13:00:20 <haleyb> #startmeeting networking 13:00:20 <opendevmeet> Meeting started Tue Aug 19 13:00:20 2025 UTC and is due to finish in 60 minutes. The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:20 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:00:20 <opendevmeet> The meeting name has been set to 'networking' 13:00:22 <haleyb> Ping list: bcafarel, elvira, frickler, mlavalle, mtomaska, slaweq, tobias-urdin, ykarel, lajoskatona, jlibosva, averdagu, haleyb, ralonsoh 13:00:23 <mlavalle> \o 13:00:25 <haleyb> o/ 13:00:31 <frickler> \o 13:00:45 <bcafarel> o/ 13:00:46 <ykarel> o/ 13:01:24 <haleyb> #announcements 13:01:36 <rubasov> o/ 13:01:36 <haleyb> We are currently in Week R-6 of Flamingo 13:01:51 <haleyb> Our next milestone in this development cycle will be Flamingo-3, week of August 25th (next week) 13:01:51 <slaweq> o/ 13:01:59 <haleyb> Final 2025.2 Flamingo release: October 3rd, 2025 13:02:06 <haleyb> #link https://releases.openstack.org/flamingo/schedule.html 13:02:51 <haleyb> This week is general library freeze, I have seen reviews out already 13:04:05 <haleyb> on that ^^ I did have a question on an os-ken change for eventlet 13:04:17 <haleyb> #link https://review.opendev.org/c/openstack/os-ken/+/938337 13:04:36 <elvira> o/ 13:04:55 <haleyb> it seems we should merge that for this cycle 13:05:52 <haleyb> of course i realize rodolfo is not here, or lajos 13:07:20 <ykarel> rodolfo will be back next week 13:07:50 <rubasov> lajos will be back Thursday 13:07:55 <haleyb> ykarel: right, it's just that os-ken release has been proposed so it really should merge this week 13:08:16 <haleyb> rubasov: ack, i will wait and discuss with Lajos, thanks for the info 13:08:58 <ykarel> i see multiple +2s there, what it was waiting for? 13:09:09 <ykarel> since feb 13:09:55 <haleyb> ykarel: really just based on rodolfo's comment - "I would wait a bit to merge this patch. This can always be selected via env var OSKEN_HUB_TYPE." 13:10:05 <haleyb> but i think it's been "a bit" 13:11:13 <haleyb> ykarel: i guess at worst we have to add an OSKEN_HUB_TYPE=eventlet to anything that gets broken 13:11:45 <ykarel> yes can run some more jobs before landing if we want 13:13:37 <ykarel> will do update test patch for sanity https://review.opendev.org/c/openstack/neutron/+/941508 13:14:32 <haleyb> ykarel: ah, yes let's update that for a check, thanks for finding 13:15:19 <haleyb> ok, back to announcements 13:16:07 <haleyb> so for library releases, neutron-lib will be next week, so please check any reviews you have there 13:17:32 <haleyb> also, i will be out next Wed to Friday, so will not be around to approve anything, the release liaiason will have to do it (slaweq) 13:17:51 <haleyb> i will remind everyone at next weeks meeting 13:18:33 <haleyb> The next OpenInfra PTG will take place October 27-31, 2025 and registration for the event is now open 13:18:34 <haleyb> #link https://ptg.openinfra.dev/ 13:18:43 <haleyb> please sign-up if you haven't already 13:19:24 <haleyb> Nominations for PTL and TC have started, and end tomorrow 13:19:28 <haleyb> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/DOC4GBLKKEE4SUY6NRU4MY5NAN526KP6/ 13:19:56 <haleyb> i did propose myself for PTL (again) so you will be stuck with me for another 6 months 13:20:29 <bcafarel> noooo (</joking> of course) 13:21:16 <haleyb> it is easy to lead such a great team :) 13:21:42 <haleyb> and final announcement 13:21:52 <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 13:22:05 <haleyb> with the RH day this week i will cancel the meeting 13:22:39 <haleyb> that was all the announcements i had, any others? 13:23:15 <haleyb> ok, moving on 13:23:19 <haleyb> #topic bugs 13:23:45 <haleyb> bcafarel was the deputy last week 13:23:52 <haleyb> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/3LE4LHA76MTTRCRZDPV6IRXRKAKTJ4YH/ 13:24:12 <haleyb> there was a fair amount of bugs 13:24:30 <haleyb> first one 13:24:35 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2120723 13:24:35 <opendevreview> yatin proposed openstack/neutron master: dnm: tests neutron master with osken native https://review.opendev.org/c/openstack/neutron/+/941508 13:24:42 <haleyb> [OVN] Metadata agent errors when returning gzip-encoded data 13:25:22 <haleyb> this is a regression in 2025.1, and they were able to test with master and still broken 13:27:39 <haleyb> not sure if anyone has seen something similar, we definitely changed the code last cycle and introduced the bug but it wasn't obvious how to fix 13:28:56 <haleyb> ok, will ping rodolfo about that next week 13:29:28 <haleyb> there were a number of medium bugs that all had patches proposed so we're good there :) 13:30:06 <haleyb> next is regarding vpnaas 13:30:11 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2120316 13:30:19 <haleyb> Dalmatian: Second IPsec site connection on same VPN service/router stays PENDING_CREATE; no agent-side rendering (works on Zed L3 agent) 13:31:01 <haleyb> this was reported on ML as well 13:32:17 <haleyb> not sure who can look at this, maybe Bodo? 13:33:26 <haleyb> i'll ping as well 13:34:16 <haleyb> there were also two RFEs, i'm not sure if the first one is just a bug 13:34:20 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2120732 13:34:27 <haleyb> [RFE] Add metadata caching for immutable values 13:34:52 <haleyb> #link https://review.opendev.org/c/openstack/neutron/+/957197 proposed 13:36:04 <haleyb> do others think adding a cache to metadata is a bad idea? or have any opinions on it being an RFE? 13:36:42 <mlavalle> at least, let's discuss it 13:37:09 <mlavalle> I don't know yet if it is a good or bad idea 13:37:44 <ykarel> +1 13:38:30 <haleyb> yeah, it definitely helps the use case/bug where it's getting repeatedly queried 13:38:32 <mtomaska_> yes I will have to read it in detail.... but doesnt nova already cache metadata? we (neutron) just retrieve it? .... i will have to look 13:40:08 <haleyb> mtomaska_: that's a good question, i don't know 13:40:32 <amorin> hi! I vote +1 for metadata cache 13:40:43 <amorin> IIRC we have a custom downstream patch regarding this 13:41:00 <mtomaska_> i can look into it but not until next week... 13:41:02 <haleyb> the only other place we use cache_method_results is in the policy code 13:41:15 <amorin> it's definitely a good idea, I will talk about it internally see if one of the team member can share 13:41:20 <haleyb> amorin: ack, is it similar to https://review.opendev.org/c/openstack/neutron/+/957197 ? 13:41:22 <mtomaska_> What is this RFE going to solve? do we have some performance issues? 13:42:29 <haleyb> mtomaska_: it would solve the case where an instance is asking for static info, and reduce the neutron->nova api calls 13:42:34 <amorin> haleyb: will check 13:42:57 <amorin> we also have some sort of rate limiting, to force users slowing down 13:43:12 <haleyb> amorin: ack, thanks, and if you could leave a comment there or the bug would be great 13:43:16 <mtomaska_> But how often does that happen? I thought it is just used when a VM is booting 13:43:34 <haleyb> mtomaska_: this is a poorly behaved set of VMs 13:43:48 <mtomaska_> ACK 13:44:00 <haleyb> i'm trying to remember, do we have rate-limit settings for haproxy? 13:44:40 <haleyb> seems we do 13:44:46 <ykarel> https://bugs.launchpad.net/neutron/+bug/1989199 13:44:56 <haleyb> rate_limit_enabled 13:45:26 <mlavalle> yes we do 13:45:48 <haleyb> i will ask if they've tried that 13:46:36 <mtomaska_> +1 on rate limiting 13:47:28 <amorin> mtomaska_: in our situation, these are VM that are used by k8s, they heavily rely on metadata, and they were ddos-ing it 13:47:41 <mlavalle> https://review.opendev.org/c/openstack/neutron/+/858879 13:48:39 <haleyb> ok, i think we have enough info there to make progress. we can move forward once we have a little more info, i'm not averse to caching but if the rate limit helps... 13:48:52 <amorin> thanks mlavalle, this is the things we pushed :) 13:49:25 <mlavalle> https://docs.openstack.org/neutron/latest/admin/config-metadata-rate-limiting.html 13:49:41 <mlavalle> amorin: yeap 13:49:56 <haleyb> amorin: great, and i'll keep an eye on things for comments on the caching proposal, thanks! 13:50:05 <mtomaska_> amorin: reading your situation I wonder if you are going to be affected by this https://review.opendev.org/c/openstack/neutron/+/955951 . just fyi 13:50:36 <amorin> thanks, will check 13:50:49 <haleyb> there was one other RFE from the same person 13:50:53 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2120740 13:51:01 <haleyb> [RFE] Set additional search domains via DHCP 13:51:24 <haleyb> i did drop a comment there since we can add this via port-specific options, but only for OVN 13:52:05 <haleyb> it definitely needs more discussion will have to put on the RFE list 13:52:43 <haleyb> mlavalle: the proposed change did scare me regarding dns - https://review.opendev.org/c/openstack/neutron/+/957573 13:54:02 <haleyb> but it could be we need a better way for such options on a subnet or network basis? instead of port-specific - something to think about if people look at the bug 13:54:23 <haleyb> any other bugs to discuss? 13:55:01 <haleyb> oh, and mtomaska_ is the deputy this week, elvira next week - is that good for both? 13:55:18 <mtomaska_> ACK . bcafarel thanks for covering this week for me 13:55:26 <mtomaska_> sorry last week :) 13:55:50 <bcafarel> no problem, I hope you do not get a busy week this week :) 13:56:02 <haleyb> +1 there 13:56:19 <haleyb> #topic community goals 13:56:23 <elvira> it's all good o/ 13:56:35 <haleyb> without lajos or rodolfo here not sure there's any updates on community 13:57:01 <mlavalle> probably not 13:57:03 <haleyb> #link https://review.opendev.org/c/openstack/nova/+/928022 is still a WIP 13:57:17 <haleyb> ok, we can move on then 13:57:29 <haleyb> #topic on-demand 13:57:42 <haleyb> i see nothing on agenda, but the floor is open 13:58:07 <mlavalle> thanks for leading us another cyclw 13:58:24 <bcafarel> +1 13:59:18 <ykarel> +1 13:59:22 <haleyb> mlavalle: np 13:59:47 <haleyb> it's a good team of people so makes it easy :) 14:00:07 <haleyb> ok, one minute left so will end, have a good week everyone! 14:00:11 <haleyb> #endmeeting