| opendevreview | Brian Haley proposed openstack/neutron master: Remove support for setting 'trusted' key in binding:profile https://review.opendev.org/c/openstack/neutron/+/986217 | 00:13 |
|---|---|---|
| opendevreview | Merged openstack/neutron master: Fix netlink test privsep entrypoint unwrapping https://review.opendev.org/c/openstack/neutron/+/1009010 | 00:20 |
| *** haleyb|out is now known as haleyb | 00:31 | |
| opendevreview | Merged openstack/neutron master: Restrict ``delete_port`` for the OVN metadata port https://review.opendev.org/c/openstack/neutron/+/1008762 | 01:36 |
| opendevreview | OpenStack Proposal Bot proposed openstack/networking-bgpvpn stable/2026.2: Imported Translations from Zanata https://review.opendev.org/c/openstack/networking-bgpvpn/+/1009140 | 02:40 |
| opendevreview | Takashi Kajinami proposed openstack/networking-bagpipe master: Drop sys_platform marker for win32 system https://review.opendev.org/c/openstack/networking-bagpipe/+/1009154 | 05:56 |
| opendevreview | Takashi Kajinami proposed openstack/neutron-fwaas master: Drop sys_platform marker for win32 system https://review.opendev.org/c/openstack/neutron-fwaas/+/1009155 | 05:56 |
| opendevreview | Takashi Kajinami proposed openstack/tap-as-a-service master: Drop sys_platform marker for win32 system https://review.opendev.org/c/openstack/tap-as-a-service/+/1009156 | 05:57 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: ml2: replace tunnel segment physical_network NULL with '' https://review.opendev.org/c/openstack/neutron/+/1007844 | 06:49 |
| ralonsoh | haleyb, hi! How are you doing with the wsgi patch? | 06:59 |
| ralonsoh | We would need it in 2026.2 too | 06:59 |
| lajoskatona | ralonsoh: Hi, the vpnaas rally job is green: https://review.opendev.org/q/topic:%22bug/2163189%22 | 07:06 |
| lajoskatona | I cut the commit in 2 to have one for rally and one for the race for status update (so rally catched it it seems) | 07:07 |
| lajoskatona | ralonsoh, mtomaska : another stadium, tap-as-a-service: https://review.opendev.org/q/topic:%22bug/2150209%22 , to introduce OVO for it, if you have time please check the series | 07:10 |
| ralonsoh | lajoskatona, nice! let me check | 07:10 |
| lajoskatona | warning: more in the pipe for vpnaas and dynamic-routing | 07:11 |
| opendevreview | Marcin Wilk proposed openstack/neutron stable/2025.2: Ensure MaintenanceWorker lock set before connect https://review.opendev.org/c/openstack/neutron/+/1009159 | 07:39 |
| -opendevstatus- NOTICE: zuul processing is broken since about 03:00 UTC, investigation is in progress, please be patient | 08:02 | |
| *** tobias-u1 is now known as tobias-urdin | 08:19 | |
| *** mikal0 is now known as mikal | 08:57 | |
| opendevreview | chanyeol yoon proposed openstack/neutron-specs master: Add OVN ACL flow sampling to Network Log https://review.opendev.org/c/openstack/neutron-specs/+/1005660 | 09:42 |
| opendevreview | chanyeol yoon proposed openstack/ovsdbapp master: nb: add support for mirror-rules in mirrors https://review.opendev.org/c/openstack/ovsdbapp/+/950742 | 10:20 |
| opendevreview | chanyeol yoon proposed openstack/neutron-lib master: Add the tap-mirror-lport API definition for TaaS https://review.opendev.org/c/openstack/neutron-lib/+/1009192 | 10:47 |
| opendevreview | Fiorella Yanac proposed openstack/neutron-tempest-plugin master: Add additional Json Tempest api tests for indirect floating IP https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/1009017 | 11:04 |
| opendevreview | chanyeol yoon proposed openstack/tap-as-a-service master: Add lport tap mirrors: mirror a port to another Neutron port https://review.opendev.org/c/openstack/tap-as-a-service/+/1009194 | 11:07 |
| opendevreview | fivetime proposed openstack/ovn-octavia-provider master: Match HM events to the exact member https://review.opendev.org/c/openstack/ovn-octavia-provider/+/1009199 | 11:16 |
| opendevreview | fivetime proposed openstack/ovn-octavia-provider master: Report a member ERROR when its monitor is deleted https://review.opendev.org/c/openstack/ovn-octavia-provider/+/1009200 | 11:16 |
