14:00:09 <haleyb> #startmeeting networking 14:00:09 <opendevmeet> Meeting started Tue Sep 26 14:00:09 2023 UTC and is due to finish in 60 minutes. The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:09 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:09 <opendevmeet> The meeting name has been set to 'networking' 14:00:21 <bcafarel> o/ 14:00:23 <obondarev> hi 14:00:37 <slaweq> o/ 14:00:44 <mlavalle> haleyb: is this your first official act as new PTL? 14:00:45 <mtomaska> o/ 14:00:45 <haleyb> hello everyone, taking over for ralonsoh today 14:00:50 <ralonsoh> thanks! 14:01:08 <haleyb> i'm the future PTL, so i've still got training wheels on :) 14:01:17 <mlavalle> :-) 14:01:40 <haleyb> #topic announcements 14:01:53 <haleyb> #link https://releases.openstack.org/bobcat/schedule.html 14:02:23 <haleyb> basically one week left, so if there is anything critical for Bobcat it needs to be merged asap 14:02:56 <haleyb> I did see this merged this morning - https://review.opendev.org/c/openstack/neutron/+/896335 14:03:01 <lajoskatona> o/ 14:03:14 <aprats-> o/ 14:03:23 <amorin> o/ 14:03:28 <frickler> \o 14:03:59 <haleyb> ralonsoh: since you're maybe here, do we need to push a new release tag for that ^^ 14:04:28 <ralonsoh> code is working now and this should be backported along with the previous patch introducing it 14:04:31 <ralonsoh> but is not needed 14:04:58 <haleyb> ack, didn't seem critical 14:05:04 <ralonsoh> nope 14:05:11 <frickler> but better do an rc2 now than an .1 later? 14:05:24 <lajoskatona> I think the release team will propse the new rc and the final for bobcat anyway if that was the question 14:05:27 <ralonsoh> we have time for a rc2 14:05:35 <ralonsoh> lajoskatona, exactly 14:05:36 <ykarel> o/ 14:05:54 <frickler> right, you can wait until the end of this week I think 14:05:59 <haleyb> lajoskatona: ah, ok, didn't know if we needed to manually propose this late 14:06:02 <bcafarel> what about https://review.opendev.org/c/openstack/neutron/+/893447 revert (still in master)? affecting cold migration 14:06:22 <lajoskatona> that should be included I suppose 14:06:24 <ralonsoh> yes, that needs to be in bobcat 14:06:33 <ralonsoh> so I'll propose the 2023.2 backport this week 14:06:42 <ralonsoh> once the master branch patch is merged 14:06:48 <haleyb> that was stuck due to previous gate failure 14:07:42 <bcafarel> ack let's keep an eye on master and 2023.2 backport then 14:08:07 <haleyb> if there are any other proposals for inclusion in bobcat please ping ralonsoh (or me) to prioritize 14:08:13 <ralonsoh> +1 14:08:34 <haleyb> any other announcements? 14:09:16 <haleyb> #topic bugs 14:09:36 <haleyb> last week's report is from bcafarel 14:09:49 <haleyb> #link https://lists.openstack.org/pipermail/openstack-discuss/2023-September/035216.html 14:10:32 <bcafarel> and good progress on the unassigned since I posted it 14:10:44 <haleyb> First one is (another) gate failure 14:10:49 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2037239 14:11:17 <haleyb> neutron-tempest-plugin-openvswitch-* jobs are randomly failing in gate with an l3-ha issue 14:11:53 <lajoskatona> ykarel suspected that an os-vif patch is behind it, but as I saw the logs I am not sure 14:12:00 <haleyb> a couple of us have looked through recent reviews and logs bug no silver pullet yet 14:12:26 <haleyb> lajoskatona: yes, i saw that review, https://review.opendev.org/c/openstack/neutron/+/896504 14:12:45 <lajoskatona> exactly 14:12:57 <haleyb> first pass seems to have failed the same way 14:13:06 <haleyb> tempest.lib.exceptions.TimeoutException: Request timed out 14:13:06 <haleyb> Details: Router 6986c001-1030-4a25-96cf-8445da63f144 is not active on any of the L3 agents 14:13:11 <ykarel> test patch with revert is still running https://zuul.openstack.org/status#896504 , results ther should confirm 14:13:34 <ykarel> in 1st attempt os-vif patch was not used, i pushed a devstack patch to get it applied 14:13:59 <haleyb> ykarel: ah, thanks, let's see if that helps 14:14:08 <lajoskatona> +1 14:14:54 <haleyb> if not i will take a look again at logs, or if anyone else has cycles... 