13:00:23 <haleyb> #startmeeting networking 13:00:23 <opendevmeet> Meeting started Tue Jul 8 13:00:23 2025 UTC and is due to finish in 60 minutes. The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:23 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:23 <opendevmeet> The meeting name has been set to 'networking' 13:00:24 <haleyb> Ping list: bcafarel, elvira, frickler, mlavalle, mtomaska, obondarev, slaweq, tobias-urdin, ykarel, lajoskatona, jlibosva, averdagu, haleyb, ralonsoh 13:00:44 <mlavalle> \o 13:00:52 <bcafarel> o/ 13:01:06 <ykarel> o/ 13:01:32 <obondarev> o/ 13:01:56 <haleyb> alright we can get started, others will probably wander in 13:02:01 <elvira> o/ 13:02:02 <haleyb> #announcements 13:02:16 <haleyb> We are currently in Week R-12 of Flamingo 13:02:31 <haleyb> Our next milestone in this development cycle will be Flamingo-3, week of August 25th 13:02:39 <haleyb> Final 2025.2 Flamingo release: October 3rd, 2025 13:03:21 <haleyb> last week was Flamingo-2 so releases for libraries, etc were done 13:04:11 <haleyb> The next OpenInfra PTG will take place October 27-31, 2025 and registration for the event is now open 13:04:16 <haleyb> https://ptg.openinfra.dev/ 13:04:51 <haleyb> Gazpacho is the next cycle name 13:05:29 <haleyb> #link https://releases.openstack.org/flamingo/schedule.html 13:05:34 <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 13:06:16 <haleyb> that was all the announcements i had, any others? 13:07:30 <mlavalle> Gazpacho, what a tasty name! 13:07:37 <haleyb> Oh, remember to check the priorities dashboard, and use the RP +1/+2, there are a few things there 13:08:19 <haleyb> mlavalle: yes, my mother used to make it when i was a kid 13:09:07 <haleyb> and i always thought it was Italian, but i was clearly wrong as it's Spanish according to mr. google 13:09:59 <elvira> It's my summer daily lunch indeed :) 13:10:13 <mlavalle> it is indeed Spanish 13:10:15 <haleyb> oh nice 13:11:04 <haleyb> i hope it is an enjoyable release (sorry for the dad joke) 13:11:15 <mlavalle> LOL 13:11:38 <mlavalle> (another dad laughing) 13:11:56 <elvira> heheh 13:12:20 <slaweq> :) 13:12:52 <haleyb> alright we can move on 13:12:57 <haleyb> #topic bugs 13:13:13 <haleyb> obondarev was the bug deputy last week, his report is at 13:13:16 <haleyb> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/HZ456UX23OJNK4NTQ46GLLWCMFTSH66J/ 13:13:43 <ralonsoh> hi (late) 13:14:03 <haleyb> ralonsoh: o/ 13:14:10 <haleyb> first bug unassigned is 13:14:26 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2115776 13:14:33 <haleyb> Neutron fails to delete records for a VM or FIP when there is A and AAAA records with the same name 13:14:52 * mlavalle will look at it next week 13:15:39 <haleyb> mlavalle: great, thanks! seems like a dual stack issue we've never encountered 13:15:57 <mlavalle> just assigned it to myself 13:16:23 <haleyb> +1 13:16:32 <haleyb> the only other unassigned is 13:16:36 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2115768 13:16:48 <haleyb> Mellanox Technologies MT2892 Family [ConnectX-6 Dx] SRIOV using Openstack - No IP inside the VM 13:17:05 <ralonsoh> I commented on this one 13:17:30 <ralonsoh> this is a standalone env, and the SRIOV ports DHCP request/reply must go to another node (GW node) 13:17:42 <ralonsoh> where the SRIOV port (external type for OVN) is bound 13:17:57 <ralonsoh> it is not possible to loop back the traffic inside the same node 13:18:16 <haleyb> yes, thanks for that, it does seem like a config issue 13:18:25 <ralonsoh> a deployment issue 13:19:16 <haleyb> we can wait for a response, otherwise nothing to do there 13:19:45 <haleyb> any other bugs to discuss? 