16:00:40 <noonedeadpunk> #startmeeting openstack_ansible_meeting
16:00:41 <openstack> Meeting started Tue Nov 17 16:00:40 2020 UTC and is due to finish in 60 minutes.  The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:44 <openstack> The meeting name has been set to 'openstack_ansible_meeting'
16:01:35 <noonedeadpunk> #topic bug triage
16:01:42 <noonedeadpunk> https://bugs.launchpad.net/openstack-ansible/+bug/1904327
16:01:43 <openstack> Launchpad bug 1904327 in openstack-ansible "msgpack version causing linuxbridgeagent to fail on ubuntu 18" [Undecided,New]
16:01:55 <openstackgerrit> James Denton proposed openstack/openstack-ansible-os_neutron master: Test OVS/OVN deployments on CentOS 8  https://review.opendev.org/762661
16:01:58 <noonedeadpunk> jamesdenton: have you ever seen smth like this?
16:02:21 <jamesdenton> can't say i have
16:03:09 <jrosser> python2 on train? is that right?
16:03:28 <noonedeadpunk> I think stdout from stein
16:03:35 <jamesdenton> i see related bug here: https://bugs.launchpad.net/oslo.privsep/+bug/1844822
16:03:35 <openstack> Launchpad bug 1844822 in oslo.privsep "linuxbridge agent crash after R ->S upgrade" [Undecided,New]
16:04:21 <noonedeadpunk> it really is
16:04:27 <jamesdenton> same OP
16:06:47 <jrosser> i guess the best we can do is prod oslo to see if they can patch it?
16:07:44 <noonedeadpunk> ok, so now it's max_buffer_size
16:08:14 <noonedeadpunk> https://msgpack-python.readthedocs.io/en/latest/api.html#msgpack.Unpacker
16:09:22 <jrosser> that matches up with this https://bugs.launchpad.net/oslo.privsep/+bug/1844822/comments/5
16:09:23 <openstack> Launchpad bug 1844822 in oslo.privsep "linuxbridge agent crash after R ->S upgrade" [Undecided,New]
16:09:54 <noonedeadpunk> but default is 1024*1024*100
16:11:55 <noonedeadpunk> Ok, well, I can bump ofc version of msgpack but it's in u-c so the only chance to do that is to use global-requirement-pins
16:12:05 <noonedeadpunk> however I'm not sure if it's our bug or not
16:12:23 <noonedeadpunk> I'd rather left it to oslo and try to figure out the best way here to proceed
16:13:06 <openstackgerrit> James Denton proposed openstack/openstack-ansible-os_neutron master: Test OVS/OVN deployments on CentOS 8  https://review.opendev.org/762661
16:13:08 <noonedeadpunk> oh, wait, it's msgpack 1.0 docs
16:14:26 <noonedeadpunk> but it does not support py2....
16:17:24 <noonedeadpunk> ok, so interesting one was https://bugs.launchpad.net/openstack-ansible/+bug/1903846
16:17:26 <openstack> Launchpad bug 1903846 in openstack-ansible " libvirt fails to start because --listen parameter is not permitted with systemd activation socket " [Undecided,New]
16:17:41 <noonedeadpunk> I haven't check it in detail yet
16:19:52 <noonedeadpunk> I'm wondering what libvirt do we get on systmes...
16:20:08 <noonedeadpunk> as it's failing since 5.7.0
16:21:04 <noonedeadpunk> bionic on train looks good
16:21:54 <noonedeadpunk> so it's from https://github.com/libvirt/libvirt/commit/3a6a725b8f575890ee6c151ad1f46ea0ceea1f3b
16:23:00 <jrosser> does this say which OS versions the bug report is for
16:24:05 <noonedeadpunk> nope, don't see it
16:24:15 <noonedeadpunk> But focal will have issues for sure
16:25:17 <noonedeadpunk> libvirt docs are pretty confusing for me.....
16:25:29 <jrosser> so do we need those socket listener services started/stopped on nova_libvirtd_listen_tcp/tls ansible vars
16:25:53 <jrosser> systemctl start libvirtd-tls.socket
16:25:56 <jrosser> that sort of thing
16:26:30 <noonedeadpunk> and we should not use libvirtd.conf for it anymore
16:26:41 <jrosser> https://opendev.org/openstack/puppet-nova/src/branch/master/releasenotes/notes/libvirt-systemd-socket-units-f7b0fc3a4f3c5219.yaml
16:26:43 <noonedeadpunk> but um... how to define address to listen on...
