13:00:21 <haleyb> #startmeeting networking 13:00:21 <opendevmeet> Meeting started Tue Oct 7 13:00:21 2025 UTC and is due to finish in 60 minutes. The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:21 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:21 <opendevmeet> The meeting name has been set to 'networking' 13:00:23 <haleyb> Ping list: bcafarel, elvira, frickler, mlavalle, mtomaska, slaweq, tobias-urdin, ykarel, lajoskatona, jlibosva, averdagu, haleyb, ralonsoh, sahid 13:00:27 <haleyb> hi everyone o/ 13:00:37 <bcafarel> o/ 13:00:38 <mtomaska> o/ 13:00:38 <elvira> o/ 13:00:44 <cbuggy> o/ 13:00:50 <ykarel> o/ 13:01:03 <mlavalle> \o 13:01:07 <slaweq> o/ 13:01:35 <haleyb> #announcements 13:01:48 <ralonsoh> hello 13:01:59 <haleyb> We are officially done with Flamingo and in Gazpacho cycle 13:02:19 <haleyb> congrats to getting another release done and thanks for all the hard work! 13:03:06 <haleyb> The next OpenInfra PTG will take place October 27-31, 2025 and registration for the event is open 13:03:13 <haleyb> #link https://ptg.openinfra.dev/ 13:03:21 <haleyb> please sign-up if you haven't 13:03:31 <haleyb> also please start thinking of any topics you want to discuss 13:03:39 <haleyb> #link https://etherpad.opendev.org/p/oct2025-ptg-neutron 13:04:38 <haleyb> Reminder: If you have a topic for the drivers meeting on Friday, please add it to the wiki @ https://wiki.openstack.org/wiki/Meetings/NeutronDrivers 13:04:39 <lajoskatona> o/ 13:05:44 <haleyb> The other thing I wanted to propose was that since we just released Flamingo, we should tag stable/2025.1, 2024.2 and 2024.1 soon 13:05:51 <bcafarel> +1 13:05:57 <haleyb> i at least know our downstream would appreciate it 13:06:24 <haleyb> i think there might only be one 2024.1 change in flight, but if you have others please propose them this week and ping me 13:07:06 <haleyb> #link https://review.opendev.org/c/openstack/neutron/+/962862 is the only one i see and it's in the gate :) 13:07:16 <haleyb> thanks bcafarel and lajoskatona for the reviews 13:07:21 <bcafarel> reviewed just in time for the meeting :) 13:07:59 <haleyb> And continue to use the priorities dashboard for important changes or things ready to merge 13:08:06 <haleyb> i see 7 changes on the list 13:08:45 <haleyb> that was all the announcements i had, any others? 13:09:30 <haleyb> ok, moving on then 13:09:43 <haleyb> #topic bugs 13:09:55 <haleyb> lajoskatona was the bug deputy last week, his report is at 13:10:01 <haleyb> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/Z2TYE772QP5LYRIKETKEP3XL3LP6LHHH/ 13:10:02 <lajoskatona> The report : https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/Z2TYE772QP5LYRIKETKEP3XL3LP6LHHH/ 13:10:07 <lajoskatona> yes that one :-) 13:10:07 <haleyb> it was a busy week 13:10:17 <lajoskatona> yes, my bad luck :-) 13:10:36 <lajoskatona> Most of the bugs has owner, so not that bad 13:11:08 <lajoskatona> One High prio bug without assignee: https://bugs.launchpad.net/neutron/+bug/2126032 13:11:22 <lajoskatona> [CI] Random failure network create with Error while executing command: HttpException: 503, Unable to create the network. No tenant network is available for allocation. 13:11:44 <haleyb> 2025.1 only? 13:11:49 <lajoskatona> As I remember we shortly discussed it yesterday on CI meeting 13:11:50 <ykarel> yes this was discussed in CI meeting, ralonsoh will be taking a look there 13:12:09 <lajoskatona> yes but ykarel said some happens on master also 13:12:25 <lajoskatona> but mostly on stable branch yes 13:12:42 <ykarel> yes seen in both master(14%) and 2025.1(86%) as per opensearch 13:13:47 <haleyb> alright, hopefully we can figure out 13:14:09 <lajoskatona> ralonsoh suggested that perhaps related to https://review.opendev.org/c/openstack/neutron/+/948200 and we have to backport it 13:14:24 <ralonsoh> I'm still investigating 13:14:35 <ralonsoh> today I didn't have time for this, I'll check it this week for sure 13:14:47 <lajoskatona> for stable of course, and for master as I remember there was a pending patch from Liu perhaps 13:15:16 <ralonsoh> this patch is merged in 2026.1, I updated and rebased it 13:15:32 <ralonsoh> https://review.opendev.org/c/openstack/neutron/+/947898 13:15:41 <lajoskatona> ahh, ok, I just found it, ok 13:16:15 <ralonsoh> I think (maybe) that this could be an issue between workers: worker 0 is in charge of updating this table 13:16:25 <lajoskatona> ok so anyway it is not totally out of control 13:16:30 <ralonsoh> but worker 1+ must read it too, maybe too soon 13:16:37 <haleyb> does it hurt to backport the first change? 