Wednesday, 2025-10-22

opendevreviewHelen Chen proposed openstack/neutron master: Fix bug/2100294 to update IPv6 gateway in OVN  https://review.opendev.org/c/openstack/neutron/+/96271001:43
open10k8shi folks, can i get feedback on this please? https://bugs.launchpad.net/neutron/+bug/212550004:18
para_hi. i have a simple dns question: does designate+neutron dns integration supports private resolvers in private networks without internet access? i cant find the doc about this, neither in designate nor in neutron, so im wondering if the feature is available. thanks a lot08:48
ralonsohpara_, I don't know if we have the right people reading this now. Maybe you can try sending a mail to opendiscuss09:19
opendevreviewElvira García Ruiz proposed openstack/neutron stable/2025.2: [OVS] Avoid duplicate logging of network packets  https://review.opendev.org/c/openstack/neutron/+/96452109:57
opendevreviewElvira García Ruiz proposed openstack/neutron stable/2025.2: [OVS] Avoid duplicate logging of network packets  https://review.opendev.org/c/openstack/neutron/+/96452109:59
opendevreviewElvira García Ruiz proposed openstack/neutron stable/2025.1: [OVS] Avoid duplicate logging of network packets  https://review.opendev.org/c/openstack/neutron/+/96452209:59
opendevreviewElvira García Ruiz proposed openstack/neutron stable/2024.2: [OVS] Avoid duplicate logging of network packets  https://review.opendev.org/c/openstack/neutron/+/96452310:00
opendevreviewDai Dang Van proposed openstack/neutron master: Add dns forwarder l2 extension  https://review.opendev.org/c/openstack/neutron/+/95139010:42
ralonsohopen10k8s, I think I found the issue10:56
ralonsohthe problem seems to be in the elevated context used to update the port10:56
ralonsohI'm trying a new approach, let me push a pathc10:56
*** iurygregory_ is now known as iurygregory10:57
opendevreviewRodolfo Alonso proposed openstack/neutron master: Fix a race condition in OVN metadata port update  https://review.opendev.org/c/openstack/neutron/+/96246311:27
opendevreviewMerged openstack/neutron master: [OVS] Avoid duplicate logging of network packets  https://review.opendev.org/c/openstack/neutron/+/96447711:52
opendevreviewMerged openstack/neutron master: [UT] Allow any call order in ``test_send_ipv4_addr_adv_notif``  https://review.opendev.org/c/openstack/neutron/+/96392111:52
opendevreviewRodolfo Alonso proposed openstack/neutron master: Fix a race condition in OVN metadata port update  https://review.opendev.org/c/openstack/neutron/+/96246313:31
opendevreviewRodolfo Alonso proposed openstack/neutron master: Fix a race condition in OVN metadata port update  https://review.opendev.org/c/openstack/neutron/+/96246313:32
UgglaHi haleyb, do you think about a neutron/nova timeslot ?14:21
haleybUggla: hi, was just thinking about this as well. from what i can see in the proposals on the neutron etherpad there were only two items that overlap - https://bugs.launchpad.net/neutron/+bug/2120732 (metadata agent caching) and deprecating pass-through APIs14:23
haleybi think we talked about the second there14:23
haleybif you think it would be useful to meet for the bug that would be good, or maybe there are items on the nova etherpad to discuss?14:24
UgglaIt seems I have this one too from lajoskatona https://bugs.launchpad.net/nova/+bug/205168514:25
haleyback, and that might just be a bug according to notes, so maybe a short x-project would work14:28
lajoskatonaUggla, haleyb: Hi, that bug is not Neutron realted, but I added the neutronclient migration topic14:28
lajoskatonanot sure if we need anxproject for that14:28
lajoskatonaI meant this bug: https://bugs.launchpad.net/nova/+bug/2051685 (After repeat of incomplete migration nova applies wrong (status=error) migration context in update_available_resource periodic job )14:29
haleybok, so there's no neutron port issue there14:29
lajoskatonano that is related to migration only and no Neutron involment suprisingly :-)14:31
Ugglaok so we only have the client stuff topic14:33
Ugglain common14:33
Ugglalajoskatona, does 30mn ok for that ?14:33
lajoskatonaI hope it will be too much :-) That will be perfect, thanks14:39
haleybclient and metadata14:39
lajoskatonaahh, ok I forgot the metadata14:39
haleybsorry i'm in a meeting so not paying attention14:39
Ugglalajoskatona, and I'll find another slot for the migration bug. Is that ok ?14:40
Ugglalajoskatona, expected duration for the migration bug ?14:40
lajoskatonaUggla: I hope in 30minutes we can have that as well, Sunday afternoon I had a chat with sean-k-mooney, and based on that for me at least it is cleaner what should be a solution for that issue14:43
Ugglalajoskatona ok cool, that finally easier to cover all the points during the Neutron/Nova session. 👍14:44
Ugglahaleyb, just need to find a common 30mn timeslot now, ping me when you are free.14:45
opendevreviewRodolfo Alonso proposed openstack/neutron master: [OVN] The external networks GW chassis must the same as the GW LRP  https://review.opendev.org/c/openstack/neutron/+/96215514:55
sean-k-mooneylajoskatona: i just triaged this as high https://bugs.launchpad.net/nova/+bug/205168515:20
sean-k-mooneylajoskatona: in a least some case nova can move the migration form error to failed15:21