14:15:25 <haleyb> next unassigned is 14:15:32 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2036603 14:17:11 <haleyb> ykarel: you think this is an advanced image issue based on last comment? 14:17:46 <ykarel> haleyb, no not saying it's due to the images used, but other tests using those images are impacted 14:17:59 <ykarel> as bug mentioned a test name, so just trying to point it's generic issue 14:18:42 <haleyb> do we know if the image changed just recently? 14:19:04 <haleyb> although i guess we didn't change it in our config 14:19:19 <opendevreview> Terry Wilson proposed openstack/neutron master: Add has_lock_periodic decorator for OVN Maintenance https://review.opendev.org/c/openstack/neutron/+/896544 14:19:58 <haleyb> either way, is there anyone that can take a closer look at that bug? 14:20:00 <ykarel> me haven't checked it if those are updated 14:20:44 <ykarel> i can check it 14:20:57 <haleyb> ykarel: ack, thanks! 14:21:02 <haleyb> next unassigned 14:21:11 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2036705 14:21:32 <haleyb> 'A port that is disabled and bound is still ACTIVE with ML2/OVN' 14:21:57 <haleyb> was confirmed it's different from ML2/OVS on irc 14:22:29 <opendevreview> Merged openstack/ovn-octavia-provider stable/2023.1: Check multiple address of a LRP plugged to LS https://review.opendev.org/c/openstack/ovn-octavia-provider/+/896267 14:24:07 <lajoskatona> I think this was on which I commented that I am not sure which one is the correct 14:24:11 <haleyb> so i guess part of that is figuring out which driver is correct, octavia assumes OVS is 14:24:40 <lajoskatona> yes, we can say that it was with OVs that is the "standard" and adopt OVN 14:25:41 <opendevreview> Jakub Libosvar proposed openstack/neutron master: Introduce ovn_nb_global config section https://review.opendev.org/c/openstack/neutron/+/896545 14:25:57 <haleyb> I will add that comment, will need owner to work on it 14:26:26 <lajoskatona> +1 14:27:25 <opendevreview> Terry Wilson proposed openstack/neutron master: Add has_lock_periodic decorator for OVN Maintenance https://review.opendev.org/c/openstack/neutron/+/896544 14:27:40 <haleyb> will move onto next 14:27:49 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2036877 14:28:02 <haleyb> 'radvd seems to crash when ipv4 addresses are supplied as nameservers to ipv6 subnets' 14:28:36 <haleyb> i just assigned this to myself, seems valid 14:29:16 <haleyb> actually saw a similar bug with dhcp recently downstream 14:29:48 <haleyb> moving onto next 14:29:53 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2037107 14:30:05 <haleyb> 'slow port creation with a large amount of networkrbacs' 14:31:08 <haleyb> ralonsoh took a look at this, could not reproduce on master but gave a lot of info on reproducing 14:31:22 <opendevreview> Jakub Libosvar proposed openstack/neutron master: Introduce ovn_nb_global config section https://review.opendev.org/c/openstack/neutron/+/896545 14:32:11 <haleyb> there was just a reply this morning, i'll assume rodolfo will reply in time 14:32:39 <haleyb> and last unassigned from last week 14:32:44 <haleyb> #link https://bugs.launchpad.net/neutron/+bug/2036423 14:32:54 <haleyb> 'subnet's gateway ip can be unset while attached to router' 14:33:16 <haleyb> i looked at this initially and noticed a couple of things 14:33:56 <haleyb> #1 the api doc seems wrong, looks like a copy/paste from POST was done in PUT section 14:34:38 <haleyb> # setting gateway_ip to None is valid, and setting it does not remove the router interface (maybe that wasn't assumed) 14:35:12 <haleyb> i could not reproduce, but submitter just replied it might only happen on multinode 14:35:28 <haleyb> anyone have a comment at least on the API behavior? 