16:01:41 <johnsom> #startmeeting Octavia
16:01:42 <openstack> Meeting started Wed Feb 19 16:01:41 2020 UTC and is due to finish in 60 minutes.  The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:45 <openstack> The meeting name has been set to 'octavia'
16:01:45 <johnsom> Hi sorry I am late
16:02:13 <johnsom> rm_work is unable to attend today
16:02:49 <ataraday_> hi
16:02:55 <gthiemonge> Hi
16:03:04 <johnsom> #topic Announcements
16:03:37 <johnsom> The only announcement I have this week is we are now working towards feature freeze
16:03:41 <johnsom> #link https://releases.openstack.org/ussuri/schedule.html
16:04:13 <cgoncalves> hi
16:04:51 <johnsom> Please make sure any new features you want in Ussuri are on track to merge before the milestone 3 date in April
16:05:08 <johnsom> This is a license to bug people for reviews!
16:05:46 <johnsom> However, any new features landing at this point will be looked at closely and likely tagged to merge in the "V" release.
16:06:08 <johnsom> I.e. we need to focus on getting our committed features in.
16:06:26 <johnsom> Any questions / comments about that?
16:07:06 <ataraday_> we still hope to get jobboard in U?
16:07:26 <johnsom> Yes, I think that is probably our #1 feature priority for U
16:08:07 <johnsom> We may not make it the default driver, but we should make sure all of the parts are merged.
16:08:34 <johnsom> Do we all agree on that?
16:08:41 <ataraday_> Yeah, just code, not making the default
16:09:33 <johnsom> Any other comments from the team?
16:10:20 <johnsom> Any other announcements this week?
16:10:48 <johnsom> #topic Brief progress reports / bugs needing review
16:11:35 <johnsom> I have had some internal/downstream stuff to work on over the last week, but that is wrapped now.
16:12:21 <johnsom> Failover is functional for the v1 driver and passing the tests. I am now working on adding the required tests and doing additional manual testing.
16:12:58 <johnsom> My plan is to wrap that up and push to merge for v1 while I port it over to the v2 side. Getting it merged for v1 will also allow us to start the backports.
16:13:30 <johnsom> There are some retries left that I think we want to add as well, which I might try to get in or work on in another patch series later.
16:13:39 <cgoncalves> stable gates should be green again. please recheck any patches previously failed
16:14:15 <johnsom> Ah, yeah, there have been a series of "fix things broken by the py2 transition".
16:14:22 <gthiemonge> cgoncalves: yes, it looks better today
16:14:43 <johnsom> The move to native mock was one of those
16:15:14 <johnsom> However there is some question of if that issue isn't broken in python 3.8 native mock as well.
16:15:30 <johnsom> I have a TODO to open a python bug
16:16:09 <johnsom> Any other updates?
16:17:15 <ataraday_> review for jobboard is welcome :)
16:17:16 <cgoncalves> just a quick update that the devstack patch to allow override of LIBVIRT_CPU_MODE is now merged
16:17:31 <johnsom> ataraday_ I did comment on the etherpad, however it looks like you had already found solutions to those issues.
16:17:43 <cgoncalves> this enables us to use KVM whenever possible
16:18:17 <cgoncalves> #link https://review.opendev.org/703324
16:18:23 <johnsom> cgoncalves So this is the rename of the old setting in devstack?  Does it have any functional differences? Does it do a better job detecting support?
16:18:41 <ataraday_> johnsom, Thanks! Yeah, just verify to failover issue should be solved with failover refactor
16:19:11 <johnsom> Yeah, those should be resolved. We should keep those listed and test when I do the v2 port of the code
16:19:20 <johnsom> Hopefully that will go quickly
16:19:28 <cgoncalves> johnsom, yes. it allows us to set the cpu mode in our gate. if KVM is enabled, it will use it otherwise fallback to QEMU/TCG
16:21:01 <johnsom> Cool, I am all for using the nested virt if we can. It makes a huge difference
16:21:35 <cgoncalves> we'll be using "host-passthrough" whenever KVM is enabled. "host-model" fails on vexxhost
16:22:00 <cgoncalves> host-passthrough works in all nested virt capable nodepool providers
16:22:14 <johnsom> Nice
16:22:39 <cgoncalves> hopefully it will help us greatly in the future as more tempest tests are added
16:23:16 <johnsom> cgoncalves Are you updating our tempest jobs to set that? Is there an open patch for that?
16:23:51 <cgoncalves> yes
16:23:52 <cgoncalves> #link https://review.opendev.org/#/c/702921/
16:24:05 <cgoncalves> I'll update it now that the devstack patch merged
16:24:11 <johnsom> Cool, thanks!
16:24:32 <johnsom> Any other updates before we move on to open discussion?
16:24:57 <cgoncalves> shout-out to ataraday_ for the continued work in jobboard and to gthiemonge for the awesome reviews!
16:25:20 <johnsom> Yes, lots of great work from the team. Very appreciated.
16:26:17 <johnsom> #topic Open Discussion
16:26:27 <johnsom> Any other topics this week?
16:29:22 <johnsom> Ok then, have a great week everyone!
16:29:33 <johnsom> #endmeeting