Wednesday, 2026-04-22

opendevreviewRico Lin proposed openstack/neutron master: [OVN] Mark agents down when Chassis is deleted but Chassis_Private remains  https://review.opendev.org/c/openstack/neutron/+/98449500:44
opendevreviewBrian Haley proposed openstack/ovn-octavia-provider master: Use project_id in all test code  https://review.opendev.org/c/openstack/ovn-octavia-provider/+/98574302:33
opendevreviewHelen Chen proposed openstack/neutron-specs master: Propose spec for BGP EVPN Type-5 Route Support  https://review.opendev.org/c/openstack/neutron-specs/+/98225604:28
cardoehaleyb: During the Ironic/Neutron cross session I was talking about "cost" of connectivity as well as saying that VXLAN VNI 20000 might have multiple uses / meanings and the fact that Neutron has 1 pool of VNIs unlike VLANs where there's a pool per physical_network (which is what the spec I'm gonna update covers)... Well now that I see Helen's spec above... https://review.opendev.org/c/openstack/neutron-specs/+/98225605:59
cardoeThat's literally the case.05:59
cardoeSo her second picture shows what I was talking about "cost" wise.. Machines can live on the left or the right. Let's say nova is asked to build 2 servers and it randomly picks them left and right, that's the "costly" option. But if it built them both on one side or the other then its "cheaper".06:01
cardoeThe other part about the extra pools. Well there's "evpn_vni_auto_ranges" which is adding another VNI pool to neutron but hardcoding it to just 1 more. You06:08
cardoeYou'll get other asks for more pools in the future.06:08
cardoeAnd maybe abusing network-segment-range isn't correct for a pool06:35
opendevreviewRodolfo Alonso proposed openstack/neutron master: Pass ``external_ids`` in the ``address_set_add`` command  https://review.opendev.org/c/openstack/neutron/+/97792308:24
opendevreviewRodolfo Alonso proposed openstack/neutron-tempest-plugin master: Update all authentication algorithms to be sha256  https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/98577408:40
opendevreviewRodolfo Alonso proposed openstack/neutron-vpnaas master: Remove deprecated sha1, md5, and des algorithm references  https://review.opendev.org/c/openstack/neutron-vpnaas/+/98238308:50
opendevreviewAdam Harwell proposed openstack/neutron master: Fix unbounded thread pool in OVN MaintenanceThread  https://review.opendev.org/c/openstack/neutron/+/98572208:58
rm_work[m]ralonsoh: ^^ addressed, ended up just using your test verbatim, seems good08:59
ralonsohlet me check09:04
rm_workNo worries, not in a rush. We have an internal patching system so we’re good for now, will need to wait for backports too09:20
ralonsohrm_work, unit test works fine, please fix the pep8 error: https://zuul.opendev.org/t/openstack/build/aafa48fdfd044dfa9081ca89c1c2d28709:20
rm_workwtf I ran pep8 locally grrrrrr09:21
opendevreviewMartin Morgenstern proposed openstack/neutron stable/2026.1: Add logging decorator for OVN maintenance methods  https://review.opendev.org/c/openstack/neutron/+/98578309:22
opendevreviewMartin Morgenstern proposed openstack/neutron stable/2025.2: Add logging decorator for OVN maintenance methods  https://review.opendev.org/c/openstack/neutron/+/98578409:23
opendevreviewMerged openstack/neutron-fwaas master: Start using context.project_id  https://review.opendev.org/c/openstack/neutron-fwaas/+/98494909:28
opendevreviewAdam Harwell proposed openstack/neutron master: Fix unbounded thread pool in OVN MaintenanceThread  https://review.opendev.org/c/openstack/neutron/+/98572209:36
rm_work[m]weird, idk why my pep8 is not catching it locally, couldn't get a fail... but fixed according to the upstream zuul result /shrug09:46
rm_work[m]commented, happy to go with that current patch but wonder if maybe the thing opus suggested might be a good middle-ground to real testing vs determinism09:57
opendevreviewSteve Traylen proposed openstack/python-neutronclient master: Remove redundant import of reno for docs  https://review.opendev.org/c/openstack/python-neutronclient/+/98578910:05
opendevreviewMerged openstack/neutron-lib master: objects: Prepare for oslo.versionedobjects 3.10.0  https://review.opendev.org/c/openstack/neutron-lib/+/98563610:57
opendevreviewRodolfo Alonso proposed openstack/neutron-lib master: Reduce DB retry decorator retries to 1 in unit tests  https://review.opendev.org/c/openstack/neutron-lib/+/98579811:07
opendevreviewRodolfo Alonso proposed openstack/neutron-lib master: Add security-groups-default-statefulness API definition  https://review.opendev.org/c/openstack/neutron-lib/+/98435411:07
rm_work[m]ralonsoh: you're ok with the test as-is (basically the one you wrote)? not concerned about noise? I guess AI could be wrong but it definitely SOUNDED legit :D who knows with this magic code slot-machines tho lol12:36
rm_work[m]i don't care so much about the test as long as you're happy with it, since my test is monitoring it in real usage for the last 24h and it has been perfect so far12:37
ralonsohrm_work[m], I asked Claude to rewrite this test but I finally did it manually12:42