14:36:01 <lajoskatona> +1 to fix the API ref 14:37:23 <haleyb> lajoskatona: ack, assuming you also believe setting gateway_ip to None is valid :) 14:37:46 <haleyb> https://docs.openstack.org/api-ref/network/v2/index.html?expanded=update-subnet-detail,create-subnet-detail#subnets 14:37:58 <haleyb> actually the API does say it can be null... 14:38:18 <haleyb> but if not specified on create one will be selected 14:38:55 <lajoskatona> Is it really a valid usecase to set it to None? 14:39:32 <amorin> yes 14:39:35 <amorin> 'none': This subnet will not use a gateway 14:39:44 <lajoskatona> thanks 14:39:45 <amorin> openstack subnet set --help will show this to you 14:40:42 <opendevreview> Terry Wilson proposed openstack/neutron master: Add has_lock_periodic decorator for OVN Maintenance https://review.opendev.org/c/openstack/neutron/+/896544 14:40:52 <lajoskatona> I assume if we set it to None at Create time it works as expected and the PUT is problematic 14:41:11 <haleyb> i can update the api ref and will see if i can reproduce in multinode 14:41:43 <haleyb> lajoskatona: that's a good question, my testing was mostly was changing it will PUT 14:41:51 <haleyb> s/with 14:42:38 <haleyb> lajoskatona: if you're interesting in learning more about it... :) j/k 14:43:28 <haleyb> moving on, any other bugs to talk about? 14:43:31 <lajoskatona> haleyb: perhaps 2nd part of this week I ll have some time for it 14:43:46 <haleyb> ack 14:43:59 <haleyb> So this week lajoskatona is the deputy, next week will be slaweq 14:44:11 <slaweq> ok, good for me 14:44:33 <lajoskatona> I am on it 14:45:02 <haleyb> slaweq, lajoskatona: thanks 14:45:07 <haleyb> #topic specs 14:46:04 <haleyb> No real change from last week, just the OVN interconnect RFE, did have a comment so needs an update 14:46:59 <haleyb> we can open the 2024.1 folder so just need to get https://review.opendev.org/c/openstack/neutron-specs/+/891342 merged 14:47:16 <opendevreview> yatin proposed openstack/neutron-tempest-plugin master: [DNM] check os-vif https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/896549 14:47:49 <haleyb> moving on 14:47:50 <opendevreview> yatin proposed openstack/neutron master: [DNM] Test os-vif revert https://review.opendev.org/c/openstack/neutron/+/896504 14:48:02 <haleyb> #topic community_goals 14:48:32 <haleyb> I'll skip the first one, "Add support for the service role in neutron API policies" 14:48:48 <haleyb> 2) neutron client deprecation 14:49:00 <haleyb> lajoskatona: an update from last week? 14:49:19 <lajoskatona> nothing to tell the truth, I was sucked into downstream work mostly 14:50:21 <haleyb> lajoskatona: ack, and i did not click on all the "active patch" links in the wiki, will go through by next week 14:50:49 <haleyb> ok, onto last section 14:50:52 <haleyb> #topic on_demand 14:51:34 <haleyb> since it's been pretty quiet i figure people have been waiting for a turn to speak? (haleyb is snarky this morning) 14:51:43 <bcafarel> :) 14:51:48 <aprats-> Hi! About https://review.opendev.org/c/openstack/neutron/+/890459 did someone had the time to take at look at this ? 14:53:02 <amorin> ^ we are running this code in production since few weeks now, so we think it's ready for review 14:53:34 <haleyb> aprats-: sorry, i did not have time to look yet, and i think ralonsoh had some thoughts based on last weeks meeting 14:55:10 <haleyb> anything else? 14:55:23 <haleyb> ykarel: there is a CI meeting today, right? 14:55:32 <ykarel> yes 14:55:36 <amorin> I have one here: https://review.opendev.org/c/openstack/neutron/+/884711 14:55:36 <ykarel> it's video today 14:56:01 <haleyb> ykarel: and it's in :5 right? 14:56:02 <bcafarel> ack 14:56:07 <opendevreview> Arnaud Morin proposed openstack/neutron master: Register GMR with config https://review.opendev.org/c/openstack/neutron/+/884842 14:56:17 <ykarel> haleyb, yeah right 14:56:25 <amorin> and this one ^ 14:56:32 <lajoskatona> amorin, aprats-: I suppoe when the release push is over review will go back to normal again 14:56:44 <amorin> make sense, thanks :) 14:56:48 <aprats-> thx :) 14:57:12 <haleyb> amorin: ack, will take a look this week 14:58:19 <haleyb> thanks everyone for attending, need to end for next meeting 14:58:22 <haleyb> #endmeeting