13:20:15 <haleyb> this week the deputy is ralonsoh, next week is sahid 13:20:25 <ralonsoh> ack 13:21:50 <haleyb> and i have removed obondarev from the rotation, so the rotation is every 13 weeks/3 months now, which still is not too bad 13:22:43 <haleyb> ok, moving ong 13:22:55 <haleyb> #topic specs 13:22:59 <haleyb> #link https://review.opendev.org/q/project:openstack/neutron-specs+status:open 13:23:10 <haleyb> I just wanted to highlight we have a number of new and old specs 13:23:28 <ralonsoh> pfff too many... 13:23:40 <haleyb> a number of those i had to -1 since they are targeting old releases and need a rebase 13:23:57 <haleyb> some are newer and i just need to set aside a little time to review 13:24:23 <haleyb> i would encourage others to review as well, and for submittors to rebase if they intend to work on them 13:25:42 <haleyb> #topic community goals 13:26:02 <haleyb> i do not see lajos here 13:27:01 <haleyb> but he has made progress, there are only a couple of reviews in flight 13:27:51 <haleyb> and finally the main event(let) 13:28:01 <haleyb> i have now exhausted my dad jokes for the month 13:28:05 <ralonsoh> yes, I've been working on removing eventlet from UTs 13:28:06 <ralonsoh> hehehe 13:28:36 <ralonsoh> I have a patch that is working fine (there are 6 tests that need some review, but no problem) 13:28:40 <ralonsoh> the problem: FTs 13:29:03 <ralonsoh> all related to loopincall, threads in agents and all this stuff is stuck during the execution 13:29:10 <ralonsoh> so the FTs never end 13:29:31 <ralonsoh> I need first to start fixing the OVS related, then the OVN ones (NB, SB dbs, ovn controller) 13:29:38 <ralonsoh> and then the L3 agent 13:29:46 <ralonsoh> I thought that was going to be easier... 13:29:48 <ralonsoh> that's all 13:29:50 <haleyb> so like a missing join() or something? 13:30:07 <ralonsoh> most of the times these threads never end 13:30:15 <ralonsoh> so we need something to finish the inner loop 13:30:29 <ralonsoh> sometimes we have "while true..." 13:30:43 <ralonsoh> and a thread is impossible to kill in python (at least a kernel thread) 13:30:52 <haleyb> ah :( 13:32:00 <haleyb> if you can get one fixed i can help apply it elsewhere 13:32:17 <ralonsoh> yes, I need "the first one" fixed, and then the rest of them 13:32:22 <ralonsoh> (same as I did in UTs) 13:32:39 <haleyb> is there a bug filed? or is it not always happening? 13:32:50 <ralonsoh> one sec (better the patch) 13:33:01 <ralonsoh> https://review.opendev.org/c/openstack/neutron/+/952258 13:33:11 <ralonsoh> so in this patch I'm removing eventlet from UTs 13:33:19 <ralonsoh> as commented, a couple are failing (trivial) 13:33:32 <ralonsoh> but becasue FT and UT FW share some libraries 13:33:39 <ralonsoh> I'm also removing eventlet from FTs 13:33:47 <ralonsoh> and you can check the output in the CI 13:34:24 <ralonsoh> (neutron-tempest-plugin-openvswitch also failing because of some nasty things I need to fix) 13:35:38 <haleyb> i will look at the review and CI there 13:36:51 <haleyb> thanks for the update 13:37:11 <haleyb> #topic on-demand 13:37:29 <haleyb> any other things to discuss? nothing listed in agenda 13:37:41 <mlavalle> I have a question 13:38:12 <haleyb> sure 13:38:28 <mlavalle> Last week I said that I want to help with https://bugs.launchpad.net/neutron/+bug/2115327. At least some of the test will require a whitebox test framework 13:38:52 <mlavalle> Is this what we should use https://opendev.org/x/whitebox-neutron-tempest-plugin? 13:39:32 <ralonsoh> or n-t-p 13:39:44 <ralonsoh> whitebox allows you to, for example, restart the Neutron API 13:39:50 <ralonsoh> this is something no possible in n-t-p 13:40:04 <ralonsoh> so if you need to change the API config (or agent config), use whitebox 13:40:40 <mlavalle> yes, that is what I mean by "some of the tests will require a whitebox test framework" 13:40:56 <haleyb> but it is not run in the check or gate queues from what i can tell 13:41:10 <ralonsoh> I think we have it in periodic 13:41:14 <ralonsoh> if not, I'll add it 13:41:43 <haleyb> ack 13:42:03 <haleyb> and this is the etherpad for the fullstack migration 13:42:03 <ralonsoh> (we don't have it, I'll add it) 13:42:07 <haleyb> #link https://etherpad.opendev.org/p/migrate_Neutron_fullstack_to_Tempest 13:42:29 <mlavalle> then I assume that if a test doesn't require whitebox, we prefer to migrate it to n-t-p, right? 13:42:57 <ralonsoh> yes 13:43:14 <mlavalle> ack, that's all from me 13:44:24 <haleyb> thanks for the discussion 13:44:28 <haleyb> any others? 13:45:28 <haleyb> ok, thanks for attending everyone, and have a good week 13:45:31 <haleyb> #endmeeting