| *** liuxie is now known as liushy | 07:24 | |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Make resource "tag" case sensitive https://review.opendev.org/c/openstack/neutron/+/952819 | 09:04 |
|---|---|---|
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [eventlet-removal] Remove the usage of eventlet from the periodic worker https://review.opendev.org/c/openstack/neutron/+/952117 | 09:06 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] Use stateless NAT rules for FIPs https://review.opendev.org/c/openstack/neutron/+/951511 | 09:53 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Add non-repair mode to the existing method in OVN sync class https://review.opendev.org/c/openstack/neutron/+/953081 | 09:53 |
| opendevreview | Lajos Katona proposed openstack/neutron master: Delete tunnel endpoints if endpoint agent is deleted https://review.opendev.org/c/openstack/neutron/+/952988 | 10:00 |
| opendevreview | Dai Dang Van proposed openstack/neutron master: Add dns forwarder l2 extension https://review.opendev.org/c/openstack/neutron/+/951390 | 10:27 |
| LarsErikP | Hi, does anyone here know who to ping about actually getting this released from proposed to updates in UCA? https://bugs.launchpad.net/ubuntu/+source/neutron-dynamic-routing/+bug/2090992 | 10:40 |
| LarsErikP | I guess there is some canonincal-people that needs to do something here? | 10:40 |
| slaweq | LarsErikP: maybe ask haleyb, he is from Canonical IIRC | 11:15 |
| LarsErikP | tyty! | 11:17 |
| LarsErikP | haleyb|out: ? ^ :P | 11:18 |
| lajoskatona | slaweq: Hi, welcome back :-) | 11:49 |
| lajoskatona | slaweq: as you already jumped to sRBAC, I have another related topic for which your expertise should be useful: https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/948465 | 11:50 |
| lajoskatona | slaweq: I had a chat last week with frickler (but see his comments in the patch also): https://meetings.opendev.org/irclogs/%23openstack-neutron/%23openstack-neutron.2025-06-19.log.html#openstack-neutron.2025-06-19.log.html#t2025-06-19T12:24:24 | 11:50 |
| lajoskatona | slaweq: his main concern is that n-d-r API is admin only and by adding sRBAC that can chang this behaviour | 11:51 |
| slaweq | hi lajoskatona I think that the whole idea behind S-RBAC is to make it possible to define some rules for the operators and not hard code things | 12:08 |
| slaweq | so I don't really get the point of the "admin only " APIs and why that would prevent migrating to the new policies | 12:09 |
| slaweq | lajoskatona I commented in the patch too | 12:13 |
| ierdem | Hello, we have a kolla-ansible yoga environment with 9 controller nodes and more than 1000 compute nodes. When I upgrade neutron and opensvwitch to a newer version (doesn't matter which, in my case Antelope) access to all virtual machines is lost, but if there are existing ping flows, new ones can not created. | 12:14 |
| ierdem | This continues until openvswitch deploy is completed, i.e. neutron_openvswitch restart. In some cases, the access outage can be prolonged. When I send an upgrade to neutron alone, I don't experience any outage. How can I prevent this, is there a way to send a hot-upgrade without any outage? Or is it normal to have an outage? | 12:14 |
| lajoskatona | slaweq: thanks, I check it | 12:16 |
| ykarel | Hi i have a clash with internal meeting so couldn't chair CI meeting today, somehow i mistook the timing | 12:58 |
| ykarel | bcafarel, lajoskatona, slawek, mlavalle, mtomaska, ralonsoh, ykarel, jlibosva, elvira ^ | 12:58 |
| ykarel | wrt failures below are some new ones | 13:01 |
| ykarel | unit test test_spawn_rate_limited_metadata_proxy | 13:02 |
| ykarel | - https://025b8612ea74b6df403c-82c7c26c1c9d82596e9efade16533251.ssl.cf1.rackcdn.com/openstack/8ae8a7c942354bb780425a17cd25e199/testr_results.html | 13:02 |
| ykarel | - https://9f4bad0085cac2a65c00-b3e1428302d70b58654d7733ebfb21f6.ssl.cf2.rackcdn.com/openstack/7cc99183a92c44b0b05d7ea7215c3146/testr_results.html | 13:02 |
| ykarel | - related to https://review.opendev.org/c/openstack/neutron/+/952220 | 13:02 |
| ykarel | functional fails | 13:02 |
| ykarel | - test_floatingip_mac_bindings | 13:02 |
| ykarel | - https://4e87414b0fe3ca1de111-0dd3b493df614cc7e41a46754664ea83.ssl.cf5.rackcdn.com/openstack/a858fb193a21437987d557b6de06dbdc/testr_results.html | 13:02 |
| ykarel | - Timeout in 15 sec, but it took ~ 17 seconds | 13:02 |
| ykarel | - test_fip_connection_for_address_scope | 13:02 |
| ykarel | - https://launchpad.net/bugs/2115026 | 13:02 |
| ykarel | will report issues tomorrow for these if someone don't beat me to that | 13:03 |
| ykarel | and will send cancellation mail for today, sorry for last minute change | 13:03 |
| lajoskatona | ykarel: ack, thanks for headsup | 13:03 |
| slaweq | ykarel ack and thx for the heads up | 13:05 |
| *** haleyb|out is now known as haleyb | 13:12 | |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Use NetworkMTUSubnetConflict from neutron-lib https://review.opendev.org/c/openstack/neutron/+/953095 | 13:16 |
| opendevreview | Takashi Kajinami proposed openstack/ovn-octavia-provider master: Replace LOG.warn https://review.opendev.org/c/openstack/ovn-octavia-provider/+/953055 | 13:17 |
| haleyb | LarsErik1: the movement of something from -proposed to -updates is supposed to be just time-based, usually around 2 weeks or so. Is it version 2:24.0.0-0ubuntu1.1~cloud0 you are asking about? | 13:20 |
