Tuesday, 2025-12-02

opendevreviewOleksandr Kozachenko proposed openstack/neutron master: Fix a race condition in OVN metadata port update  https://review.opendev.org/c/openstack/neutron/+/96246302:03
opendevreviewOleksandr Kozachenko proposed openstack/neutron master: Fix a race condition in OVN metadata port update  https://review.opendev.org/c/openstack/neutron/+/96246302:10
opendevreviewJimin Shin proposed openstack/neutron master: Add information to network ip availabilities  https://review.opendev.org/c/openstack/neutron/+/96508302:24
opendevreviewBrian Haley proposed openstack/neutron master: Change some unit test code to use project_id  https://review.opendev.org/c/openstack/neutron/+/96808404:34
opendevreviewBrian Haley proposed openstack/neutron master: Change some functional test code to use project_id  https://review.opendev.org/c/openstack/neutron/+/96808504:35
opendevreviewBrian Haley proposed openstack/neutron master: Change test code to use project_id in dictionaries  https://review.opendev.org/c/openstack/neutron/+/96841404:35
*** ykarel_ is now known as ykarel07:35
opendevreviewTobias Urdin proposed openstack/neutron master: Cleanup stale DHCP ports not bound to host  https://review.opendev.org/c/openstack/neutron/+/96877008:02
ralonsohhi folks! The 2nd bug fix for the OVS manager FT test: https://review.opendev.org/c/openstack/neutron/+/96910708:45
ykareladded a comment08:52
ralonsohthanks!08:55
opendevreviewRodolfo Alonso proposed openstack/neutron master: [FT] Wait for the manager to be created (2)  https://review.opendev.org/c/openstack/neutron/+/96910709:11
opendevreviewMerged openstack/neutron stable/2025.2: Don't prevent setting CIDRs as allowed_address_pair for port  https://review.opendev.org/c/openstack/neutron/+/96912609:12
opendevreviewMerged openstack/neutron stable/2025.1: Don't prevent setting CIDRs as allowed_address_pair for port  https://review.opendev.org/c/openstack/neutron/+/96912709:12
opendevreviewMerged openstack/neutron stable/2024.2: Don't prevent setting CIDRs as allowed_address_pair for port  https://review.opendev.org/c/openstack/neutron/+/96912809:12
opendevreviewRodolfo Alonso proposed openstack/neutron master: [FT] Add extra logs to ``test_create_bridges`` test  https://review.opendev.org/c/openstack/neutron/+/96920209:36
opendevreviewSlawek Kaplonski proposed openstack/neutron master: [Docs] Add doc about Virtual IP addresses in Neutron ML2/OVN  https://review.opendev.org/c/openstack/neutron/+/96881210:04
zigoralonsoh: Hi there! I wonder, what's missing for dynamic-routing eventlet removal to work?11:24
zigoIs it that you don't understand what's going on?11:25
ralonsohzigo, I don't know, this is why I wrote this mail requesting help on this issue. It seems (maybe) to be a problem in the os-ken library, according to the logs from https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/95674711:26
ralonsohbut I didn't spend more time on this11:27
zigoralonsoh: os-ken is still full of eventlet stuff...11:27
zigoShouldn't we start by it first?11:28
ralonsohzigo, no, we migrated it to threading. Actually in os-ken 4.0.0, "native" is the default backend11:28
zigoIf one sets the env var OSKEN_HUB_TYPE=native, right?11:29
ralonsohbut of course, there could still be some bugs on untested code branchs, like bgp: https://zuul.opendev.org/t/openstack/build/bbfba96502fb40aeb3083a05e2ca08bf/log/controller/logs/screen-q-dr-agent.txt#123611:29
ralonsohbranches*11:29
zigo(which seems to be the default)11:29
ralonsohyes, this is now the default one with 4.0.011:29
zigoI'll have to fix my neutron-ipv6-bgp-injector then ! :)11:30
zigoralonsoh: BTW, I was thinking, but not sure how, to get my neutron-ipv6-bgp-injector integrated with neutron-dynamic-routing. See https://salsa.debian.org/openstack-team/services/neutron-ipv6-bgp-injector/11:31
zigoYour thoughts?!?11:31
zigoIs it better to keep it a standalone project?11:31
zigoWhat it does: advertize each /128 over BGP, so that our switch can block traffic for IPs that aren't provisionned on the OpenStack side, eliminating a lot of noise trafic, like scans ...11:32
ralonsohis that a generic bgp tool? can we create functional test for it?11:33
ralonsohalso is it possible to have it also for ipv4?11:33
ralonsohthat could be added (maybe) as a tool in n-d-r11:33
ralonsohyou can propose a RFE11:33
zigoFor IPv4, we use dynamic-routing, but it wouldn't do the work for IPv6.11:34
zigoie: for IPv6, only /64 are advertized.11:34
lajoskatonazigo, ralonsoh: I plan to go back to ovs-agent and os-ken vs threading this week11:35
zigoFor integration in n-d-r, it would need to add a subnet table so one could add ipv6 subnets with the API (instead of the config file).11:35
zigolajoskatona: great ! :)11:35
zigoAnd thanks.11:35