rm_work[m]haha yeah that ends up happening to me a lot too12:42
lajoskatonahaleyb, all: the first topic for today is LFX insights, and when I today open the link (https://insights.linuxfoundation.org/project/OpenStack/repository-group/neutron) I have just this msg: "This project hasn't been onboarded to LFX Insights."12:43
rm_work[m]i still like it for brainstorming / research or sometimes just to get me started12:43
lajoskatonaLast week I was happy with it (even sent to my managers), but seems it was improved in the last week :-)12:43
rm_work[m]well if in the future it starts behaving badly/flaky, the alternative test is in the comments of the CR12:43
opendevreviewRodolfo Alonso proposed openstack/neutron-tempest-plugin master: Update all authentication algorithms to be sha256  https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/98577412:48
haleyblajoskatona: sigh, it was working the other day12:52
haleybthere is always stackalytics i guess for reference12:53
lajoskatonahaleyb: I hope we have some working crystal ball 12:54
opendevreviewRodolfo Alonso proposed openstack/neutron stable/2026.1: Fix unbounded thread pool in OVN MaintenanceThread  https://review.opendev.org/c/openstack/neutron/+/98581513:02
opendevreviewRodolfo Alonso proposed openstack/neutron stable/2025.2: Fix unbounded thread pool in OVN MaintenanceThread  https://review.opendev.org/c/openstack/neutron/+/98581613:02
opendevreviewRodolfo Alonso proposed openstack/neutron stable/2025.1: Fix unbounded thread pool in OVN MaintenanceThread  https://review.opendev.org/c/openstack/neutron/+/98581713:02
*** ykarel_ is now known as ykarel13:29
ykarelralonsoh, hi, in the call you mentioned something about process is changed regarding the movement to x/namespace , can you share more context about this?13:29
ykarelor link the document or mailing list where this is described13:30
ralonsoh#link https://review.opendev.org/c/openstack/neutron/+/98255713:30
ralonsohhttps://review.opendev.org/c/openstack/neutron/+/982557/2/neutron/extensions/stateful_security_group.py13:40
ralonsohykarel, let me check that later13:40
ykarelack13:41
stephenfinhaleyb: I see you're discussing neutronclient work at the moment. I'm in the nova room for the afternoon, but am available here if there's anything I can help me. I still need to get back to lajoskatona's patch but otherwise I think I'm on top of things.14:13
stephenfin*I can help with14:13
haleybstephenfin: yes, it's next on the list, think it's mostly about the migration of things into OSC and osc plugin code into the client, i actually haven't looked at the second much myself14:15
ralonsohykarel, I'm trying to find it written somewhere. I had this conversation in the #openstack channel14:34
ralonsohykarel, from gemini: https://paste.opendev.org/show/bL9AO054QcmjedToD3Cd/. The only reference I found is https://review.opendev.org/c/openstack/project-config/+/84713514:37
ykarelralonsoh, i was mainly looking for some official doc or announcement, like i am reading https://github.com/openstack/governance/blob/e74fe14bbff38b87cd50f57a7d4efdadff860f7d/reference/dropping-projects.rst#id3 but not clear if x/ is still an option or not14:42
ralonsohykarel, yeah, I'm trying to find it too. The point is that /x namespace doesn't belong to openstack but to opendev14:43
ralonsohI think this is the problem14:43
stephenfinralonsoh: I don't have context, but fungi recently told me that the x/ namespace projects were exclusively for things that had no owner during the setup of opendev (I think that was it?) and new projects aren't allowed there (there's a job apparently). If there's a deliverable you require on living there, it should probably live elsewhere14:48
stephenfinYou may already know all of that. If so, sorry for the noise!14:48
ralonsohstephenfin, actually that was the topic: what to do with unattended projects14:49
ralonsohI thought moving to /x was now forbidden, but I can be wrong14:49
stephenfinIs it a project that you (neutron) used to maintain that's no longer maintained, or a project that you simply depend on which no longer has a clear owner?14:51
fungicorrect, we don't want to add any more repositories to the x/ namespace, but you can implicitly create a new namespace just by prepending a repository name with it14:53
ralonsohfungi, ok, nice to know this!14:54
ykarelfungi, stephenfin thx that clears our doubts here, do we also document this somewhere?14:54
fungiwe can also fairly easily (though not instantly) move repositories fro x/ to another namespace (new or existing)14:55
fungihttps://docs.opendev.org/opendev/infra-manual/latest/creators.html#decide-status-and-namespace-of-your-project would be a good place to expand if it's unclear14:56
lajoskatonafungi, stephenfin, ralonsoh: an example from the past is networking-l2gw under x/ as I remember it was a Neutron stadium but when nobody maintained it, it was moved there, and it still lives there (I maintain it mostly when I have time or when it has some serious issue)14:58
fungiit could have just as easily moved to oldtron/networking-l2gw or something14:59