sean-k-mooneybut that is potitlally unsafe to do if the admin has moved the vm to the same host again15:21
sean-k-mooneybecasue https://opendev.org/openstack/nova/src/commit/a2fa7b1fa197b91aacac3aed9ba138905018465d/nova/compute/manager.py#L11940-L1199315:21
sean-k-mooneywill try and delete the instance files as part fo the clean up15:22
sean-k-mooneythe commit that intoduced this was https://opendev.org/openstack/nova/commit/255b3f2f918843ca5dd9b99e109ecd2189b6b74915:22
sean-k-mooneyand it made some assumtion that intoduce a race condtion15:22
sean-k-mooneynamely the assumed that the admin woudl not migrate the vm back to the same host before the cleanup ran on the dest15:23
sean-k-mooneyin many cases that might eb true but nova is not guarding agaisnt that case and this is not valid to do in all cause where the migration errored15:23
lajoskatonasean-k-mooney: thanks for your time and the good discussion :-)15:25
sean-k-mooneytalking atbout this at the sumit helped too15:27
sean-k-mooneyi was not really aware we did https://opendev.org/openstack/nova/commit/255b3f2f918843ca5dd9b99e109ecd2189b6b74915:27
sean-k-mooneyi.e. maed error a transiatn sate instead of terminal with a perodic15:27
sean-k-mooneyis this causing problem on the neutron side or si it purly on the nova side15:28
sean-k-mooneyi see the current logic as having 2 undesireable sideffect15:29
sean-k-mooneyfirst a vm can be accounted for twice once form the errord migration and once from the second migration that succeeded15:30
sean-k-mooneythat woudl cause the duplicate cpu pinning issue that was orgianlly reported15:30
sean-k-mooneybut also the _cleanup_incomplete_migrations function could alos lead to data lose in that case.15:31
lajoskatonasean-k-mooney: there are open patches on top of  https://review.opendev.org/c/openstack/nova/+/745281 , perhaps related to this logic ?15:31
sean-k-mooneyoh the other in the relation chang15:32
sean-k-mooneylooking at them quickly no15:33
sean-k-mooneythey are for other edgecases15:33
sean-k-mooneythe dont appare to fix the bug intoduced by the first patch15:34
sean-k-mooneyi think to "fix" this we need to do two things, 1 make sure we dont double count in update_aviable_resouces by checkign if the instnace is on the current host now.15:35
sean-k-mooneyand 2 do not do clean up if the insnatce is assigend to the current host, just set the migration to failed15:35
sean-k-mooneylet me put that in the bug report15:36
sean-k-mooneylajoskatona: i captured my toughts in https://bugs.launchpad.net/nova/+bug/2051685/comments/7 https://review.opendev.org/c/openstack/nova/+/747899/7 bight be a good impovement in general but its not a fix for the bug15:44
lajoskatonasean-k-mooney: cool, thanks, just thought to save it somewhere because I will forget for next week :-)15:46
lajoskatonasean-k-mooney: I try to check  https://review.opendev.org/c/openstack/nova/+/747899/7 and understand what it is trying to fix, thanks for pointing15:47
sean-k-mooneyits kind of simple15:47
sean-k-mooneyits trying to change how nova treat inprogress live migration after an agent restart15:48
sean-k-mooneyinstead of setting them to error it trys to revert the migraiton and set it to faild15:48
sean-k-mooneyim not sure that a good idea but that what its tryign to do15:48
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Add thread to discover BGP connections for bridge  https://review.opendev.org/c/openstack/neutron/+/96450717:10
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Add thread to discover BGP connections for bridge  https://review.opendev.org/c/openstack/neutron/+/96450717:53
haleybUggla: sorry for the slow response. we have times in our ptg schedule wednesday afternoon after 15:30, or any time thursday18:08
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Keep bridge mappings consistent  https://review.opendev.org/c/openstack/neutron/+/96442918:15
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Introduce Main router and chassis router commands  https://review.opendev.org/c/openstack/neutron/+/95989020:32
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Introduce BGP chassis commands  https://review.opendev.org/c/openstack/neutron/+/96042220:32
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: OVN agent BGP extension  https://review.opendev.org/c/openstack/neutron/+/96089520:33
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Detect BGP bridge creation  https://review.opendev.org/c/openstack/neutron/+/96442520:34
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Keep bridge mappings consistent  https://review.opendev.org/c/openstack/neutron/+/96442920:34
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Add thread to discover BGP connections for bridge  https://review.opendev.org/c/openstack/neutron/+/96450720:34
opendevreviewJakub Libosvar proposed openstack/neutron master: bgp: Add event to create BGP chassis resources  https://review.opendev.org/c/openstack/neutron/+/96185421:36
opendevreviewTerry Wilson proposed openstack/neutron master: Handle virtual port parent lookup via Neutron DB  https://review.opendev.org/c/openstack/neutron/+/95717223:25

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