16:00:13 <gthiemonge> #startmeeting Octavia
16:00:13 <opendevmeet> Meeting started Wed Apr 12 16:00:13 2023 UTC and is due to finish in 60 minutes.  The chair is gthiemonge. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:13 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:13 <opendevmeet> The meeting name has been set to 'octavia'
16:00:16 <gthiemonge> Hi
16:00:18 <johnsom> o/
16:00:28 <tweining> o/
16:00:34 <oschwart> o/
16:00:41 <QG> Hello !
16:01:26 <gthiemonge> #topic Announcements
16:01:30 <gthiemonge> * Bobcat Release Schedule: R-25
16:01:42 <gthiemonge> the first commit of the amphorav1-removal chain is merged!
16:01:51 <gthiemonge> #link https://review.opendev.org/q/topic:amphorav1-removal
16:02:01 <gthiemonge> johnsom: I updated the next commit
16:02:05 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/876035
16:02:24 <gthiemonge> you gave me a -1, could you take a look again? (and other folks please ;-)
16:02:33 <johnsom> Ack
16:02:40 <gthiemonge> thank you!
16:03:04 <tweining> ok
16:03:26 <gthiemonge> i don't have any other announcements, did I miss something?
16:05:42 <tweining> do we need to update our zuul jobs definition to test py310?
16:06:22 <tweining> I think we're only testing 3.9 right now
16:06:45 <johnsom> It should be done at the main zuul job
16:06:49 <gthiemonge> I don't think we need it
16:06:59 <gthiemonge> yeah we should inherit from "openstack-python3-jobs"
16:07:10 <johnsom> It is already there
16:07:15 <gthiemonge> https://opendev.org/openstack/octavia/src/branch/master/zuul.d/projects.yaml#L8
16:07:27 <gthiemonge> oh right
16:07:32 <johnsom> openstack-tox-py310
16:07:57 <tweining> thanks
16:08:25 <tweining> I was looking at the "octavia-tox-tips" jobs
16:08:57 <tweining> anyway, that is not an announcement
16:09:11 <johnsom> py39 for tips should be fine, py39 is still in the PTI for bobcat
16:09:12 <gthiemonge> hmm
16:09:37 <gthiemonge> yeah I think we need only one tips job
16:10:52 <gthiemonge> #topic CI Status
16:11:05 <gthiemonge> we are still running our jobs on Ubuntu Focal
16:11:25 <gthiemonge> I'm trying to get the list of the Ubuntu Jammy "broken" hosts
16:11:29 <gthiemonge> we could report them to vexxhost
16:11:41 <gthiemonge> (context: https://bugs.launchpad.net/octavia/+bug/2015244)
16:11:52 <gthiemonge> 2nd item:
16:12:03 <gthiemonge> the periodic job octavia-amphora-image-build is broken (my fault)
16:12:11 <gthiemonge> This commit should fix it:
16:12:18 <gthiemonge> #link https://review.opendev.org/c/openstack/octavia/+/879824
16:12:27 <gthiemonge> johnsom: could you take a look?
16:13:15 <johnsom> Done
16:13:29 <gthiemonge> thanks!
16:15:05 <gthiemonge> #topic Brief progress reports / bugs needing review
16:17:13 <tweining> I managed to get py-spy to work with the CI!
16:17:22 <gthiemonge> nice
16:18:33 <gthiemonge> I proposed a patch that would fix some potential CI issues
16:18:49 <tweining> mmh, looking for the patch but can't find it...
16:18:51 <gthiemonge> (issues that I didn't see upstream, but we've faced them downstream here)
16:18:57 <gthiemonge> #link https://bugs.launchpad.net/octavia/+bug/2015572
16:19:01 <gthiemonge> #link https://review.opendev.org/c/openstack/octavia/+/879898
16:19:19 <gthiemonge> My proposal: for each new subnet plugged into the amphora, the amphora-agent sends a GARP packet
16:19:31 <gthiemonge> Please ready the story and review the patch
16:19:34 <gthiemonge> thanks
16:20:16 <gthiemonge> tweining: https://review.opendev.org/q/topic:py-spy
16:20:19 <gthiemonge> ?
