Wednesday, 2026-01-21

*** liuxie is now known as liushy03:04
slaweqhaleyb: hi, just FYI, together with elvira1 we have RFE to discuss: https://bugs.launchpad.net/neutron/+bug/2138746 - can it be added to the drivers meeting agenda for this or next week?08:29
ralonsohlajoskatona, elvira hi! I have a question about https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/973626/10/zuul.d/master_jobs.yaml#36309:35
ralonsohwhy taas should be running only in OVS?09:36
lajoskatonaralonsoh: tap-services and tap-flows are working only with OVS currently and SRIOV09:45
ralonsohlajoskatona, but I don't understand, we added support for ovn09:48
ralonsohhttps://github.com/openstack/tap-as-a-service/commit/a2523e1e29f8f0142fa6e13eeda69af3e9cc59b509:48
lajoskatonaralonsoh: only for tap-mirrors which works both with native OVS and OVN, and now elvira extended the OVN part with the both direction which was added alter to the OVN core itself10:00
opendevreviewElvira García Ruiz proposed openstack/neutron master: Enable tap-as-a-service OVN extension  https://review.opendev.org/c/openstack/neutron/+/97392010:04
ralonsohlajoskatona, but there are specific tests for this10:04
ralonsohI don't know why TestTaaSTrafficScenarios shouldn't be executed10:04
lajoskatonaralonsoh: the naming of the scenario modules not that straight forward, scenario/test_taas.py i.e. has test for sriov, test_traffic_impact.py for tap_services and tap_flows on OVS, and test_tap_mirror is for tap_mirrors executed on OVS and OVN10:16
lajoskatonaralonsoh: at least that is my memory, perhaps I have to revisit these tests and jobs for them10:16
ralonsohooook, I think that deserves at least a note in the OVS taas extension10:17
ralonsohjust to be aware of this10:17
ralonsohlajoskatona, can you push this patch for n-t-p? just to have this info10:18
ralonsohelvira, ^^10:18
ralonsohI'll +2 your patch now10:18
opendevreviewMerged openstack/neutron master: Bump oslo.config to 10.2.0  https://review.opendev.org/c/openstack/neutron/+/97249110:25
opendevreviewMerged openstack/neutron master: Fix ``LogicalRouterPortEvent`` when a LR is deleted  https://review.opendev.org/c/openstack/neutron/+/97358410:25
elviraHi, just read all the conversation. Thanks a lot for the inputs! is there anything I should change from the patches? I already removed taas from enabled extensions in neutron ovn 10:28
elviraI think maybe the post-config in ovn job is not needed anymore since I guess taas tests will be skipped anyway in OVN, not sure what you think ralonsoh 10:30
ralonsohelvira, maybe it could be nice to have a comment in the taas plugin in the OVS job10:33
ralonsohin another patch10:33
elvira ralonsoh: to note that the post_config is needed for that? or to say that we only support taas extension in ML2/OVS?10:35
ralonsohthe last, that taas is only for OVSA10:36
ralonsohOVS*10:36
elviraok! will do10:36
elviraI will also remove the post-config in OVN. I can always put it back if something fails but I think that was failing only on taas tests iiuc10:37
wilkmarcinralonsoh: I'd appreciate it if you shared your comments in https://review.opendev.org/c/openstack/neutron/+/971876, whenever you have time. Thank you in advance10:39
ralonsohwilkmarcin, let me check that today. I would also want terry wilson to check it too10:40
wilkmarcinThank you, ralonsoh10:42
opendevreviewElvira García Ruiz proposed openstack/neutron-tempest-plugin master: Add both direction tap-as-a-service scenario test  https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/97362611:17
opendevreviewElvira García Ruiz proposed openstack/neutron-tempest-plugin master: Add comment highlighting taas is ML2/OVS only  https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/97404011:19
opendevreviewMaor Blaustein proposed x/whitebox-neutron-tempest-plugin master: Fix allows multi value configuration options  https://review.opendev.org/c/x/whitebox-neutron-tempest-plugin/+/97404111:28
opendevreviewAlexey Stupnikov proposed openstack/neutron master: Use RFC 5398 reserved AS numbers in Neutron documentation  https://review.opendev.org/c/openstack/neutron/+/97404211:28
*** ralonsoh_ is now known as ralonsoh12:26
haleybslaweq, elvira: i will have to read the docs on PVLAN but i think you can add to this week drivers agenda, i might be out next friday14:49
haleybslaweq: and while i have your attention can you look at https://review.opendev.org/c/openstack/neutron/+/971628 - neutron-lib bump14:50
haleybor lajoskatona ^^14:50
lajoskatonahaleyb: cool, I wf+1-ed it14:54
opendevreviewMerged x/whitebox-neutron-tempest-plugin master: Fix allows multi value configuration options  https://review.opendev.org/c/x/whitebox-neutron-tempest-plugin/+/97404115:05
ralonsohhaleyb, hello! Now in the correct channel15:09
ralonsohthis is about https://review.opendev.org/c/openstack/neutron/+/97386615:09
ralonsohYou are right, this is the fix that solves the problem in the driver15:09
ralonsohbut you know sometimes these fixes take sometime to land in some D/S repositories15:10
