16:00:41 <gthiemonge> #startmeeting Octavia
16:00:41 <opendevmeet> Meeting started Wed Nov 16 16:00:41 2022 UTC and is due to finish in 60 minutes.  The chair is gthiemonge. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:41 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:41 <opendevmeet> The meeting name has been set to 'octavia'
16:00:47 <gthiemonge> o/
16:00:55 <johnsom> o/
16:00:58 <tweining> o/
16:01:15 <matfechner> o/
16:03:04 <gthiemonge> #topic Announcements
16:03:12 <gthiemonge> ** Antelope-1
16:03:16 <johnsom> It is MS1 week
16:03:23 <gthiemonge> this week is Antelope-1 Milestone
16:03:25 <gthiemonge> yep
16:03:39 <johnsom> I will be proposing a MS1 tag for octavia-tempest-plugin this week
16:03:47 <gthiemonge> johnsom: thanks
16:04:26 <gthiemonge> we have also planned to review/merge some RFEs before this milestone
16:04:30 <fkr> o/
16:04:37 <fkr> (sorry, a tad late)
16:04:43 <gthiemonge> I think that things are moving well for the cpu-pinning RFE from tweining
16:05:20 <gthiemonge> (and it's not a big issue if it is not merged this week, at least we got interesting feedback on it)
16:05:55 <tweining> yes, I think I will not change things and wait for reviews now
16:07:17 <gthiemonge> ack
16:07:33 <tweining> different topic: I read that call for papers for the next summit is open
16:07:33 <gthiemonge> any other announcements that I have missed?
16:08:17 <johnsom> #link https://lists.openstack.org/pipermail/openstack-discuss/2022-November/031201.html
16:08:44 <gthiemonge> interesting
16:09:19 <johnsom> Note this link for how to setup a forum session:
16:09:22 <johnsom> #link https://cfp.openinfra.dev/app/vancouver-2023/20/
16:10:07 <johnsom> Hmm, I wonder if that is the correct link, but it is what they had in the email
16:11:23 <gthiemonge> ack
16:11:55 <tweining> wow, 1000 submission on average.
16:12:51 <johnsom> If anyone is interested in submitting a session, feel free to ask me questions. I have done a few over the years.
16:14:52 <gthiemonge> thanks
16:15:01 <gthiemonge> #topic CI Status
16:15:21 <gthiemonge> FYI the (periodic) FIPS job is failing with timeouts
16:15:27 <gthiemonge> https://zuul.openstack.org/builds?job_name=octavia-v2-dsvm-scenario-fips&project=openstack/octavia
16:15:43 <gthiemonge> I proposed to split it into 2 jobs (one with the traffic tests, one with the non-traffic tests)
16:15:48 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/864391
16:16:00 <gthiemonge> a test showed that it would fix those failures
16:16:04 <gthiemonge> https://zuul.opendev.org/t/openstack/buildset/d9eaf090ccba4ddc80925eadf69983c2
16:16:42 <gthiemonge> (note: there's a another FIPS job in the check pipeline that uses only the tls_barbican scenario tests)
16:16:52 <johnsom> Nice
16:19:43 <gthiemonge> #topic Brief progress reports / bugs needing review
16:21:05 <gthiemonge> Cores: we have 3 pages of open backports: https://review.opendev.org/q/project:openstack/octavia+status:open+branch:%255Estable/.*
16:21:06 <johnsom> I have finished the Barbican secrets consumer patch. It will not pass the Barbican tests until the Barbican team fixes a bug in the client. But I fixed that locally and had successful test runs.
16:21:13 <gthiemonge> johnsom: rm_work: tweining: gthiemonge: ^
16:21:34 <johnsom> #link https://review.opendev.org/c/openstack/octavia/+/864308
16:21:34 <gthiemonge> johnsom: great!