13:16:52 <ralonsoh> it's backported up to 2025.2 13:17:10 <ralonsoh> sorry, 2025.1 13:17:17 <ralonsoh> https://review.opendev.org/q/Ibc42f900214e1f7631e266bccd083a2ef4111585 13:17:42 <ralonsoh> give me this week to investigate 13:17:52 <haleyb> oh yeah, didn't notice 13:17:58 <haleyb> np, thanks for looking 13:18:59 <haleyb> next unassigned one is 13:19:03 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2126637 13:19:10 <haleyb> [OVN] Router "after_update" event contains wrong information with GW multihoming API 13:19:37 <ralonsoh> easy to reproduce and to debug, I have no time to check it now 13:20:04 <haleyb> right, if anyone has cycles please take a look 13:20:42 <haleyb> next one 13:20:46 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2126685 13:20:52 <haleyb> test_add_external_port_avoid_flapping functional test failure 13:21:54 <lajoskatona> this one was opened by jlibosva 13:22:07 <lajoskatona> by opensearch not frequent 13:22:25 <ralonsoh> hmmm we are still checking ('up', '=', False), instead of up!=True 13:22:42 <ralonsoh> "up" can be an empty list 13:24:50 <haleyb> ralonsoh: so is that a possibly easy fix? 13:25:08 <ralonsoh> maybe, just a suggestion to check 13:25:13 <ralonsoh> I have no time this week 13:26:38 <haleyb> ack i'll see if i have any time 13:26:59 <haleyb> i think that was it for unassigned 13:27:08 <lajoskatona> And we have 2 for eventlet, as I see I forgot to mark these as unassigned.... 13:27:22 <lajoskatona> https://bugs.launchpad.net/neutron/+bug/2126790 13:27:28 <lajoskatona> https://bugs.launchpad.net/neutron/+bug/2126791 13:27:31 <haleyb> lajoskatona: right, i saw you took them both ? 13:27:38 <lajoskatona> one for dhcp-agent and one for ovs-agent 13:27:53 <lajoskatona> I started to investigate the dhcp-agent bug 13:28:23 <lajoskatona> but not the ovs-agent :-) 13:28:49 <haleyb> oh, are they most likely related? 13:29:28 <ralonsoh> At least in the DHCP could be related to the thread pool 13:29:44 <lajoskatona> The pattern should be similar I suppose, ralonsoh suggested that the way we use threadpoolexecutor should be wrong, and I suppose the way we start threads .... 13:29:45 <ralonsoh> we have the same in L3, but in this case there is a check to avoid spawning new threads 13:29:54 <ralonsoh> def _process_routers_loop(self): 13:29:54 <ralonsoh> LOG.debug("Starting _process_routers_loop") 13:29:54 <ralonsoh> while not self._exiting: 13:29:54 <ralonsoh> self._pool.submit(self._process_update) 13:30:05 <ralonsoh> we have the " self._exiting" in the L3 agent 13:30:09 <ralonsoh> this is not present in the DHCP 13:30:17 <ralonsoh> this is the first thing to check and try 13:30:39 <ralonsoh> this is what we have in the DHCP agent 13:30:40 <ralonsoh> def _process_loop(self): 13:30:40 <ralonsoh> LOG.debug("Starting _process_loop") 13:30:40 <ralonsoh> while True: 13:30:40 <ralonsoh> self._pool.submit(self._process_resource_update) 13:30:45 <ralonsoh> (just a while true) 13:30:57 <lajoskatona> +1 13:31:13 <lajoskatona> yes I started to list these things and set up env for it 13:31:21 <haleyb> thanks for the advice ralonsoh 13:32:04 <haleyb> ok, is that all the bugs? 13:32:26 <lajoskatona> yes I think we covered them 13:32:36 <lajoskatona> the others are progressing or has owner 13:32:54 <haleyb> this week ralonsoh is the deputy, next week is sahid 13:32:58 <ralonsoh> ok 13:33:29 <haleyb> i do not see sahid here, will ping offline 13:33:52 <haleyb> moving on 13:34:01 <haleyb> #topic specs 13:34:02 <haleyb> #link https://review.opendev.org/q/project:openstack/neutron-specs+status:open 13:34:14 <haleyb> just a reminder there might be specs in need of review 13:34:43 <haleyb> when i start doing PTG planning i will go through that list as it has not moved much 13:35:16 <haleyb> #topic community goals 13:35:31 <haleyb> lajoskatona: any neutronclient updates? 