16:27:27 <noonedeadpunk> well, ok, it's valid and has high priority
16:27:48 <jrosser> yeah so its config_template and [socket] a.b.c.d:port sort of format
16:28:18 <noonedeadpunk> yeah, seems like that
16:28:19 <jrosser> or systemd_service with overrides
16:28:48 <jrosser> https://www.freedesktop.org/software/systemd/man/systemd.socket.html
16:29:27 <noonedeadpunk> it was so easier in libvirtd.conf :)
16:30:16 <noonedeadpunk> also what I'm wondering if we can just move to systemd for previous branches - if this new way won't break pre 5.7 libvirt
16:32:07 <noonedeadpunk> I think that's most interesting unporcessed ones.
16:34:36 <jrosser> we should also be close to a victoria RC?
16:35:06 <noonedeadpunk> #topic ffice hours
16:35:30 <noonedeadpunk> well, regarding rc - we need branching for that and I'm pretty stuck with fixing upgrade jobs for magnum
16:35:47 <noonedeadpunk> but I've placed beta
16:36:01 <jrosser> is there something to look at for the magnum problem?
16:36:07 <noonedeadpunk> https://review.opendev.org/#/c/762821/
16:36:31 <noonedeadpunk> So I've got https://review.opendev.org/#/c/758461/
16:36:59 <noonedeadpunk> but https://review.opendev.org/#/c/755569/5 still fails
16:37:14 <noonedeadpunk> and it's not only about magnum - octavia fails the same way
16:37:35 <noonedeadpunk> they both don't have ability to sent endpoint in tempest plugin
16:38:10 <noonedeadpunk> magnm fails on tempest https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_bf0/755569/5/check/openstack-ansible-upgrade-aio_metal-ubuntu-focal/bf0ca28/logs/openstack/aio1-utility/stestr_results.html
16:39:05 <noonedeadpunk> I'm about to jsut abandon 758461 and set upgrade jobs non-oting for magnum and octavia....
16:39:22 <noonedeadpunk> will probably try to spend some little time on fixing....
16:39:35 <noonedeadpunk> but really concidering just setting then to nv
16:40:51 <jrosser> interesting that octavia fails, becasue i have some memory of fixing the tempest plugin
16:41:07 <jrosser> that was perhaps for no ssl verification rather than endpoint
16:42:14 <johnsom> We have this: https://github.com/openstack/octavia-tempest-plugin/blob/master/octavia_tempest_plugin/config.py#L48
16:43:00 <noonedeadpunk> oh, will try this out now
16:45:34 <noonedeadpunk> hm, we already have it https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/roles/bootstrap-host/templates/user_variables_octavia.yml.j2#L24
16:45:46 <noonedeadpunk> but not during test https://zuul.opendev.org/t/openstack/build/d3e6a58048f944aea338979c65626418/log/logs/etc/host/tempest/tempest.conf.txt#164
16:45:48 <noonedeadpunk> interesting
16:46:03 <openstackgerrit> Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_octavia master: Fix linter errors  https://review.opendev.org/755576
16:49:18 <jrosser> why are those missing?
16:50:18 <noonedeadpunk> maybe I depended on patch that wasn't rebased on top of the fix
16:51:17 <noonedeadpunk> it was added not so far ago or for octavia we're re-defining
16:51:22 <noonedeadpunk> https://opendev.org/openstack/openstack-ansible/commit/eae210982469d3fdb190e61acb6d19df5f008374
16:51:55 <noonedeadpunk> probably it was more than that
16:52:10 <noonedeadpunk> when amphora was trying to reach octavia api or smth like this...
16:56:55 <noonedeadpunk> but from my point this is the major things that stops us at the moment
16:58:00 <openstackgerrit> Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Added Netplan examples for a Openstack Infrastructure node.  https://review.opendev.org/760090
16:58:46 <noonedeadpunk> and also we need to get to https://review.opendev.org/#/c/752040/3 one day...
16:58:57 <noonedeadpunk> as it's really a bug we need to look at
17:00:28 <noonedeadpunk> #endmeeting