haleybralonsoh: ah yes, we had tracked that down a few months back during a fire drill with a customer15:10
ralonsohand it could happen that the customer doesn't have the most updated drivers15:10
ralonsohso this is why I'm suggesting this patch15:10
haleybralonsoh: the reason we didn't go with a neutron change is there are other places in the code that just always poke the interface and don't check state15:11
ralonsohI can add a TODO to remove it in some releases15:11
ralonsohhaleyb, but for SR-IOV this is the only method that changes the link state15:11
ralonsohand nova-cpu is not doing it15:11
haleybplus we found the offending commit from intel15:11
ralonsohhaleyb, so you don't think this should go into the neutron repo?15:12
ralonsoheven with a TODO to remove it?15:12
slaweqhaleyb sure, I will add it to the agenda then15:14
slaweqand sorry for not checking earlier this patch but I was in the meeting15:14
slaweqand lajoskatona already approved it :)15:14
slaweqhaleyb are you still using https://wiki.openstack.org/wiki/Meetings/NeutronDrivers for the drivers meeting?15:15
haleybralonsoh: i'm on the fence since i was involved in the triage of the driver, i'll explain...15:15
haleybthere was a change to the i40e driver to make sure things were (re)initialized properly, and while it broke this case, it seemed that not blindly setting link UP might potentially cause an issue in the driver15:18
haleybwe worked with intel on the fix and verified in a partner cloud15:18
haleybi think the neutron change would have worked too, but since the i40e driver was seemingly the only one broken we went with the targeted change15:19
haleybplus we could release the kernel quicker15:19
ralonsohhaleyb, ok, I'll push this patch only in our D/S repos15:19
ralonsohI'll abandon it15:19
haleybralonsoh: ack15:22
slaweqok, I've added it to the On Demand agenda in https://wiki.openstack.org/wiki/Meetings/NeutronDrivers#Agenda15:28
slaweqthx haleyb 15:28
opendevreviewElvira García Ruiz proposed openstack/neutron-tempest-plugin master: Add comment highlighting taas is ML2/OVS only  https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/97404015:31
opendevreviewsean mooney proposed openstack/os-vif master: Add TAP device pre-creation support for OVS/OVN  https://review.opendev.org/c/openstack/os-vif/+/97123116:08
opendevreviewMerged openstack/neutron master: Bump neutron-lib to 3.23.0  https://review.opendev.org/c/openstack/neutron/+/97162817:04
opendevreviewMerged openstack/neutron master: bgp: Detect BGP bridge creation  https://review.opendev.org/c/openstack/neutron/+/96442518:15
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Keep bridge mappings consistent  https://review.opendev.org/c/openstack/neutron/+/96442918:17
opendevreviewLajos Katona proposed openstack/neutron master: [eventlet-removal] Handle systemctl stop for DHCP agent  https://review.opendev.org/c/openstack/neutron/+/96790218:53
opendevreviewLajos Katona proposed openstack/neutron master: WIP: [eventlet]: Make ovs-agent stop again  https://review.opendev.org/c/openstack/neutron/+/96710518:53
opendevreviewBrian Haley proposed openstack/neutron master: Use project_id in HANetworkConcurrentDeletion()  https://review.opendev.org/c/openstack/neutron/+/97424720:13
opendevreviewBrian Haley proposed openstack/neutron-lib master: Remove tenant_id from HANetworkConcurrentDeletion()  https://review.opendev.org/c/openstack/neutron-lib/+/97424820:16
opendevreviewBrian Haley proposed openstack/neutron master: Use project_id key in port plugin code  https://review.opendev.org/c/openstack/neutron/+/97298220:40
opendevreviewBrian Haley proposed openstack/neutron master: Stop using tenant_* in context objects  https://review.opendev.org/c/openstack/neutron/+/97425321:03
cardoeTheJulia, slaweq, haleyb: The PVLAN stuff actually layers really nicely/easily with the VXLAN stuff that TheJulia and I have been working on. Because the private VLAN or Vi is just a lower segment to the Vp which is the upper segment.21:58
cardoeCisco's own docs reference each other.22:00
cardoeThe spec and the docs that I submitted can cover PVLAN with a simple sed.22:01
cardoeThe ML2 driver that I submitted to the neutron repo and TheJulia has now moved over to networking-baremetal can cover PVLAN by changing the config.22:02
cardoeWe'd love to work with you folks.22:05
opendevreviewBrian Haley proposed openstack/neutron-lib master: Remove tenant_id from HANetworkConcurrentDeletion()  https://review.opendev.org/c/openstack/neutron-lib/+/97424822:29
haleybcardoe: great, can you sync with elvira and/or slaweq on it? they were going to present at drivers meeting this week22:29
cardoeYep.22:32
cardoeWe can also update that data center networking doc to include an example of this as well.22:32
cardoeUnless what you're really looking for is doing this with VM then it wouldn't make sense for that doc.22:32
* haleyb shrugs22:33

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!