| opendevreview | Slawek Kaplonski proposed openstack/neutron master: Don't prevent setting CIDRs as allowed_address_pair for port https://review.opendev.org/c/openstack/neutron/+/968646 | 06:37 |
|---|---|---|
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] Add 'provider:physical_network' field if there is physnet https://review.opendev.org/c/openstack/neutron/+/968769 | 08:52 |
| mnasiadka | ralonsoh: Do you have an idea if there’s a version of https://bugzilla.redhat.com/show_bug.cgi?id=2238969 that has been created in the new Jira tracker? | 09:23 |
| ralonsoh | let me check | 09:24 |
| ralonsoh | mnasiadka, that was reported in https://bugs.launchpad.net/neutron/+bug/2032817 too | 09:27 |
| ralonsoh | and fixed here: | 09:27 |
| ralonsoh | https://review.opendev.org/c/openstack/neutron/+/937026 | 09:27 |
| ralonsoh | https://review.opendev.org/c/openstack/neutron/+/937026/16/releasenotes/notes/ovn_router_gateway_mtu-ac69e07b30093193.yaml | 09:27 |
| ralonsoh | (since 2025.1) | 09:27 |
| mnasiadka | ralonsoh: well, the check_pkt_len stuff in OVS/OVN is not fixing that completely, some of my customers are seeing very low bandwidth when using 9k Geneve internal MTU and regular 1500 on outside - so I’ll probably do some analysis and fire off a ticket to FDP | 09:45 |
| mnasiadka | I understand it’s not Neutron’s fault, just wanted to follow up on something - but I see I need to create a new FDP ticket | 09:46 |
| ralonsoh | mnasiadka, don't you have a case open? | 09:46 |
| mnasiadka | ralonsoh: well, I’m not an RH customer :) so most probably it should be an email to ovs-discuss :) | 09:47 |
| opendevreview | Tobias Urdin proposed openstack/neutron master: Cleanup stale DHCP ports not bound to host https://review.opendev.org/c/openstack/neutron/+/968770 | 09:49 |
| ralonsoh | mnasiadka, do you have the flag ovn_emit_need_to_frag=True? | 09:49 |
| mnasiadka | ralonsoh: yes | 09:53 |
| mnasiadka | Without that it’s even worse :) | 09:53 |
| ralonsoh | so then there is fragmentation but at a cost | 09:54 |
| ralonsoh | maybe that could be improved (somehow) by OVS folks | 09:54 |
| mnasiadka | I’ll run some tcpdumps to see why the bandwidth is so low - like kilobytes per second on a 10G link | 09:56 |
| opendevreview | Dr. Jens Harbott proposed openstack/neutron master: Don't prevent setting CIDRs as allowed_address_pair for port https://review.opendev.org/c/openstack/neutron/+/968646 | 10:37 |
| frickler | ah, the MTU stuff. I think this was only partially fixed indeed | 10:51 |
| frickler | downstream we also use 1500 on the inside for everything because of that | 10:52 |
| mnasiadka | frickler: HPC land doesn’t like 1500 ;-) | 11:02 |
| frickler | at the risk of going offtopic: is the delta really still as important as it was 20 years ago? also, can you do different tenant networks, then? | 11:04 |
| mnasiadka | I don’t know, but surely there is something to it - if it is still THAT important - I have no clue | 11:15 |
| ralonsoh | bcafarel, hello! some trivial backports: https://review.opendev.org/q/topic:%22bug/2132982%22. Thanks in advance | 11:22 |
| *** srelf__ is now known as continuity | 11:31 | |
| mnasiadka | frickler: actually I found https://lore.kernel.org/all/20240516200941.16152-1-aconole@redhat.com/ which is only in RHEL 9.7 (when you use EL or clones) - don’t know if that is in Ubuntu-land in some recent kernel | 11:37 |
| frickler | mnasiadka: interesting patch, looks pretty simple for such a kind of problem. but also this is only for OVS, not for OVN, right? it talks about L3 tunnels, but that's likely different from OVNs gateway functionality? | 12:13 |
