16:00:09 <gthiemonge> #startmeeting Octavia
16:00:09 <opendevmeet> Meeting started Wed Jul 19 16:00:09 2023 UTC and is due to finish in 60 minutes.  The chair is gthiemonge. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:09 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:09 <opendevmeet> The meeting name has been set to 'octavia'
16:00:11 <gthiemonge> hi!
16:00:15 <johnsom> o/
16:00:16 <oschwart> o/
16:01:08 <gthiemonge> #topic Announcements
16:01:17 <gthiemonge> * Reminder: vPTG October 2023 Dates & Registration Kickoff
16:01:19 <tweining> o/
16:01:23 <gthiemonge> https://lists.openstack.org/pipermail/openstack-discuss/2023-July/034363.html
16:02:36 <gthiemonge> * Proposals for reforming extended maintenance
16:03:24 <opendevreview> Omer Schwartz proposed openstack/octavia master: WIP Fix vip ports leak on fully-populated load balancer create  https://review.opendev.org/c/openstack/octavia/+/887593
16:03:36 <gthiemonge> FYI there's a proposal to replace the EM with "Unmaintained" branches
16:03:49 <gthiemonge> https://lists.openstack.org/pipermail/openstack-discuss/2023-July/034350.html
16:03:55 <gthiemonge> https://review.opendev.org/c/openstack/governance/+/888771
16:04:11 <gthiemonge> I haven't read all that stuff yet, but we need to have an eye on it
16:05:12 <tweining> I like the idea of reducing the number of maintained releases
16:05:28 <johnsom> Yeah, seven is way too many
16:06:53 <gthiemonge> any other announcements folks?
16:07:05 <johnsom> nope
16:07:19 <gthiemonge> alright
16:07:20 <tweining> maybe that octavia-lib 3.3.0 has been released? or was that a news last week already?
16:07:44 <gthiemonge> I don't remember
16:08:01 <gthiemonge> yeah this release provides the support for the HSTS RFE
16:08:04 <tweining> anyway, I'm happy about it. thanks for the reviews
16:08:32 <tweining> but it's quite a few more patches for HSTS still
16:08:37 <gthiemonge> guys, I'm skipping the CI status topic... nothing to report
16:09:36 <gthiemonge> #topic Brief progress reports / bugs needing review
16:10:07 <gthiemonge> FYI I rebased some old octavia-tempest-plugin commits (almost forgotten): "Add member test with multiple subnets on same network"
16:10:12 <gthiemonge> https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/815313
16:10:26 <gthiemonge> it helped me to indentify a bug with the amphorav1 driver
16:10:30 <gthiemonge> I proposed a fix:
16:10:37 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/888669
16:10:49 <gthiemonge> (the bug doesn't break anything, but the tests detects the incorrect behavior)
16:11:12 <gthiemonge> (note: it affects only stable branches because amphorav1 was removed from master)
16:11:47 <tweining> I've updated a few of my HSTS patches after I received some feedback.
16:12:20 <tweining> https://review.opendev.org/q/topic:hsts-haproxy the pytest patches are not directly related but it made debugging tests so much easier for me.
16:13:20 <gthiemonge> hopefully they will be merged soon, the feature looks good to me
16:13:36 <tweining> seems the main patch passes the Zuul tests now. I think all patches are good now
16:13:49 <gthiemonge> i'm going to test them again
16:14:21 <tweining> yeah, I will do some manual testing as well next week
16:16:15 <gthiemonge> #topic Open Discussion
16:16:21 <gthiemonge> I have one topic
16:16:34 <gthiemonge> There's a new bug "Reboot of Amphora does not work with altname in ip link"
16:16:44 <gthiemonge> https://bugs.launchpad.net/octavia/+bug/2027995
16:17:11 <gthiemonge> Basically, rebooting an amphora is not handled correctly, the vip/member interfaces are not moved to the amphora-haproxy ns
16:17:31 <gthiemonge> (reproduced on centos, don't know if ubuntu is affected)
16:17:41 <gthiemonge> so this line doesn't work:
16:17:47 <gthiemonge> https://opendev.org/openstack/octavia/src/branch/master/octavia/amphorae/backends/agent/api_server/templates/amphora-netns.systemd.j2#L29
16:18:17 <gthiemonge> yummy
16:18:20 <johnsom> Is that still needed with the new amphora-agent management of the interfaces?
16:18:46 <tweining> uff, that line looks complicated
16:18:56 <gthiemonge> johnsom: yeah because the script doesn't move the ifaces to the namespace, basically it replaces ifup
16:19:17 <gthiemonge> so I think we need to replace that line with a script (python/shell?)
16:19:45 <gthiemonge> we could fix the line, but I would prefer another solution
16:20:00 <johnsom> I wrote that a long, long time ago...
16:20:22 <gthiemonge> Assigned to -> johnsom ;-)
16:21:17 <gthiemonge> I'll check maybe we can handle it in "amphora-interface" I don't remember if we store everything we need to move them
16:21:59 <johnsom> I think it should move into the amphora agent. Historically everything that was needed was stored in  /var/lib/octavia/plugged_interfaces.sorted
16:22:27 <johnsom> Which comes from /var/lib/octavia/plugged_interfaces
16:22:43 <gthiemonge> but it needs to be done before "amphora-interface up all" (we can move it to the amphora-agent too)
16:22:47 <johnsom> But it seems like the agent should handle this  on startup now
16:23:12 <johnsom> Right
16:23:17 <gthiemonge> ack
16:23:21 <gthiemonge> makes sense
16:23:48 <gthiemonge> I'll add a comment to the launchpad
16:24:24 <gthiemonge> thanks for the feedback tweining johnsom
16:24:53 <gthiemonge> any other topic for today folks?
16:25:21 <tweining> not from me
16:25:26 <oschwart> nothing from me
16:25:44 <johnsom> Nope
16:25:57 <gthiemonge> alright! thank you!
16:26:03 <gthiemonge> #endmeeting