| opendevreview | fivetime proposed openstack/ovn-octavia-provider master: Handle Service_Monitor rows created offline https://review.opendev.org/c/openstack/ovn-octavia-provider/+/1009201 | 11:16 |
| opendevreview | fivetime proposed openstack/ovn-octavia-provider master: Apply HM events after the LB operation reported https://review.opendev.org/c/openstack/ovn-octavia-provider/+/1009202 | 11:16 |
| opendevreview | fivetime proposed openstack/ovn-octavia-provider master: Keep the HM mapping of members sharing an address https://review.opendev.org/c/openstack/ovn-octavia-provider/+/1009203 | 11:16 |
| opendevreview | OpenStack Proposal Bot proposed openstack/neutron-fwaas-dashboard stable/2026.2: Imported Translations from Zanata https://review.opendev.org/c/openstack/neutron-fwaas-dashboard/+/1009205 | 11:20 |
| opendevreview | OpenStack Proposal Bot proposed openstack/neutron-vpnaas-dashboard stable/2026.2: Imported Translations from Zanata https://review.opendev.org/c/openstack/neutron-vpnaas-dashboard/+/1009212 | 11:41 |
| opendevreview | Merged openstack/neutron-fwaas-dashboard stable/2026.2: Imported Translations from Zanata https://review.opendev.org/c/openstack/neutron-fwaas-dashboard/+/1009205 | 11:46 |
| opendevreview | Merged openstack/neutron stable/2026.1: Ensure MaintenanceWorker lock set before connect https://review.opendev.org/c/openstack/neutron/+/1008327 | 12:38 |
| opendevreview | Merged openstack/neutron-vpnaas-dashboard stable/2026.2: Imported Translations from Zanata https://review.opendev.org/c/openstack/neutron-vpnaas-dashboard/+/1009212 | 12:45 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: ovn: Add upgrade check for stale OVN Metadata agent tags https://review.opendev.org/c/openstack/neutron/+/1008016 | 12:52 |
| haleyb | ralonsoh: i should get to it today, yesterday was mostly offline | 13:03 |
| opendevreview | Merged openstack/networking-bgpvpn stable/2026.2: Imported Translations from Zanata https://review.opendev.org/c/openstack/networking-bgpvpn/+/1009140 | 13:06 |
| ralonsoh | nice, thanks! | 13:07 |
| opendevreview | chanyeol yoon proposed openstack/tap-as-a-service master: Add lport tap mirrors: mirror a port to another Neutron port https://review.opendev.org/c/openstack/tap-as-a-service/+/1009194 | 13:51 |
| opendevreview | Maor Blaustein proposed x/whitebox-neutron-tempest-plugin master: Add EVPN-BGP FRR gateway HA failover test https://review.opendev.org/c/x/whitebox-neutron-tempest-plugin/+/1009097 | 13:53 |
| opendevreview | Merged openstack/tap-as-a-service master: Drop sys_platform marker for win32 system https://review.opendev.org/c/openstack/tap-as-a-service/+/1009156 | 13:55 |
| opendevreview | Merged openstack/networking-bagpipe master: Drop sys_platform marker for win32 system https://review.opendev.org/c/openstack/networking-bagpipe/+/1009154 | 14:10 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: ml2: derive segment IDs from ``networksegments`` https://review.opendev.org/c/openstack/neutron/+/1007849 | 14:10 |
| otherwiseguy | haleyb: on https://review.opendev.org/c/openstack/neutron/+/989565, I checked with some of our OVN devs and they uniformly rejected using PB.chassis as a signal that HA_CHassis_Group priorities should be renumbered. They also rejected BFD status as an indicator, since the link between A and B being down doesn't mean A or B is down. | 14:11 |
| otherwiseguy | I felt bad, so burned some tokens for a non-preemptive mode in OVN and linked it in the comment which may work? | 14:13 |
| otherwiseguy | I haven't tested it other than the in-patch tests. But neutron would just set options:preempt=false on the HACG and OVN would prefer the Chassis that was currently up and maintain the order of the backups. | 14:14 |
| otherwiseguy | https://github.com/otherwiseguy/ovn/commit/36ff23aa46ab425f67bef0a765ff1c0c91d85080 | 14:14 |
| haleyb | otherwiseguy: thanks for the update and the patch. I know Marcin had added a PTG topic on this, i'll ping him and our OVN team to see what we can do here. | 14:17 |
| haleyb | We just won't be able to backport that :( | 14:17 |