13:35:36 <lajoskatona> few 13:35:55 <lajoskatona> for horizon the hanging patches can now be reviewed as we started the new cycle 13:36:18 <lajoskatona> https://review.opendev.org/q/topic:%22bug/1999774%22+status:open 13:36:41 <lajoskatona> 2 of them is in stadium dashboards 13:36:51 <lajoskatona> Otherwise I spent some time with nova 13:37:11 <lajoskatona> https://review.opendev.org/q/topic:%22sdk_for_neutron%22 13:37:31 <lajoskatona> and I lost with the FIP patch 13:37:58 <lajoskatona> as there is stille the "proxy API" in nova that actually calls neutron API 13:38:30 <lajoskatona> so if you call nova/os-floating-ips (Ithink this is the endpoint) nova will call neutron/v2/floating-ips 13:38:56 <lajoskatona> and there I have to rethink how to handle different projects and users etc 13:39:05 <lajoskatona> but no time this week for that 13:39:12 <lajoskatona> that's it for neutronclient 13:39:33 <haleyb> and they have no plan to deprecate that API to just have people use neutron API? 13:40:28 <lajoskatona> that should be a question even during the PTG, but I fear that is a kind of Pandora's box 13:41:11 <haleyb> nova-network is long dead :) 13:41:13 <lajoskatona> I am sure that life would be easier for everybody without these ancient proxy APIs but I am not familiar with how nova has to handle these APIsi 13:41:23 <lajoskatona> yes, exactly 13:42:17 <lajoskatona> I run a round with nova team to make them think about this (perhaps today nova meeting i s perfect....) 13:42:39 <haleyb> i put in ptg agenda 13:42:47 <lajoskatona> +1 13:42:59 <lajoskatona> will we have x-project with Nova? 13:43:26 <opendevreview> Ashish Gupta proposed openstack/neutron-tempest-plugin master: Testing https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/963287 13:43:27 <haleyb> i believe so, think i'm behind on emails 13:43:53 <haleyb> ok, and finally eventlet 13:44:21 <ralonsoh> finding the patch... 13:44:25 <haleyb> https://review.opendev.org/c/openstack/neutron/+/952258 13:44:33 <ralonsoh> yes, this one 13:44:49 <ralonsoh> please, review it, we need that to "almost" remove eventlet 13:44:58 <ralonsoh> there are some small bits, but minimal 13:45:04 <haleyb> it looks like just CI issues perhaps 13:45:30 <ralonsoh> yes, that will introduce CI instabilities, in particular the FT one 13:46:21 <haleyb> so please review that ^^ change 13:47:04 <haleyb> #topic on-demand 13:47:13 <haleyb> ralonsoh: i see you added a topic last week? 13:47:35 <ralonsoh> MariaDB 11.8 collation default changes. 13:47:42 <ralonsoh> There is a mail thread about this 13:47:49 <ralonsoh> and there is also a etherpad 13:47:54 <ralonsoh> https://etherpad.opendev.org/p/gazpacho-collations-charsets 13:47:59 * mlavalle added the patch to pile 13:48:08 <ralonsoh> I suggest to open a LP bug to investigate it 13:48:21 <haleyb> who is running that work? 13:48:24 <ralonsoh> I also suggest to have a Neutron CI job forcing this MariaDB version 13:48:38 <ralonsoh> maybe better to ask this in openstack-tc 13:49:52 <ralonsoh> we had something related last cycle with https://review.opendev.org/c/openstack/neutron/+/952819 13:50:32 <ralonsoh> in this case, what mariadb 11.8 is forcing is all collations and charset case sensitive 13:50:41 <ralonsoh> that could have some unexpected side effects 13:51:31 <ralonsoh> in any case, these side effects can take time to emerge, like https://bugs.launchpad.net/neutron/+bug/2114819 (a customer using case sensitive tags) 13:52:18 <haleyb> right. i'll see if i can find if this will be discussed at ptg 13:53:15 <haleyb> any other things to discuss? 13:53:33 <lajoskatona> I have a quick question 13:53:50 <lajoskatona> Does anybody go to the Paris Summit next weekend? 13:54:09 <ralonsoh> nope 13:54:09 <mlavalle> I wish I would 13:54:24 <haleyb> not me 13:55:05 <lajoskatona> as a suprise one level of management suggested for us to go and approved the travel, and I hoped that I can meet some Neutron folks there :-) 13:56:10 <lajoskatona> that's it 13:57:11 <haleyb> enjoy Paris! :) 13:57:32 <lajoskatona> :-) 13:57:34 <haleyb> alright, that's all i had, have a good week everyone 13:57:39 <haleyb> #endmeeting