haleyb#startmeeting networking14:01
opendevmeetMeeting started Tue Dec  2 14:01:17 2025 UTC and is due to finish in 60 minutes.  The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot.14:01
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:01
opendevmeetThe meeting name has been set to 'networking'14:01
haleybPing list: bcafarel, elvira, frickler, mlavalle, mtomaska, slaweq, tobias-urdin, ykarel, lajoskatona, jlibosva, averdagu, haleyb, ralonsoh14:01
elvirao/14:01
ralonsohhedllo14:01
ralonsohhello*14:01
lajoskatonao/14:01
rubasovo/14:02
haleyb#announcements14:02
cbuggyo/14:02
haleybWe are in week R-17 of Gazpacho14:02
haleybOur next milestone in this development cycle will be gazpacho-2, on January 8, 202614:03
haleyb#link https://releases.openstack.org/gazpacho/schedule.html14:03
haleybSo we should start getting any new features merged14:04
haleybContinue to use the priorities dashboard for important changes or things ready to merge14:04
haleyb#link https://tinyurl.com/59z278km14:04
haleybthink there was only one patch there yesterday14:05
haleybReminder: If you have a topic for the drivers meeting on Friday, please add it to the wiki @ https://wiki.openstack.org/wiki/Meetings/NeutronDrivers14:05
haleybwas there anything important i missed being out most of last week?14:06
haleyblooks like CI has improved :)14:06
haleybalright can move on14:07
haleyb#topic bugs14:07
haleybelvira was bug deputy last week, her report is at14:07
haleyb#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/NO5J73HGRWUDT6AOKVTG5QZ4Z7HICJI6/14:07
haleyblooks like the critical and high ones have owners and fixes proposed/committed14:08
haleybfirst medium one14:09
haleyb#link https://bugs.launchpad.net/neutron/+bug/213234214:09
haleybOVN LB / FIP PF does not reply to ARP when VIP is from different subnet than router14:09
haleybovn-octavia-provider issue it seems14:09
haleyblooks like OVN behavior changed14:10
haleybSince i see froyo commented I'll assume he will help address it14:11
haleybnext one is 14:12
haleyb#link https://bugs.launchpad.net/neutron/+bug/213321214:12
haleyb[S-RBAC] Regular user can't never send project_id in request body, even with custom policy rules which allows that14:12
slaweqI will work on this one14:12
haleybslaweq: thanks for picking this up, did it need a discussion at all?14:12
slaweqI'm not sure if we need to talk about it really14:13
haleybok, will look for the patch :)14:13
slaweqwe should probably add new rule in policy to let operators allow to do that for specific roles14:13
slaweqbut I think it will be better to discuss in the patch review once it will be pushed to gerrit14:14
slaweqI plan to work on it this week14:14
haleyback, thanks14:15
haleybfinal unassigned one is14:15
haleyb#link https://bugs.launchpad.net/neutron/+bug/213238214:15
haleyb[RFE] Refactor the SR-IOV agent to use the oslo.service launcher14:15
ralonsohyeah, this is just an improvement14:16
ralonsohwe use oslo.serive launch for all agents except SR-IOV14:16
ralonsohit could be desirable to have this14:16
haleybfor a second i thought it was eventlet related14:16
ralonsohnononono14:16
haleybi'm assuming any change would look like what was done in other agents14:16
ralonsohyes, that is the point14:17
ralonsoh(and removing the manual signal handling, for example)14:17
haleybi can take a closer look later14:18
ralonsohthanks!14:18
haleybthat was all the unassigned/new bugs from last week, any others to discuss?14:19
haleybthis week slaweq is the bug deputy, next week is myself14:19
haleybslaweq: is that ok for you?14:19
slaweqsure14:20
haleybok, moving on14:21
haleyb#topic specs14:21
haleyb#link https://review.opendev.org/q/project:openstack/neutron-specs+status:open14:21
haleybi really need to review that list14:22
haleybwe can move on14:23
haleyb#topic community goals14:23
haleyblajoskatona: can you give a neutronclient update?14:23
lajoskatonasome, I hope I can push today a nova patch for ports14:23
haleyb#link https://review.opendev.org/c/openstack/nova/+/962604 is probably one of them14:24
haleyb#link https://review.opendev.org/q/topic:%22sdk_for_neutron%2214:24
lajoskatonaand perhaps refresh for some of the older patches as I found some small mistakes during the work on port calls migration to SDK14:24
lajoskatonayes thanks, those are the nova patches14:24
lajoskatonathat's it for novaclient14:25
lajoskatonaAnd Miro started to work on moving OSC related code from n-client to OSC, I have to go to test that as well14:26
haleybah yes14:26
lajoskatonahttps://review.opendev.org/c/openstack/python-openstackclient/+/96344514:26
haleyblajoskatona: have you had many reviews on the nova patches? just want to make sure they don't get ignored14:27
lajoskatona0 (zero)14:27
lajoskatonaI have t poke them a little14:27