| otherwiseguy | haleyb: yeah. I tried to come up with *something* neutron only, and the OVN team didn't really like anything (which makes sense, my initial thought was "this seems like something only OVN can do well)" but I had hoped I could come up with *something*. | 14:19 |
| haleyb | it is what it is | 14:19 |
| haleyb | otherwiseguy: there was this other change that is seemingly related - https://review.opendev.org/c/openstack/neutron/+/1008752 - don't know if you had seen that or had an opinion | 14:19 |
| otherwiseguy | The demoting the chassis everywhere because of one PB change seemed particularly problematic. But maybe your OVN team will have different opinions/ideas. I don't feel like an expert on the HA stuff, so me not coming up with an idea isn't indicative that there is no idea. :D | 14:21 |
| * otherwiseguy looks | 14:21 | |
| haleyb | otherwiseguy: we'll have to ask them, have a face-2-face soon | 14:23 |
| otherwiseguy | haleyb: I've gotta find a way to get our agent liveness under control. with ovn-monitor-all=True, bumping nb_cfg ends up doing compute^2 ovsdb notifications. It makes a reasonable agent_down_time at any scale pretty much impossible. | 14:26 |
| otherwiseguy | haleyb: the agent db stuff I'm working on can take care of the controller side (it's just that the patch that added watching Chassis_Private for live migration info broke the second patch's ability to only register Chassis_Private in one worker). | 14:28 |
| otherwiseguy | But I think OVN improvements are going to be necessary to make nb_cfg bumping scalable with ovn-monitor-all=true | 14:28 |
| otherwiseguy | s/ovn/OVS/ | 14:28 |
| haleyb | otherwiseguy: oh, compute^2 is not good. And i'll review https://review.opendev.org/c/openstack/neutron/+/818850 and friends whenver they are ready :) | 14:30 |
| otherwiseguy | I always incorrectly assumed that Chassis_Private was excluded from ovn-monitor-all, but it wasn't. They tried https://mail.openvswitch.org/pipermail/ovs-dev/2025-February/421142.html in 2025 and had to revert. | 14:30 |
| ralonsoh | otherwiseguy, should we add a short topic in the PTG for https://review.opendev.org/c/openstack/neutron/+/818850? | 14:31 |
| otherwiseguy | ralonsoh: it'd probably be a good idea. | 14:31 |
| ralonsoh | ok, I'll add it right now | 14:32 |
| otherwiseguy | ralonsoh++ | 14:32 |
| otherwiseguy | That first patch can stand on its own w/o Chassis_Private improvements. The follow-up, I have some work in progress to move to a separate binary instead of using MaintenanceWorker, e.g. agent-health-monitor. That way it could be managed by an external orchestrator, scaled separately etc. | 14:34 |
| otherwiseguy | That second patch as-is screws around too much with MainteananceWorker's ovsdb connection, etc. | 14:35 |
| otherwiseguy | haleyb: https://review.opendev.org/c/openstack/neutron/+/1008752 will take me a while to get through. I will say that with our own deployment tools I've found that we SIGTERM ovn-controller and that in general that is bad and that we should do everything we can to gracefully stop those, e.g. PreStop ovn-ctl stop_controller or ovn-appctl ovn-appctl exit etc. | 14:39 |
| otherwiseguy | Regardless if the patch is a good idea, I'm excited that someone found my snuck in Transaction.add_op() in python-ovs :) | 14:39 |
| otherwiseguy | as an aside, any place we call ovsdb-client to avoid having to register a table, e.g. MAC_Bindings, we should be able to use that add_op() to re-use our existing connection. | 14:59 |
| opendevreview | Quentin GROLLEAU proposed openstack/os-ken stable/2026.1: Fix deadlock in OfctlService when switch reconnects during message retry https://review.opendev.org/c/openstack/os-ken/+/1009293 | 15:35 |
| opendevreview | Quentin GROLLEAU proposed openstack/neutron stable/2026.1: Re-install extension flows lost during ovs restart https://review.opendev.org/c/openstack/neutron/+/1009294 | 15:35 |
| opendevreview | Quentin GROLLEAU proposed openstack/neutron-lib stable/2026.1: Extend L2AgentExtension API: handle_switch_restart https://review.opendev.org/c/openstack/neutron-lib/+/1009292 | 15:35 |