| LarsErik1 | haleyb: yeah, it's that version | 13:20 |
| LarsErik1 | I would argue that 2week has now passed since 29.05 though :P | 13:21 |
| haleyb | LarsErik1: you are in luck since i have a meeting with that team in 2 hours, i will put it on the agenda since it's related to a case | 13:21 |
| LarsErik1 | oh great! Thanks =) | 13:22 |
| haleyb | LarsErik1: well, ~2 weeks is sometimes more depending on how busy the team is, and i do not have the keys to the castle where these things are | 13:23 |
| LarsErik1 | okok! Thanks anyways =) It was my impression that this should just happen automatically when everything was verfied, but I guess there has to be a human in the loop somewhere? | 13:24 |
| opendevreview | Lajos Katona proposed openstack/neutron master: [eventlet-removal] remove eventlet references from dhcp_agent https://review.opendev.org/c/openstack/neutron/+/950499 | 13:25 |
| haleyb | sometimes not knowing how the sausage is made is a good thing | 13:25 |
| *** LarsErik1 is now known as LarsErikP | 13:28 | |
| LarsErikP | haleyb: hahahahaa | 13:29 |
| LarsErikP | true that :P | 13:29 |
| LarsErikP | But it's nice to know someone's actually cooking the sausage ;) | 13:29 |
| haleyb | there are definitely people here making and cooking on the grill | 13:32 |
| LarsErikP | great! Never under-estimate the value of a good barbeque :P | 13:33 |
| ralonsoh | ykarel, I'll check this errors, thanks! | 13:35 |
| opendevreview | Arnau Verdaguer proposed x/whitebox-neutron-tempest-plugin master: [DVR] Add sleep after adding/removing FIPs https://review.opendev.org/c/x/whitebox-neutron-tempest-plugin/+/951111 | 14:06 |
| opendevreview | Merged openstack/neutron stable/2025.1: Consider logging options when using OVNdbsync https://review.opendev.org/c/openstack/neutron/+/952583 | 14:11 |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Remove compat logic for neutron-lib < 3.9.0 https://review.opendev.org/c/openstack/neutron/+/953111 | 14:17 |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Remove compat logic for neutron-lib < 3.9.0 https://review.opendev.org/c/openstack/neutron/+/953111 | 14:48 |
| opendevreview | Takashi Kajinami proposed openstack/neutron master: Remove compat logic for neutron-lib < 3.9.0 https://review.opendev.org/c/openstack/neutron/+/953111 | 14:58 |
| zigo | Hi there! We have busy network nodes, each with 250+ virtual router. When I restart neutron-l3-agent, mostly all routeurs (if not all) are failing-over to the other HA virtual router. Is there a way to avoid this during the l3 restart ? (note: I'm still on Victoria and OVS in this deployment, obivously with HA routers by default). | 20:40 |
| zigo | I'm seeing when I restart the l3-agent: | 20:44 |
| zigo | function update_all_ha_network_port_statuses took 97.292s seconds to run wrapper /usr/lib/python3/dist-packages/oslo_utils/timeutils.py:386 | 20:44 |
| zigo | Does this mean I need to increase agent_down_time to more than this ? It's currently at 75 ... | 20:44 |
| zigo | Cause reading the agent log when it was restarting, I could "feel" that it was stuck, and then saw it as down for a short time in the "openstack network agent list --agent-type l3" | 20:48 |
| zigo | So, my thinking: the l3 agent is then doing a full sync, triggering the failover ? Or is it something else that triggered the failover ? | 20:48 |
| haleyb | zigo: sorry, been a while since i touched the l3-agent, but there is a neutron.conf setting allow_automatic_l3agent_failover ? | 21:41 |
| haleyb | of course according to https://bugs.launchpad.net/neutron/+bug/2050236 that might have been broken for some time | 21:41 |
| zigo | haleyb: Opposite way. :) | 21:44 |
| zigo | In my case, we have an l3 agent with many virtual routers in it. | 21:44 |
| zigo | Somehow, restarting it makes all virtual routers to switch to standby. | 21:45 |
| zigo | I would like to avoid this behavoir, because 1/ this makes an "empty" network node, and 2/ it breaks all ongoing connections. | 21:45 |
| zigo | Note we're not using DVR, because we have all our compute nodes using bgp-to-the-host, which Neutron + OVS doesn't support. | 21:46 |
| zigo | So the issue is only on our network nodes. | 21:46 |
| zigo | FYI, we have 12 network nodes on that cluster. | 21:47 |
| haleyb | zigo: understood. i have a faint memory from my cloud provider days but can't think of what we did (yet) | 21:47 |
| zigo | :) | 21:50 |
| zigo | Ok. | 21:50 |
| haleyb | zigo: so in your case the active router interface moves to the other node? | 21:51 |
| zigo | Right. | 21:51 |
| zigo | All of them, as much as I can tell. | 21:51 |
| zigo | I'll experiment with our preprod cluster to see what happens when I provision 1024 virtual routers there and restart the l3 agent. | 21:51 |
| zigo | I'll see if I can reproduce it, and if increasing agent_down_time helps... | 21:51 |
| * zigo goes to sleep | 21:52 | |
| zigo | Thanks. | 21:52 |
| haleyb | there is an ha_vrrp setting as well, testing all the knobs is the best answer i guess | 21:53 |
| opendevreview | Kevin Carter proposed openstack/neutron master: feat(OVN): add new mode to only add when missing https://review.opendev.org/c/openstack/neutron/+/941824 | 22:56 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!