| mnaser | tkajinam: fyi this may interest you https://bugs.launchpad.net/neutron/+bug/2099450 :) | 03:08 |
|---|---|---|
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Fix wrong rejex for [ovs] ovsdb_connection https://review.opendev.org/c/openstack/neutron/+/942412 | 05:21 |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Fix wrong rejex for [ovs] ovsdb_connection https://review.opendev.org/c/openstack/neutron/+/942412 | 05:22 |
| tkajinam | mnaser, ^^^ | 05:24 |
| tkajinam | I really don't know why these defaults do not agree by the way | 05:24 |
| mnaser | Me either. It seems like they are duplicated, but this should do the trick for now (tm) | 05:41 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: osken: bump version to 3.0.0 https://review.opendev.org/c/openstack/neutron/+/942420 | 08:17 |
| opendevreview | Tobias Urdin proposed openstack/neutron master: wip: Use quota driver from allowed address pairs https://review.opendev.org/c/openstack/neutron/+/942361 | 08:56 |
| sahid | lajoskatona: o/ | 08:59 |
| sahid | I was looking into your comment https://review.opendev.org/c/openstack/neutron/+/939117/4/neutron/privileged/agent/ovsdb/native/helpers.py | 08:59 |
| sahid | I think you made a good point, removing and creating it again is perhaps not good | 08:59 |
| sahid | that is said we are normally at time where wr launch the agent and we can consider that a full resync will happen | 09:01 |
| lajoskatona | sahid: Hi | 09:32 |
| lajoskatona | sahid: yeah, it sound safer to keep it and update if something is missing or not what is in the neutron cfg file values | 09:34 |
| sahid | lajoskatona: sure | 09:52 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: log warning when ovs is reported dead https://review.opendev.org/c/openstack/neutron/+/920753 | 09:53 |
| zigo | Hi there! | 10:05 |
| zigo | We're having some issue with some Windows VMs. They use the virtio-net driver. When I ping something outside the VM, I'm seeing some packet loss between the interface within the VM and its associated TAP. | 10:05 |
| zigo | Is there a way to select smoething else than virtio-net for windows VMs ? | 10:05 |
| zigo | I hate trying debugging MS crap, seriously ... :/ | 10:06 |
| priteau | zigo: you have various options available via hw_vif_model depending on your hypervisor, see https://docs.openstack.org/glance/latest/admin/useful-image-properties.html | 10:20 |
| zigo | Thanks, having a look. | 10:21 |
| priteau | I don't work with Windows VMs, but I would assume most of the emulated NICs would work, although some might require installing extra drivers | 10:22 |
| zigo | Ok, I see: hw_vif_model=<something>, with something being e1000, e1000e, ne2k_pci, pcnet, rtl8139, virtio, vmxnet3 and igb... | 10:24 |
| zigo | Probably defaulting to virtio, I guess. | 10:25 |
| priteau | I think it depends on the other metadata you have on the image. Nova uses libosinfo to pick the best hardware type for your selected operating system. | 10:39 |
| sahid | o/ neutron team | 11:45 |
| sahid | any chance to have this easy one? https://review.opendev.org/c/openstack/neutron/+/942420 | 11:45 |
| tore | When DVR is enabled, and a tenant network port is both assigned a floating IP, and is a member of an Octavia load balancer, traffic to the FIP ends up centralised (going through the CR-LRP port). is it possible to detect this situation from looking in the MySQL database? I end up programming the wrong MAC address in my EVPN agent for this corner case | 11:57 |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Fix wrong rejex for [ovs] ovsdb_connection https://review.opendev.org/c/openstack/neutron/+/942412 | 13:24 |
| lajoskatona | sahid: done | 13:51 |
| lucadelmonte | hello, not sure if this is the right channel for this question, i have an openstack 2024.2 cluster deployed with kolla, i am using ovn and genericswitch mech driver for ironic BM provisioning, i have an issue with neutron vif created by ironic, basically neutron-server reports the port as down and then gets plugged in again in an endless loop, https://pastebin.com/XQG5dP8p, do you have any advice on how to fix | 14:02 |
| sahid | lajoskatona: thank you ! | 14:24 |
| opendevreview | Lajos Katona proposed openstack/neutron master: [eventlet-deprecation] Remove eventlet from DHCP agent https://review.opendev.org/c/openstack/neutron/+/942393 | 15:06 |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Use openvswitch interface driver by default https://review.opendev.org/c/openstack/neutron/+/942161 | 15:30 |
| shermanm | lucadelmonte: admittedly I haven't tried ovn+genericswitch, but what do you have for `mechanism_drivers` in your ml2_conf.ini? the order there matters, and I've seen similar issues in the past. | 15:45 |
| shermanm | with ovs (not ovn), ours looks like: `mechanism_drivers = openvswitch,genericswitch,baremetal` | 15:45 |
| frickler | tempest with ovn startup failing, haven't seen that before, I assume it is just a one-off, but maybe someone wants to take a look anyway https://zuul.opendev.org/t/openstack/build/fe0b8cd7c6e445c3896be315823bd177 | 15:52 |
| lucadelmonte | shermanm: the order is mechanism_drivers = baremetal,ovn,genericswitch, but i was trying to debug until now and i see that the event which is pluggin the port is fired from PortBindingChassisUpdateEvent, not sure if i have some misconfiguration on the HA of ovn chassis | 16:10 |
| lucadelmonte | _shermanm no luck when changing the order, same "loop" | 16:16 |
| shermanm | lucadelmonte: that was unfortunately my only "bright idea", especially since ngs is reporting that it bound the port successfully, then neutron is trying again anyway | 17:45 |
| opendevreview | Mohammed Naser proposed openstack/neutron master: Add MTU for all non-external ports https://review.opendev.org/c/openstack/neutron/+/942489 | 20:22 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!