| opendevreview | Lajos Katona proposed openstack/python-neutronclient master: Delete network log and loggable resources https://review.opendev.org/c/openstack/python-neutronclient/+/1009298 | 16:05 |
| opendevreview | Brian Haley proposed openstack/neutron-fwaas master: DNM: Test CI https://review.opendev.org/c/openstack/neutron-fwaas/+/1009307 | 16:22 |
| opendevreview | Maor Blaustein proposed x/whitebox-neutron-tempest-plugin master: Add EVPN-BGP FRR gateway HA failover test https://review.opendev.org/c/x/whitebox-neutron-tempest-plugin/+/1009097 | 16:56 |
| opendevreview | Brian Haley proposed openstack/neutron-fwaas master: DNM: Test CI https://review.opendev.org/c/openstack/neutron-fwaas/+/1009307 | 17:10 |
| cardoe | otherwiseguy: for OVN-ctl stop_controller is there an amount of time I should wait? Asking for OpenStack-helm. I’m on mobile so apologies for the autocorrect. | 17:50 |
| opendevreview | Brian Haley proposed openstack/neutron-fwaas master: DNM: Test CI https://review.opendev.org/c/openstack/neutron-fwaas/+/1009307 | 18:21 |
| otherwiseguy | cardoe: I just started looking at it, mostly in the context of looking for a solution to the live migration patch that uses Chassis_Private info to signal nova that the network is ready. To truly know that *everything* can reach a live migrated VM, you need to know a lot more than just the target host is ready--really everything that can possibly reach that node needs to be wired. In an ideal world, | 18:39 |
| otherwiseguy | NB_Global.hv_cfg is "every controller has been updated with the proposed network", but the frequency of SIGTERM leaving Chassis "up" that don't get updated means its fairly likely that random chassis that don't matter could be down so hv_cfg isn't updated, etc. | 18:39 |
| otherwiseguy | I haven't run across scenarios where ovn-controller hangs trying to shut down, but I also haven't tried to find a lot. | 18:40 |
| otherwiseguy | tl;dr I don't know what a good timeout is :D | 18:40 |
| otherwiseguy | cardoe: and I should point out that as it is something I just started looking at, I don't yet know if I've missed any side effects with the chassis being cleaned up. (if there are, they are likely bugs on our part) | 18:45 |
| cardoe | Alright. Well if you find something I’ll be happy to get it fixed in Kolla and OSH so we don’t get bad bug reports in neutron. | 18:52 |
| opendevreview | Brian Haley proposed openstack/neutron-fwaas master: DNM: Test CI https://review.opendev.org/c/openstack/neutron-fwaas/+/1009307 | 19:48 |
| opendevreview | Brian Haley proposed openstack/neutron master: Revert "iptables: Rely on iptables-restore -w lock instead of oslo lock" https://review.opendev.org/c/openstack/neutron/+/1009356 | 20:09 |
| opendevreview | Brian Haley proposed openstack/neutron-fwaas master: DNM: Test CI https://review.opendev.org/c/openstack/neutron-fwaas/+/1009307 | 20:09 |
| opendevreview | Brian Haley proposed openstack/neutron master: Revert "iptables: Rely on iptables-restore -w lock instead of oslo lock" https://review.opendev.org/c/openstack/neutron/+/1009356 | 20:15 |
| *** priteau_ is now known as priteau | 21:38 | |
| opendevreview | Brian Haley proposed openstack/neutron-fwaas master: Fix neutron-fwaas-v2-dsvm-tempest-multinode CI failure https://review.opendev.org/c/openstack/neutron-fwaas/+/1009380 | 21:43 |
| opendevreview | Merged openstack/python-neutronclient master: Delete onboard_network_subnets https://review.opendev.org/c/openstack/python-neutronclient/+/995993 | 21:44 |
| haleyb | otherwiseguy: oh, compute^2 is not good. And i'll review https://review.opendev.org/c/openstack/neutron/+/818850 and friends whenver they are ready :) | 22:02 |
| opendevreview | Brian Haley proposed openstack/neutron-fwaas master: Fix neutron-fwaas-v2-dsvm-tempest-multinode CI failure https://review.opendev.org/c/openstack/neutron-fwaas/+/1009380 | 22:49 |
| opendevreview | Rocky proposed openstack/neutron master: Emit vif-plugged on OVN port re-bind during rebuild https://review.opendev.org/c/openstack/neutron/+/994438 | 23:25 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!