lajoskatona:-)15:00
fungiright after the openstack/ namespace eviction, we were still feeling things out. that was also before openstack instituted the mandatory retirement policy, i think15:01
otherwiseguyhaleyb: on https://review.opendev.org/c/openstack/neutron-specs/+/983896/2/setup.py do we know why the #! line was removed? And since it is still marked as executable, is that an issue?15:17
opendevreviewMerged openstack/neutron-specs master: Remove outdated comment in setup.py  https://review.opendev.org/c/openstack/neutron-specs/+/98389615:17
haleybotherwiseguy: i guess it doesn't have to be executable, could do some git-fu and fix that i guess15:19
otherwiseguyhaleyb: just didn't know if the globally managed part was linked to the removing #! part at all, or if it was a separate change. But in any case it seemed like executability and #! should be linked :)15:21
otherwiseguyI thought I had time to discuss while tests ran, but they were very quick :p15:22
opendevreviewBrian Haley proposed openstack/neutron-specs master: Change mode of setup.py to 644  https://review.opendev.org/c/openstack/neutron-specs/+/98583515:26
haleybotherwiseguy: yeah, it merged quick, think was just to bring in-line with neutron. but ^^^15:26
haleybarg, and i can't type15:26
opendevreviewBrian Haley proposed openstack/neutron-specs master: Change mode of setup.py to 664  https://review.opendev.org/c/openstack/neutron-specs/+/98583515:26
otherwiseguytyping is overrated15:26
haleyband i can't type again15:27
haleybit's odd i used chmod 664 but change says mode 100644 so i'm confused15:27
haleybwhatever15:28
otherwiseguyhaleyb: ai is telling me that the 100 prefix signifies "file type is a regular file"15:29
otherwiseguyand you can trust the computers implicitly, obviously. they mean us no harm.15:29
haleybyeah, the fact it thinks it's 644 when it's not, well whatevs15:30
haleybi've got my aluminum hat on now so skynet can't track me15:31
opendevreviewDoug Goldstein proposed openstack/neutron master: allow physical_network to be set for provider networks on VXLAN  https://review.opendev.org/c/openstack/neutron/+/98583715:33
opendevreviewDoug Goldstein proposed openstack/neutron master: allow physical_network to be set for provider networks on VXLAN  https://review.opendev.org/c/openstack/neutron/+/98583715:34
cardoehaleyb: yeah git uses some extra digits in front to carry other data about the file but that's a normal file15:35
otherwiseguyhaleyb: when I check it out, it's 64415:37
haleyb-rw-rw-r-- is what it shows in my copy, wtf15:38
otherwiseguyWeird. But when I check out w/o patch 755, w/ patch 644, which is at least consistent. Something w/ parent directory permissions or something?15:39
haleybdrwxrwxr-x15:39
haleybi'll see if i can hack it to 66415:40
otherwiseguyLooks like git only tracks executability15:40
otherwiseguyhaleyb: so anything else will just be whatever you had.15:41
haleybgit only has two modes apparently, 755 and 644, as long as nothing breaks15:43
ykarelhaleyb, can you revisit https://review.opendev.org/c/openstack/neutron/+/98436415:58
haleybykarel: done16:07
ykarelthx haleyb 16:07
opendevreviewBrian Haley proposed openstack/neutron stable/2026.1: Ignore designate NotFound errors during PTR delete  https://review.opendev.org/c/openstack/neutron/+/98585017:06
opendevreviewBrian Haley proposed openstack/neutron stable/2025.2: Ignore designate NotFound errors during PTR delete  https://review.opendev.org/c/openstack/neutron/+/98585117:06
opendevreviewBrian Haley proposed openstack/neutron stable/2025.1: Ignore designate NotFound errors during PTR delete  https://review.opendev.org/c/openstack/neutron/+/98585217:07
opendevreviewBrian Haley proposed openstack/neutron stable/2024.2: Ignore designate NotFound errors during PTR delete  https://review.opendev.org/c/openstack/neutron/+/98585317:08
opendevreviewBrian Haley proposed openstack/neutron unmaintained/2024.1: Ignore designate NotFound errors during PTR delete  https://review.opendev.org/c/openstack/neutron/+/98585417:08
opendevreviewMerged openstack/neutron master: [functional fips] Switch venv prep to python3.12  https://review.opendev.org/c/openstack/neutron/+/98436418:10
opendevreviewBrian Haley proposed openstack/neutron-tempest-plugin master: Use project_id key in api and scenario tests  https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/98273618:46
opendevreviewMerged openstack/neutron master: Fix unbounded thread pool in OVN MaintenanceThread  https://review.opendev.org/c/openstack/neutron/+/98572220:11
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Update main router policies when chassis is added  https://review.opendev.org/c/openstack/neutron/+/98589521:36
opendevreviewMerged openstack/neutron master: Add documentation for BGP multinode tempest job  https://review.opendev.org/c/openstack/neutron/+/98431822:42
opendevreviewMerged openstack/neutron master: Add Neutron mark to localnet ports  https://review.opendev.org/c/openstack/neutron/+/98441123:48

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!