonge> let's target that for the next release
16:04:52 <gthiemonge> it's not a big requirement
16:04:57 <johnsom> Is that important for debugging that we should try to slip it in? or again  mostly dev focus?
16:05:00 <gthiemonge> and it's backportable
16:05:49 <tweining> well, I think nobody except me really misses that either
16:06:10 <gthiemonge> I already gave a +2, I can update the hash of the release if we merge it today or tomorrow
16:06:21 <johnsom> I will fully admit that I have not looked at lib in a while.
16:06:42 <johnsom> Ok, I will take a look at repr after the meeting
16:07:02 <gthiemonge> so yeah octavia-lib this week
16:07:13 <gthiemonge> and next week is the caracal-3 milestone
16:07:24 <gthiemonge> feature freeze & final release for client libraries
16:07:25 <johnsom> Which means feature freeze.
16:07:50 <gthiemonge> right
16:07:55 <johnsom> As you know SR-IOV support for the VIPs is a major feature we are trying to get into caracal.
16:08:12 <johnsom> Sadly, due to a bunch of other issues, it's coming in close to the wire.
16:08:19 <gthiemonge> right
16:08:49 <gthiemonge> there's an attempt to fix those issues
16:08:54 <johnsom> If you can please have a look at the first two patches and give initial feedback (thank you gthiemonge for already doing this) I would appreciate it.
16:08:54 <gthiemonge> https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/861743
16:09:03 <gthiemonge> but the code doesn't look right
16:09:08 <gthiemonge> I'll update it tomorrow
16:09:29 <gthiemonge> yeah if we fix the issues with the gates, I think SR-IOV could be merged soon
16:09:34 <tweining> does that SRIOV stuff need a special setup for testing?
16:09:35 <johnsom> I hope to have an initial patch posted for the last functional part today. I am working through bugs
16:10:16 <johnsom> Yes, you have to have a NIC that supports SR-IOV until April when the next LTS will have a qemu driver we can use for testing.
16:11:12 <johnsom> So, at this point it's code review and unit/functional tests
16:11:35 <gthiemonge> and check that it doesn't break non-SRIOV stuff :D
16:11:42 <johnsom> +1 to that
16:13:08 <johnsom> Everything should be behind a flavor profile flag
16:15:01 <gthiemonge> I have another announcement
16:15:20 <gthiemonge> * PTL 2024.2 Election Season
16:15:23 <gthiemonge> it's already there
16:15:31 <gthiemonge> PTL election for the next cycle
16:15:33 <johnsom> Ah yes, nominations are open
16:17:17 <gthiemonge> and that's it
16:17:23 <gthiemonge> any other announcements?
16:17:33 <johnsom> That is all I have
16:17:51 <tweining> nothing from me
16:18:01 <gthiemonge> #topic CI Status
16:18:08 <gthiemonge> we already talked about that
16:18:18 <gthiemonge> there are some issues with haproxy in the amphora in the CI
16:18:35 <gthiemonge> I'll spend some time on it tomorrow, on the octavia-tempest-plugin patch
16:18:41 <gthiemonge> we need to fix it for next week
16:18:57 <johnsom> I saw you ran the DNM patch last night, any insights?
16:19:45 <gthiemonge> I wanted to catch the moment when haproxy recreates a worker, but it's difficult to synchronize /bin/free and haproxy :D
16:19:59 <johnsom> Yeah
16:20:18 <gthiemonge> I may also try a DNM patch with a smaller value for haproxy maxconn, just to check if it's really a mem issue
16:21:21 <gthiemonge> we could also change the connection-limit in the test classes that are imapcted
16:21:37 <gthiemonge> I'll keep you updated
16:21:41 <johnsom> +1
16:22:17 <gthiemonge> #topic Brief progress reports / bugs needing review
16:23:34 <johnsom> I am making good progress on wrapping up the SR-IOV work. I have rules being pushed down to the amphora now on LB create (for VRRP protocol for example).
16:23:38 <gthiemonge> the patches for the bugs with the neutron endpoint settings are still open:
16:23:44 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/905794
16:23:51 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/907426
16:23:59 <gthiemonge> I need to re-review the first one
16:24:15 <johnsom> I just need to finish debugging it, put in the tasks in the other flows to also update the rules as needed and finish up test coverage.
16:24:37 <johnsom> #link https://review.opendev.org/q/project:openstack/octavia+topic:%22sriov%22
16:25:05 <johnsom> Those are the two that have already been posted. The -1 zuul vote is the potential memory issue we talked about above.
16:25:29 <tweining> I had a look at the 905794 one today, but I didn't fully understand it
16:25:44 <gthiemonge> johnsom: question, do we need a python-octaviaclient patch?
16:26:05 <gthiemonge> for the type of the vnic?
16:26:23 <johnsom> No, when I pivoted I no longer need lib/client/sdk changes. It's all keyed off of the flavor profile now
16:26:25 <gthiemonge> or maybe it's not mandatory
16:26:49 <gthiemonge> ack
16:26:59 <johnsom> Oh, for the field I added to the LB VIP. Hmm,
16:27:33 <johnsom> The client just ignores it, but yeah, I guess we should update that so it shows up
16:27:54 <gthiemonge> ok
16:27:57 <johnsom> I can probably get that hammered out today
16:28:23 <johnsom> I keep forgetting I added that field
16:32:22 <gthiemonge> #topic Open Discussion
16:33:07 <gthiemonge> any other topics before we close the meeting?
16:33:25 <tweining> no
16:33:39 <johnsom> I can't think of anything.
16:34:08 <oschwart> nothing from me
16:34:13 <gthiemonge> ok!
16:34:17 <gthiemonge> thank you guys
16:34:21 <gthiemonge> have a good week
16:34:24 <gthiemonge> #endmeeting