| opendevreview | Miguel Lavalle proposed openstack/neutron master: [PoC][DNM] Router flavors and service type for OVN https://review.opendev.org/c/openstack/neutron/+/883988 | 01:02 |
|---|---|---|
| opendevreview | Miguel Lavalle proposed openstack/neutron master: [PoC][DNM] Router flavors and service type for OVN https://review.opendev.org/c/openstack/neutron/+/883988 | 01:04 |
| opendevreview | Takashi Kajinami proposed openstack/neutron-lib master: Remove unmaintained API report tooling https://review.opendev.org/c/openstack/neutron-lib/+/899009 | 02:12 |
| frickler | haleyb: FYI the dbperf session is planned for today at 15 UTC https://etherpad.opendev.org/p/oct2023-ptg-os-dbperf-crossproject | 04:58 |
| opendevreview | Luis Tomas Bolivar proposed openstack/ovn-bgp-agent master: Add initial support for local OVN cluster instead of kernel-networking https://review.opendev.org/c/openstack/ovn-bgp-agent/+/881779 | 08:35 |
| opendevreview | Luis Tomas Bolivar proposed openstack/ovn-bgp-agent master: Add initial support for local OVN cluster instead of kernel-networking https://review.opendev.org/c/openstack/ovn-bgp-agent/+/881779 | 09:40 |
| opendevreview | Luis Tomas Bolivar proposed openstack/ovn-bgp-agent master: Add initial support for local OVN cluster instead of kernel-networking https://review.opendev.org/c/openstack/ovn-bgp-agent/+/881779 | 09:57 |
| opendevreview | Luis Tomas Bolivar proposed openstack/ovn-bgp-agent master: Add support at NB BGP Driver for exposing CR-LRPs https://review.opendev.org/c/openstack/ovn-bgp-agent/+/898198 | 10:25 |
| opendevreview | Elvira García Ruiz proposed openstack/neutron master: [WIP] Fix log creation not a sync operation https://review.opendev.org/c/openstack/neutron/+/899037 | 11:39 |
| slaweq | frickler haleyb unfortunately I will not be able to attend this session today, sorry | 13:08 |
| haleyb | slaweq: i can attend, or did you want to reschedule it? | 13:47 |
| opendevreview | Lucas Alvares Gomes proposed openstack/neutron master: [OVN] OVN SB DB sync to account for neutron_sync_mode value https://review.opendev.org/c/openstack/neutron/+/899069 | 14:08 |
| opendevreview | Lucas Alvares Gomes proposed openstack/neutron master: [OVN] OVN SB DB sync to account for neutron_sync_mode value https://review.opendev.org/c/openstack/neutron/+/899069 | 14:11 |
| slaweq | haleyb I don't think it can be rescheduled, I just wanted to give You a heads up that I will not be able to attend it | 14:19 |
| haleyb | slaweq: ack | 14:20 |
| opendevreview | Lucas Alvares Gomes proposed openstack/neutron master: [OVN] OVN SB DB sync to account for neutron_sync_mode value https://review.opendev.org/c/openstack/neutron/+/899069 | 15:25 |
| opendevreview | Harald Jensås proposed openstack/neutron master: [OVN] DB sync host/physnet - filter on agent_type https://review.opendev.org/c/openstack/neutron/+/899077 | 15:47 |
| mlavalle | haleyb: are meeting tomorrow morning normally? | 17:10 |
| mlavalle | are we^^^ | 17:10 |
| haleyb | mlavalle: i didn't plan on it, but maybe i forgot to send email, will do that | 17:11 |
| haleyb | mlavalle: oh, and did you have any problem being bug deputy this week? | 17:14 |
| mlavalle | haleyb: no, that's ok | 17:23 |
| opendevreview | Harald Jensås proposed openstack/neutron master: [OVN] DB sync host/physnet - filter on agent_type https://review.opendev.org/c/openstack/neutron/+/899077 | 18:20 |
| opendevreview | Harald Jensås proposed openstack/neutron master: [OVN] DB sync host/physnet - filter on agent_type https://review.opendev.org/c/openstack/neutron/+/899077 | 18:52 |
| dansmith | haleyb: the performance.json problem was due to the venv switch in devstack recently | 20:32 |
| dansmith | here's an updated run from tempest-full: https://31acd40a51a5aad5461b-3e7bc19bdc58c96c3f64630625f9105c.ssl.cf5.rackcdn.com/899089/1/check/tempest-full-py3/79483bd/controller/logs/performance.json | 20:32 |
| dansmith | neutron selects: 404578, nova selects: 41356 | 20:32 |
| haleyb | dansmith: oh, i remember disabling that for neutron? | 20:32 |
| dansmith | even with the caching problem keystone seems to have, they're still 25% of neutron: 135291 | 20:33 |
| dansmith | haleyb: oh yeah? | 20:33 |
| haleyb | dansmith: i remember something when that broke our world | 20:33 |
| dansmith | are you talking about performance.json or the global venv switch? | 20:34 |
| haleyb | global venv switch | 20:34 |
| dansmith | the latter broke lots of stuff, the former has been *mostly* trouble-free | 20:34 |
| dansmith | ah yeah, okay | 20:34 |
| dansmith | so your perf.json probably has this data in it then | 20:34 |
| haleyb | dansmith: ack, and adding those two variables to local.conf will generate that DB section, right? in another recent gate job i see 'db: []' | 20:40 |
| dansmith | perf stats is on by default, so if you just disabled global venv, it should be generating the db stuff already | 20:41 |
| dansmith | with global venv enabled, my patch is required | 20:42 |
| dansmith | and if you disabled stats gathering, then obviously you need that on too | 20:42 |
| haleyb | dansmith: MYSQL_GATHER_PERFORMANCE=True correct? i see we always disable that | 20:43 |
| dansmith | yeah | 20:43 |
| dansmith | did you disable it because of some structural problem or just a suspicion of an impact to performance? | 20:44 |
| * haleyb shrugs | 20:44 | |
| dansmith | it only updates stats like once a minute so it should be extremely lightweight | 20:44 |
| haleyb | ah, guess we were seeing random CI issues so turned it off | 20:45 |
| * haleyb now chases another thread | 20:45 | |
| dansmith | ack, well, ideally we wouldn't need it enabled all the time of course, but it does offer the opportunity to check on a given patch if any of those numbers increases a lot from the status quo :) | 20:46 |
| haleyb | https://bugs.launchpad.net/neutron/+bug/2015065 which is still open | 20:46 |
| dansmith | yeah that's still a problem AFAIK, so probably not related I think | 20:49 |
| dansmith | "disabling $thing that has a thread makes it slightly less frequent" is often code for "this is not the problem it just perturbs the chaos a bit" :) | 20:49 |
| haleyb | not sure how much to wander into that chaos, but we can always add an experimental job with it enabled while we look | 20:54 |
| dansmith | yep, but also know that nova and the other major projects have it enabled :) | 20:58 |
| haleyb | yeah, we'll need to work on that bug, i'll at least propose a revert for now | 21:00 |
| opendevreview | Brian Haley proposed openstack/neutron master: Revert "Disable mysql gather performance in jobs" https://review.opendev.org/c/openstack/neutron/+/899052 | 21:02 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!