16:00:20 <gthiemonge> #startmeeting Octavia
16:00:20 <opendevmeet> Meeting started Wed Dec  4 16:00:20 2024 UTC and is due to finish in 60 minutes.  The chair is gthiemonge. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:20 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:20 <opendevmeet> The meeting name has been set to 'octavia'
16:00:24 <gthiemonge> o/
16:00:28 <tweining> o/
16:00:35 <johnsom> o/
16:01:52 <gthiemonge> #topic Announcements
16:02:34 <gthiemonge> well, no announcements from me this week, I was on PTO :D anything to mention here?
16:02:51 <tweining> mmh, I can't think of anything
16:03:32 <johnsom> I don't think I have anything. Just a reminder that Library freeze is coming up the week of Feb 17th
16:04:29 <tweining> working on the rate limiting rfe might become my personal christmas project ;)
16:04:43 <gthiemonge> lol
16:05:16 <gthiemonge> happy holidays ;-)
16:05:34 <gthiemonge> #topic CI Status
16:05:47 <johnsom> Yeah, getting OVN to behave so I can work on my SR-IOV patches might become mine.
16:05:57 <gthiemonge> I think we're still stuck in octavia-tempest-plugin because of multiple failures
16:06:22 <gthiemonge> hacking, doc, unmaintained branch removal, etc..
16:06:32 <gthiemonge> we may have to squash a patchchain to move forward
16:06:46 <gthiemonge> https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/934270
16:07:05 <tweining> or we could temporarily disable pep8 just to integrate that chain?
16:07:44 <gthiemonge> i don't know which template enables pep8
16:08:33 <johnsom> It's one of the "base" jobs provided
16:08:34 <gthiemonge> could be a solution, like what we're doing here with the doc job: https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/934270/4/zuul.d/projects.yaml
16:09:46 <gthiemonge> https://opendev.org/openstack/openstack-zuul-jobs/src/branch/master/zuul.d/project-templates.yaml#L3489-L3501
16:10:26 <gthiemonge> it's the only job in tempest-plugin-jobs, IMHO it's safe to disable it temporarily
16:11:08 <gthiemonge> any objections?
16:11:27 <johnsom> Nope
16:11:59 <gthiemonge> ok, we have a plan!
16:12:27 <gthiemonge> I will update the patch
16:12:40 <gthiemonge> anything else to report about CI?
16:12:57 <johnsom> Thanks for handling that
16:13:37 <tweining> +1
16:14:04 <tweining> nothing else I think
16:14:11 <tweining> oh wait
16:15:07 <tweining> do the failure we discussed also include the openstack-tox-pip-check-reqs job failure?
16:15:21 <gthiemonge> where's that?
16:15:48 <johnsom> It's not clear if one of the unpin patches will fix that or not. It might just be broken
16:16:06 <tweining> https://review.opendev.org/c/openstack/octavia/+/936811?tab=change-view-tab-header-zuul-results-summary for instance
16:16:12 <gthiemonge> not related
16:16:42 <tweining> honestly, I did not understand that failure
16:16:51 <gthiemonge> yeah I haven't seen this failure
16:17:14 <gthiemonge> so no idea
16:17:49 <johnsom> It's falsely reporting that some requirements in requirements.txt are not used even though they are. It was triggered by the change to Noble from Jammy
16:18:44 <johnsom> IMO it's a nice to have check, but I'm not going to lose sleep if we disable it for now
16:18:48 <tweining> from my perspective it fails, but does not print any error
16:19:19 <johnsom> It's in the log:
16:19:23 <johnsom> https://www.irccloud.com/pastebin/AYjdCtJu/
16:20:04 <johnsom> It is flagging most if not all as unused requirements
16:20:09 <tweining> that does not look like an error to me. I agree that it is probably not crucial to have
16:20:45 <gthiemonge> we can create a launchpad and disable it temporarily
16:20:49 <johnsom> It just catches the case where we remove a dependency from the code and forget to remove it from requirement.txt
16:22:54 <johnsom> I can take this one
16:23:07 <gthiemonge> thanks johnsom
16:23:17 <gthiemonge> #topic Brief progress reports / bugs needing review
16:24:05 <tweining> not much progress, but I worked a little bit on the rate limiting rfe
16:24:54 <tweining> the api-ref part to be precise
16:25:15 <johnsom> I had some holiday time off, but right now I am working on setting up a multinode  devstack on bare metal to continue the SR-IOV for members work.
16:25:53 * johnsom notes, pypi is down... sigh
16:27:59 <gthiemonge> nothing new on my side
16:28:05 <gthiemonge> #topic Open Discussion
16:28:10 <gthiemonge> anything else guys?
16:28:37 <johnsom> Nope
16:30:09 <gthiemonge> ack
16:30:19 <gthiemonge> we can close the meeting
16:30:24 <gthiemonge> thank you guys!
16:30:24 <tweining> maybe https://review.opendev.org/q/topic:%22bug/2090782%22 ?
16:31:44 <tweining> yeah, I noticed it has actually not directly to do with octavia
16:31:58 <gthiemonge> I'm not convinced it should be added to octavia-lib
16:32:07 <johnsom> As we discussed at the PTG I think this is a temp fix. Really we need to go through and make the models flat or limited where it makes sense. Then remove all of this recursive model loading stuff. We have had to hack it so many times....
16:32:27 <gthiemonge> we agreed to use limited_graph as a backportable solution until we find a better way to deal with objects
16:32:38 <tweining> right
16:33:02 <johnsom> Yeah, interestingly enough, I was 90% the driver lib/agent was already flat
16:34:28 <gthiemonge> I'm wondering if there's a need for recursive objects there
16:34:59 <gthiemonge> if all the gets are limited_graph=True, maybe we can do that directly in the agent
16:35:07 <johnsom> We explicitly did not have recursive objects in the driver agent
16:35:10 <johnsom> Yeah, right
16:35:53 <gthiemonge> ok, I'll take a closer look at it
16:35:58 <johnsom> It should probably go here if anywhere: https://github.com/openstack/octavia/blob/master/octavia/api/drivers/driver_agent/driver_get.py
16:36:50 <gthiemonge> hmm yeah
16:36:54 <johnsom> Ah, I see recurse there, so I think someone requested that change, probably for OVN
16:37:49 <gthiemonge> they are updating it there: https://review.opendev.org/c/openstack/octavia/+/936857/1/octavia/api/drivers/driver_agent/driver_get.py
16:38:01 <gthiemonge> still recurse=True but limited_graph is optional
16:40:31 <gthiemonge> I will check with the ovn-provider folks too
16:40:35 <gthiemonge> anything else?
16:42:20 <gthiemonge> I guess that's all for this week
16:42:24 <gthiemonge> thank you!
16:42:29 <gthiemonge> #endmeeting