14:00:40 #startmeeting networking 14:00:40 Meeting started Tue Aug 13 14:00:40 2024 UTC and is due to finish in 60 minutes. The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:40 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:40 The meeting name has been set to 'networking' 14:00:48 Ping list: bcafarel, elvira, frickler, mlavalle, mtomaska, obondarev, slaweq, tobias-urdin, ykarel, lajoskatona, jlibosva, averdagu, amotoki, haleyb, ralonsoh 14:00:49 \o 14:00:53 o/ 14:00:54 o/ 14:00:55 hi 14:01:19 o/ 14:01:34 o/ 14:01:40 o/ 14:02:19 i know a couple of people are on pto this week, so we can get started 14:02:21 #topic announcements 14:02:24 o/ 14:02:32 We are now in Dalmatian release week (R -7) 14:03:02 Oslo libraries are entering feature freeze this week 14:03:28 i know elvira had one change that would be good to merge regarding IPv6 addresses 14:04:08 Non-client library freeze: August 22nd, 2024 (R-6 week) - next week 14:04:14 Client library freeze: August 29th, 2024 (R-5 week) 14:04:22 Dalmatian-3 milestone: August 29th, 2024 (R-5 week) 14:05:22 Since we are at the end of the cycle, I would like to start using the priorities dashboard for patches in the "ready to merge" state. This could be older changes as well as new ones. 14:05:45 Please just add them as RP+1, we will keep RP+2 for critical fixes 14:06:17 i need to make a tinyurl for the dashboard, the full link is in the meeting page 14:06:25 #link https://wiki.openstack.org/wiki/Network/Meetings 14:07:17 so if people could take a look at that page once a day (and add things) it should help us close-out some long-running bugs 14:08:13 it seemed better than adding yet another etherpad since it is dynamically updated based on states 14:08:14 +1 14:09:13 any other announcements? 14:09:36 #topic bugs 14:09:48 slaweq was the deputy last week, his report is at 14:09:51 #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/S4TBXEFSCRVSOAAODUIIQTFG7ABCTCDX/ 14:10:09 there was one critical bug 14:10:21 #link https://bugs.launchpad.net/neutron/+bug/2076430 14:10:43 but it seems more we are the victim, fix is in oslo.serialization 14:11:05 #link https://review.opendev.org/c/openstack/oslo.serialization/+/926172 14:11:13 and i see it just got pushed into the gate 14:11:26 ykarel: that did seem to fix the issue, correct? 14:11:34 https://review.opendev.org/c/openstack/oslo.serialization/+/926172 14:11:38 this one I think 14:11:51 yes, sorry I missed the link 14:11:53 haleyb, yes correct that fixes it 14:12:06 we would also need to get it released and updated in u-c 14:12:15 meanwhile we marking the job non-voting 14:12:27 https://review.opendev.org/c/openstack/neutron/+/926196 14:13:59 ykarel: right. could we mark that version bad in requirements.txt? != type logic? i can never remember if gate likes that in all cases 14:16:08 ykarel: ^^ ? 14:16:33 we can come back to it 14:16:42 next bug is 14:16:51 #link https://bugs.launchpad.net/neutron/+bug/2076328 14:17:08 it's a periodic test failure, but ykarel posted a patch earlier 14:17:23 #link https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/926201 14:17:59 next un-assigned is 14:18:06 #link https://bugs.launchpad.net/neutron/+bug/2076181 14:18:17 "Accidentally no-allocation-pool for subnet leads to poor diagnosis of build-in dhcp service issues" 14:18:54 slaweq marked as low-hanging-fruit 14:20:05 last bug looks more like an rfe 14:20:11 #link https://bugs.launchpad.net/neutron/+bug/2076122 14:20:19 dns integration: support multiple domains per instance 14:20:34 i will take a look and triage for drivers meeting 14:21:04 any other bugs to discuss? 