| mnasiadka | frickler: well this is openvswitch kmod, ovn programs flows in ovs - so it’s probably related | 12:21 |
| mnasiadka | frickler: I’ll try out Rocky 9.7 when it’s out in coming weeks to see if that gets better | 12:22 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] Add documentation/reno for the OVN external port N/S routing https://review.opendev.org/c/openstack/neutron/+/968797 | 13:58 |
| bcafarel | ralonsoh: sure will look after my last meeting | 14:42 |
| ralonsoh | thanks! | 14:42 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] DNAT registers always have the FIP external MAC https://review.opendev.org/c/openstack/neutron/+/968805 | 14:48 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: iPXE over IPv6 supported since OVN v23.06.0 https://review.opendev.org/c/openstack/neutron/+/968811 | 15:00 |
| opendevreview | Slawek Kaplonski proposed openstack/neutron master: [Docs] Add doc about Virtual IP addresses in Neutron ML2/OVN https://review.opendev.org/c/openstack/neutron/+/968812 | 15:01 |
| slaweq | ralonsoh: ^^ here is docs patch which we were talking about earlier this week | 15:02 |
| ralonsoh | slaweq, thanks! | 15:02 |
| ralonsoh | slaweq, btw, about https://review.opendev.org/c/openstack/neutron/+/968646 (that is almost merged) | 15:03 |
| ralonsoh | checking https://review.opendev.org/q/I9b54e12fbd9b930a79660f2be195641107a5754e | 15:03 |
| ralonsoh | we'll need backports up to 2024.2 | 15:03 |
| ralonsoh | just a heads-up | 15:03 |
| slaweq | yeap, and https://review.opendev.org/c/openstack/neutron/+/968646 too | 15:04 |
| slaweq | :) | 15:04 |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Fix ignored metadata options https://review.opendev.org/c/openstack/neutron/+/968242 | 15:35 |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Fix ignored metadata options https://review.opendev.org/c/openstack/neutron/+/968242 | 15:41 |
| opendevreview | Daniel Failing proposed openstack/neutron master: Improve OVNGatewayLeastLoadedScheduler runtime https://review.opendev.org/c/openstack/neutron/+/967827 | 16:21 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ovn-bgp-agent master: Fix cleanup of rules per evpn device https://review.opendev.org/c/openstack/ovn-bgp-agent/+/927816 | 16:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ovn-bgp-agent master: Fix race condition where an event could unexpose a network during sync https://review.opendev.org/c/openstack/ovn-bgp-agent/+/927791 | 16:50 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ovn-bgp-agent master: Move the check for allowed address scopes to the appropriate place https://review.opendev.org/c/openstack/ovn-bgp-agent/+/943626 | 17:00 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/neutron stable/2025.2: Append content-encoding header https://review.opendev.org/c/openstack/neutron/+/968828 | 17:19 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/neutron stable/2025.1: Append content-encoding header https://review.opendev.org/c/openstack/neutron/+/968829 | 17:20 |
| opendevreview | Merged openstack/neutron stable/2025.2: [OVN] Add 'qos-bw-minimum-ingress' ML2 extension https://review.opendev.org/c/openstack/neutron/+/968451 | 18:00 |
| opendevreview | Merged openstack/neutron stable/2025.1: [OVN] Add 'qos-bw-minimum-ingress' ML2 extension https://review.opendev.org/c/openstack/neutron/+/968452 | 18:00 |
| opendevreview | Merged openstack/neutron stable/2024.2: [OVN] Add 'qos-bw-minimum-ingress' ML2 extension https://review.opendev.org/c/openstack/neutron/+/968453 | 18:00 |
| *** EugenMayer4401801 is now known as EugenMayer440180 | 19:42 | |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!