ralonsohhttps://www.youtube.com/watch?v=3n6ZY0tYsJg14:28
ralonsoh(sorry for the joke)14:28
lajoskatona:-)14:28
haleybperfect14:29
haleybi will put them in my queue14:29
haleybralonsoh: are there any more eventlet things you know of?14:31
ralonsohjust the n-d-r issues, related to the eventlet removal14:31
ralonsohzigo is going to work on this14:31
ralonsohit could be an error in the os-ken library, a module we don't use in neutron (bgp)14:32
haleybah, this was discussion right before meeting, will look through comments14:32
ralonsohyes14:32
ralonsohapart from this, I would suggest to remove this topic from the weekly meetings because we can consider that most of the job is done (Neutron and several networking projects)14:33
ralonsohso we can handle this ad hoc, when needed14:33
haleybthere definitely could be another eventelet bug in os-ken14:33
haleybralonsoh: ack, we can just file bugs when we find these issues14:33
ralonsohso congrats everyone! we had a long journey here but we finished14:34
ralonsoh(almost)14:34
lajoskatona\o/14:35
haleybagreed, thanks for all the hard work by a number of people over the past year+14:35
haleyb#topic on-demand14:36
haleybralonsoh: i think you had an item14:36
ralonsohyes, https://review.opendev.org/c/openstack/neutron/+/96824214:36
ralonsohthis patch (thanks takashi) is restoring the "multi processing" for the metadata14:37
ralonsohit is actually using threads, is not exactly the same14:37
ralonsohbut as long as the waiting for the event is not CPU bound, this coudl improve the performance14:37
ralonsohmy only concern is that before this imlementation, we were recommending to use metadata_workers=0 to reduce the SB connections from the compute nodes14:38
ralonsohthis patch is "reverting" that14:38
haleybright, that was the main thing i noticed14:39
ralonsohso I would ask to make the assumption that metadata_workers=0 is a valid input and in this implementation metadata_workers=0 == metadata_workers=114:39
ralonsohthat needs to be documented and implemented (1 minute coding)14:39
ralonsohthat's all, you can leave a comment in the patch14:40
lajoskatonayou mean in the same patch or in a follow-up ?14:40
ralonsohsame patch14:40
ralonsohbecause Takashi is multitasking, I'll ask him if I can push a new PS14:40
ralonsohbut after your reviews and comments14:41
lajoskatona+114:41
ralonsohthanks folks!14:41
haleyband i would agree with the workers==0 comment14:42
ralonsohcool14:42
haleybany other on-demand topics?14:43
haleybok, thanks for attending and have a great week!14:44
haleyb#endmeeting14:44
opendevmeetMeeting ended Tue Dec  2 14:44:20 2025 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)14:44
opendevmeetMinutes:        https://meetings.opendev.org/meetings/networking/2025/networking.2025-12-02-14.01.html14:44
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/networking/2025/networking.2025-12-02-14.01.txt14:44
opendevmeetLog:            https://meetings.opendev.org/meetings/networking/2025/networking.2025-12-02-14.01.log.html14:44
ralonsohbye14:44
opendevreviewRodolfo Alonso proposed openstack/neutron master: iPXE over IPv6 supported since OVN v23.06.0  https://review.opendev.org/c/openstack/neutron/+/96881114:44
lajoskatonaBye14:44
zigoralonsoh: I don't think I have the needed skills to work on eventlet removal for n-d-r. :/14:46
ralonsohhaleyb, ^ maybe we need a reminder in the mail I sent, just to know if anyone else is able to move this forward14:57
haleybralonsoh: sure let me look14:59
opendevreviewRodolfo Alonso proposed openstack/neutron master: [OVN] Add documentation/reno for the OVN external port N/S routing  https://review.opendev.org/c/openstack/neutron/+/96879715:01
opendevreviewRodolfo Alonso proposed openstack/neutron master: WIP == [OVN] OVN agent configuration options list missing "extensions"  https://review.opendev.org/c/openstack/neutron/+/96926615:49
opendevreviewRodolfo Alonso proposed openstack/neutron master: Remove a missing Windows configuration file  https://review.opendev.org/c/openstack/neutron/+/96926716:06
opendevreviewMerged openstack/neutron master: [FT] Wait for the manager to be created (2)  https://review.opendev.org/c/openstack/neutron/+/96910716:23
opendevreviewMerged openstack/neutron unmaintained/2024.1: Don't prevent setting CIDRs as allowed_address_pair for port  https://review.opendev.org/c/openstack/neutron/+/96915518:08
opendevreviewMerged openstack/neutron master: [FT] Add extra logs to ``test_create_bridges`` test  https://review.opendev.org/c/openstack/neutron/+/96920220:55
opendevreviewBrian Haley proposed openstack/neutron-dynamic-routing master: [eventlet-removal] Removal of eventlet usage in this repository  https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/95674722:45

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