14:21:30 Current bug count this week: 720, up 2 from last week 14:21:51 And this week I am the bug deputy, next week will be rubasov - does that work for you? 14:21:59 ack, I'll be ooo on Mon-Tue, but I'll catch up when I'm back 14:22:07 Miguel Lavalle proposed openstack/neutron-tempest-plugin master: [DNM] Test metadata query over IPv6 only network https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/925928 14:23:00 #topic community-goals 14:23:45 neutronclient deprecation is continuing 14:23:49 #link https://review.opendev.org/q/topic:%22bug/1999774%22 14:24:08 Miguel Lavalle proposed openstack/neutron-tempest-plugin master: [DNM] Test metadata query over IPv6 only network https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/925928 14:24:12 reviews are welcome, all in horizon 14:24:47 regarding eventlet removal, we have one patch to merge 14:24:51 #link https://review.opendev.org/q/topic:%22bug/2069581%22 14:25:01 #link https://review.opendev.org/c/openstack/neutron/+/924317 14:25:37 added to my pile 14:25:39 it was depending on a different patch that was failing the gate 14:25:55 I'll look at it later today 14:26:05 ^that patch is -W until the other merges, trying to find it 14:26:27 it was one monkey-patching stuff 14:26:28 ack. good to know 14:26:58 I think I saw the latter last week 14:27:07 #link https://review.opendev.org/c/openstack/neutron/+/925376 14:28:05 yeap but I think ykarel stopped it because it broke something 14:28:40 well, we had rechecked enough that something is not right 14:29:29 yes most such failures i saw against that patch so looked specific to patch 14:30:07 but i didn't dig much on that 14:30:35 i'll look a little, guess i can discount the rally failure at least 14:31:53 #topic on-demand 14:32:04 does anyone have a topic? i saw nothing in the agenda 14:32:27 yes, I have 2 or 3 things regarding accessing metadat over IPv6 only network: 14:32:37 mlavalle: sure, go ahead 14:32:50 1) This patch is ready to go https://review.opendev.org/c/openstack/neutron/+/922264 14:33:58 2) To verify the previous point I implemented a neutron-tempest-plugin test case: https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/925928 14:35:03 3) This test case now succeeds with the ML2/OVN backend. However, I discovered it fails with ML2/OVS and ML2/LB so I filed https://bugs.launchpad.net/neutron/+bug/2076916 14:36:05 4) I'm going to skip the test case for ML2/OVS and ML2/LB while I fix the new bug. So the test patch will be ready for reviws later today 14:36:24 that's all 14:37:33 mlavalle: thanks for fixing that and finding the other bug, i could have sworn it worked when it originally merged 14:38:05 honestly, yesterday I thought it was somwething with my test case 14:38:21 it took me a while to convince myself that it was otherwise 14:39:05 makes me think going with the link-local was a bad idea, since i'm assuming scope had to do with some of it, the %eth0 14:39:36 the other thing that I want to mention is that there is also a bug in cloud-init 14:40:32 it doesn't know how to handle the % stuff 14:40:46 there is already a fix in cloud-init upstream 14:41:49 https://github.com/canonical/cloud-init/commit/2ea276548e05dd1ee465bb71fc4c0f5305ff0bf4 14:41:54 but it is cloud-init version 24.1.3-0ubuntu-20,04.5 14:42:16 ykarel: yeap, that one 14:42:57 so even after we fix Neutron, we need the cloud-init fix if we want VM's to actually be able to boot with ipv6 only networks 14:43:26 and i'm assuming cirros is using an old version of cloud-init? 14:44:19 the n-t-p test case I created uses a shell script injected with a config drive 14:44:29 the image I'm using is ubuntu 14:44:42 which has the old cloud-init version 14:45:42 ok, I think that is the complete picture of the metadata + ipv6 wreck 14:45:57 :( 14:46:28 the thing is we had never tested metadat over ipv6 only networks before 14:46:43 that is why it is surfacing now 14:46:57 that we are actually testing 14:47:06 i'm guessing when we tested manually we always used a %eth0 and claimed victory 14:47:31 and if was always a dual network with ipv4 14:47:38 it was^^^ 14:48:24 right. don't make me cry IPv6 tears they're twice as big :) 14:48:33 LOL 14:48:38 :D 14:48:47 IETF dork joke alert 14:49:14 I'll tell it to my wife and I'll get a dismissive stare 14:49:25 i guess they're 4 times as big depends on what you're counting 14:49:48 header versus address 14:49:54 but i digress 14:50:05 any other topics for discussion? 14:50:12 none from me 14:50:18 otherwise we have the CI meeting in 10 minutes, online this time i believe 14:51:06 #endmeeting