16:22:10 <tweining> I will have a look
16:24:14 <QG> I'have commented on #link https://review.opendev.org/c/openstack/octavia/+/859387
16:24:23 <gthiemonge> I proposed a fix for the bug that QG described last week:
16:24:28 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/864192
16:24:48 <gthiemonge> ^ but I don't know if we should translate exception strings, any idea johnsom?
16:25:15 <gthiemonge> QG: oh thanks, I will take a look
16:26:03 <johnsom> Translation is a good question. I don't think OpenStack is translating error messages anymore. I have only seen Dashboard and release notes
16:26:24 <gthiemonge> but an exception might be displayed in the dashboard
16:26:50 <johnsom> Yeah, we pass through a lot of exception strings in our dashboard.
16:27:21 <gthiemonge> https://codesearch.opendev.org/?q=ValidationException&i=nope&literal=nope&files=&excludeFiles=&repos=openstack/octavia
16:27:30 <johnsom> #link https://review.opendev.org/c/openstack/octavia/+/415646
16:28:42 <QG> About logs i have  a question what do you think of having in every log message the loadbalancer id ?
16:28:59 <johnsom> I guess we still have the "_" in common/exceptions, so we probably should tag those. I don't think we have done a good job at that however. We might want to audit
16:30:14 <johnsom> Well, not all logs are related to a load balancer specifically. I.e. amphora instances not yet associated with an LB.
16:30:41 <johnsom> Generally, you can take the request ID, grep for that and be able to tie it back to the root resource.
16:31:23 <QG> amphora during spawn are not associated to any LB ?
16:31:27 <QG> ohhh
16:31:29 <QG> ok
16:31:50 <johnsom> Well, no, we still have the "spares" code I think. Where we boot instances before they are assigned.
16:32:07 <gthiemonge> the spare feature was removed
16:32:21 <QG> it's more in a way of we want to centralize all the logs from a specific LB
16:32:35 <johnsom> Personally, I think the log lines are already too long. I also think it will be a bit of work to pass the LB ID down all of the flows.
16:34:06 <johnsom> Yeah, I get the idea of tracing. OpenStack has been aligned around the request ID.
16:34:37 <johnsom> We do tag all of the tenant flow logs with the LB that produced it. It's just the control plane that does not.
16:34:57 <johnsom> What do others think on this?
16:35:56 <gthiemonge> I like the idea (I spent a lot of time looking at the logs), but yeah it might be complicated to implement
16:36:02 <tweining> I'm fine with having the request id I think
16:36:17 <tweining> and yes, log lines are long already
16:36:18 <gthiemonge> about the request ID, I think that sometimes, we are losing this context
16:36:52 <johnsom> We are, we have not fully implemented the request ID
16:36:57 <gthiemonge> johnsom: would it be possible to re-use the request id from the API in the controller?
16:37:00 <gthiemonge> ah ok
16:37:07 <gthiemonge> maybe we could start by fixing it
16:37:09 <johnsom> The API is fully implemented, but the backends are not
16:37:44 <johnsom> Yeah, we should do that
16:39:47 <johnsom> "I think there is an open story for that" (tm)
16:39:51 <johnsom> grin
16:40:49 <gthiemonge> a "what"?
16:41:33 <johnsom> Maybe in an hour I can find it
16:42:04 <johnsom> #link https://storyboard.openstack.org/#!/story/1694861
16:42:10 <johnsom> Wow, I got lucky
16:42:26 <gthiemonge> nice
16:42:28 <QG> :-)
16:45:32 <gthiemonge> ok, folks, I think we are already in the "Open Discussion" topic
16:47:31 <gthiemonge> any other topics?
16:47:58 <tweining> I will spend more time on reviewing things in the next few days.
16:48:26 <gthiemonge> same here
16:48:30 <tweining> we also have a few very tiny changes that can be reviewed very quickly
16:48:40 <oschwart> o/ my latest showing up to upstream meetings
16:48:50 <oschwart> I will review patches as well
16:49:02 <tweining> better late than never ;)
16:49:24 <gthiemonge> cool, thank you folks!
16:49:39 <gthiemonge> #endmeeting