16:00:08 <johnsom> #startmeeting Octavia
16:00:08 <opendevmeet> Meeting started Wed Jul 30 16:00:08 2025 UTC and is due to finish in 60 minutes.  The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:08 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:08 <opendevmeet> The meeting name has been set to 'octavia'
16:00:16 <johnsom> Hi everyone
16:00:17 <gthiemon1e> o/
16:00:33 <johnsom> #topic Announcements
16:00:52 <johnsom> The 2026.1 Gazpacho release schedule has been proposed:
16:00:59 <johnsom> #link https://6b6a75d201a9c42f31a1-68f6595b812925233ee703f28dff66ba.ssl.cf5.rackcdn.com/openstack/b1bdf51c9a9c46efbe95fc8e26c59558/docs/gazpacho/schedule.html
16:02:03 <johnsom> Otherwise, I don't think I have any other announcements other that the gate things I will cover in my progress report.
16:02:36 <johnsom> #topic Brief progress reports / bugs needing review
16:02:59 <johnsom> I have been looking into the odd gate job failures around log offloading.
16:04:24 <johnsom> With help from Greg we found that Ubuntu has some package order bugs that can lead to /var/log removing write for the syslog group. This is bad and rsyslog modules may not be able to create new log files. In our gates it impacted the Octavia log offloading and haproxy on the devstack host.
16:04:47 <johnsom> So, I proposed moving our logs into their own directory to solve this issue.
16:05:35 <gthiemon1e> LGTM!
16:05:58 <johnsom> I also discovered why I was sometimes getting permission denied errors in nova when booting instances. It appears somehow the group ownership of /dev/kvm is swiitched to systemd-resolved group from KVM. Changing it back to kvm seems to resolve the issue.
16:06:26 <johnsom> Forward looking, more reviews and a few bugs I want to look into.
16:07:04 <gthiemon1e> someone pinged me about a old patch taht is still open
16:07:10 <gthiemon1e> https://review.opendev.org/c/openstack/octavia/+/897466
16:07:47 <johnsom> Yeah, surprised that is still open. I was looking at it before the meeting.
16:07:57 <gthiemon1e> interesting bug, a race condition when deleting a LB with multiple listeners, the fix optimizes it by calling a function only once during the deletion
16:08:12 <johnsom> +1
16:08:37 <gthiemon1e> I have another patch:
16:08:42 <gthiemon1e> https://review.opendev.org/c/openstack/octavia/+/946866
16:08:43 <johnsom> There are a number of optimizations that can be done in the flows. The transition to Job Board created some inefficiencies.
16:09:09 <gthiemon1e> the port_id is part of the additional_vip object, but it's never returned by the API
16:09:27 <johnsom> ACK
16:09:50 <gthiemon1e> (it will be required to make the octavia-tempest-plugin tests work with other providers)
16:10:31 <gthiemon1e> (the amphora driver only uses one VIP port, but we cannot assume that the other providers do)
16:11:04 <johnsom> I will take a look
16:12:09 <johnsom> Any other updates this week?
16:12:30 <johnsom> #topic Open Discussion
16:12:32 <gthiemon1e> nop
16:12:36 <johnsom> How about any other topics?
16:12:47 <gthiemon1e> nothing from me
16:13:14 <johnsom> Oh, I should mention, I have removed the centos 8 jobs as there are no longer images available.
16:13:45 <johnsom> Likewise, I have removed CentOS 9 Stream and Rocky 9 from the master branch as python 3.9 is no longer supported
16:14:25 <johnsom> It leaves FIPS untested, but that is what it is.
16:14:27 <gthiemon1e> ack
16:14:47 <johnsom> No point in running jobs that are going to fail in devstack setup
16:15:04 <johnsom> Ok, I think that is what I have. Thanks!
16:15:17 <gthiemon1e> thanks johnsom!
16:15:18 <johnsom> #endmeeting
16:15:34 <johnsom> #endmeeting