16:20:26 <tweining> sorry for the CR-1, I think I should have CR+1'ed it so that it gets better visibility
16:20:38 <tweining> thanks
16:20:58 <gthiemonge> I'll update it with the annotations
16:21:23 <tweining> ... which brings me to the next topic I've been working on: static code checks with mypy
16:22:42 <tweining> I just started working on it and there is still a lot of work to do. However, I suggest we put more care into adding more type annotations in new code.
16:23:55 <QG> Please review the RFE for Let's encrypt  #link https://review.opendev.org/c/openstack/octavia/+/877281
16:24:30 <gthiemonge> QG: ack
16:24:51 <gthiemonge> tweining: do you plan to add annotations to the existing code?
16:25:07 <tweining> QG: yeah, it's on my list, but this is a big one
16:25:38 <tweining> gthiemonge: I am actually in process of doing it where mypy is complaining.
16:26:29 <tweining> Just to clarify: Not everything needs to have annotions. mypy can derive that information from context and other sources often times.
16:27:35 <tweining> https://review.opendev.org/c/openstack/octavia/+/879749 is the WIP mypy patch BTW
16:28:21 <QG> Thanls !
16:28:21 <tweining> I'm done with octavia.cmd now.
16:28:29 <QG> thanks ! ^^
16:29:12 <gthiemonge> tweining: if we add annotations to a class, does mypy check that they are correct in an inherited class? I'm thinking about our plugins/interfaces (network driver/rest api driver/etc...)
16:29:48 <tweining> yes, I think that should work. mypy is quite smart.
16:30:44 <gthiemonge> ack
16:32:15 <gthiemonge> #topic Open Discussion
16:32:21 <gthiemonge> I have one topic here
16:32:36 <gthiemonge> I proposed a patch that enables zookeeper for jobboard in devstack
16:32:42 <tweining> https://review.opendev.org/c/openstack/openstacksdk/+/867968 is an example where mypy would also need some help. (that change was for helping PyCharm mainly)
16:32:50 <gthiemonge> https://review.opendev.org/c/openstack/octavia/+/862671
16:33:10 <gthiemonge> johnsom raised a good point: why are we installing it from the source instead of using packages
16:33:26 <gthiemonge> I'm using centos and unfortunately there is no package for zookeeper in the default repositories
16:33:43 <gthiemonge> so I'm pulling zookeeper from upstream
16:34:10 <gthiemonge> this week, I tried to add a specific case to my patch for ubuntu (but it seems that the zookeeper package doesn't come with a systemd init file :/)
16:34:24 <johnsom> That is odd
16:34:35 <gthiemonge> if we want to support both methods (deb and source) I need to generate systemd file and config file for both distribs (and of course the content of those files would be different)
16:35:00 <gthiemonge> I don't know if it's worth it to add the support for .deb (I would rather avoid duplicate things)
16:35:41 <gthiemonge> (BTW I didn't mention it in the commit, but I stole this code from the monasca-api repo, I'll fix that)
16:36:10 <gthiemonge> any advices/comments on this?
16:36:11 <johnsom> Yeah, that is fair. If it gets messy because the distros are not complete, we can continue to use the soruce
16:36:22 <gthiemonge> ack
16:36:48 <gthiemonge> I'm setting W-1 because I found out that I messed with the service name
16:37:38 <gthiemonge> thanks for your feedback johnsom
16:38:27 <gthiemonge> any other topics folks?
16:39:02 <tweining> oh, I think I know why I didn't find the py-spy patch I was looking for (the link that gthiemonge posted does not contain the one I mean)
16:39:28 <tweining> when two patches have the same Change-Id will one replace the other?
16:39:39 <gthiemonge> yes
16:40:01 <gthiemonge> (and the same branch)
16:40:48 <tweining> it's the same target branch if you mean that, but a different topic
16:41:21 <gthiemonge> gerrit uses the change-id and the branch name to identify a unique patch
16:41:38 <gthiemonge> (and the repo of course)
16:41:51 <gthiemonge> the topic is only a metadata
16:42:14 <gthiemonge> BTW you can add hashtags now :D
16:42:41 <opendevreview> Tom Weininger proposed openstack/octavia master: DNM: profile w/ sqlalchemy2 using py-spy  https://review.opendev.org/c/openstack/octavia/+/880190
16:43:56 <tweining> mmh, ^ should have a different change-id now. still I don't see the other patch
16:44:40 <gthiemonge> let's have this discussion offline (after the meeting) :D
16:44:47 <gthiemonge> thank you folks
16:44:49 <gthiemonge> #endmeeting