14:01:27 #startmeeting networking 14:01:27 Meeting started Tue Feb 14 14:01:27 2023 UTC and is due to finish in 60 minutes. The chair is ralonsoh. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:01:27 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:01:27 The meeting name has been set to 'networking' 14:01:29 o/ 14:01:29 hello all 14:01:32 hi 14:01:36 o/ 14:01:37 \o 14:02:12 o/ 14:02:31 o/ 14:02:37 o/ 14:02:46 ok, let's start 14:02:48 * mlavalle is an internal company meeting. will lurk in the background 14:02:56 #topic announcements 14:03:02 Antelope / 2023.1 schedule: https://releases.openstack.org/antelope/schedule.html 14:03:21 this week is R-5 14:03:23 o/ 14:03:31 and that is antelope-3 milestone 14:03:50 remember this he final release for client libraries 14:03:54 and the feature freeze 14:04:09 the RC1 will be in 2 weeks 14:04:35 also remember that you have time to propose your candidacy for Neutron PTL (and other projects) 14:04:52 please check the calendar in https://governance.openstack.org/election/ 14:05:21 and a couple of needed links for the next PTG 14:05:28 Bobcat etherpad: https://etherpad.opendev.org/p/neutron-bobcat-ptg 14:05:35 \o/ 14:05:46 please start adding your topics to the Bobcat Neutron etherpad 14:05:58 +1 14:06:10 during the next meetings, we'll check them and make the schedule 14:06:45 I'll remember this link every week 14:07:05 I also want to talk about the Vancouver meeting 14:07:21 but, if you don't mind, I'll move this topic to the "on demand" section 14:07:41 ok, I anyway added the advertisement for Vancouver forum proposals 14:07:55 lajoskatona, of course! 14:07:59 I mean I added to the on-demand agenda 14:08:36 right, I didn't see that 14:08:37 cool 14:09:08 and as usual, please check https://openinfra.dev/live/#all-episodes 14:10:00 ok, I think I'm not missing anything 14:10:17 related to the release schedule or anything 14:10:46 ok, let's move to the next topic then 14:10:52 #topic bugs 14:11:08 last week report is from elvira 14:11:16 #link https://lists.openstack.org/pipermail/openstack-discuss/2023-February/032167.html 14:11:32 so far, there are some bugs still not assigned 14:11:48 #link https://bugs.launchpad.net/neutron/+bug/2006953 14:11:59 I think ykarel has sent a patch recently 14:12:08 yes 14:12:11 right: https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/873708 14:12:20 ok thanks! 14:12:43 ahhh ok, I see the issue, thanks! 14:13:05 then next one is 14:13:06 #link https://bugs.launchpad.net/neutron/+bug/2006683 14:13:15 but I think ykarel sent another patch 14:13:25 no lajoskatona sent it 14:13:30 ah yes 14:13:34 and is merged now: https://review.opendev.org/c/openstack/neutron-fwaas/+/873360 14:13:39 (cool) 14:13:40 but not full fix 14:14:14 what else is missing? 14:14:17 i commented there https://bugs.launchpad.net/neutron/+bug/2006683/comments/3, can you check if that makes sense 14:14:26 exactly, I saw this in the logs but I am sure that it will not solve all issues, just cleans th log before the exact traceback 14:15:22 ykarel: I check your comment later, I missed that 14:15:34 ok so we are still seeing txn errors 14:15:42 due to the slqalchemy migration 14:16:14 yeah, but as I see sqalchemy2 will not be "included", at least the req bump patch is still red 14:16:33 and bauzas from Nova asked to skip it 14:16:34 lajoskatona, ack, we can check few more runs if it's sees again with your fix included 14:16:54 yes the current failures are with sqlalchmey-1.4 14:17:29 in any case, we need to check what is the txn that is failing, or the operation 14:17:37 let's keep an eye on this one 14:17:41 +1 14:17:44 +1 14:18:01 the last one I had is 14:18:04 #link https://bugs.launchpad.net/neutron/+bug/2006145 14:18:57 this one is for neutron-dynamic-routing as I see 14:18:58 maybe ltomasbo ^^ can help us on this one 14:19:39 I'll take a look, too. 14:19:58 but I would say that what they are testing could not be supported 14:20:00 I see the use case, otoh the current system is kind of working as designed 14:20:17 they are manually stopping the RPC communication 14:20:28 frickler, it is? 14:21:01 well the current implementation assumes that rmq is always available 14:21:06 right 14:21:08 that's why we have HA for it 14:21:30 but if operators really need to shut it down for upgrades, we could try to adapt for that 14:21:49 so that could be a RFE, not a bug 14:22:00 yes, I agree to that 14:22:10 perfect, do you mind commenting on the bug? 14:22:22 I also wonder how everything else in openstack handles an rmq outage 14:22:34 frickler, ralonsoh: I know when using FRR you have some options to keep those routes. As NDR has its own speaker implementation, I suppose either that needs to be enabled (or added if missing) 14:22:34 I'll have a look at the code and then comment, yes 14:23:18 other agents (OVS, DHCP, etc) keep running if no info from the server is needed 14:23:32 * mlavalle back in the meeting 14:23:35 of course, if they need to request something from the server, the operation will stop 14:23:47 don't they do periodic refresh, too? 14:24:08 they do heartbeats 14:24:15 but that doesn't stop the agent 14:24:23 if the need to do a resync, then yes 14:25:10 in any case, I would need to check the ndr agent code too 14:25:37 yes, let's dig a bit and come back to it next week or outside the meeting 14:25:48 frickler, ltomasbo thanks for the info 14:26:19 This week ykarel is the deputy, next week will be bcafarel (mtomaska can't next week) 14:26:23 yes 14:26:45 ralonsoh: ack thanks for the reminder! 14:26:59 just a heads-up: we are experiencing random issues in the CI, more frequent as before 14:27:11 we will talk about this during the CI meeting 14:27:24 I know ykarel has addressed some of them during this week 14:27:45 but we need, somehow, to stabilize the CI 14:28:21 ok, let's move to the next topic 14:28:25 #topic specs 14:28:32 the only one open (an active) is 14:28:35 #link https://review.opendev.org/c/openstack/neutron-specs/+/870030 14:28:53 I've been following the discussion between Dmitrii and lajoskatona 14:29:46 I still need to check some of the new API changes and the architecture 14:29:59 but I don't think we should merge it in this cycle 14:30:01 as I see there are open topics for those items 14:30:09 right 14:30:19 agree, but the progress and responsibility is good as I see 14:30:52 for sure and that will be attended during the next cycle without nay doubt 14:31:04 IMO, the spec is close to be ready 14:31:07 but not this week 14:31:18 so I'll add a comment today 14:31:48 that's all on this topic 14:31:56 let's move to the next 14:31:59 next one 14:32:00 # 14:32:05 (uffff) 14:32:18 #topic community_goals 14:32:25 1) Consistent and Secure Default RBAC 14:32:31 #link https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/867518 14:32:42 slaweq has done a fantastic job during this release 14:32:53 should we merge this patch ^^ now? 14:33:06 or should we wait 2 weeks? 14:33:28 slaweq ^ 14:34:11 (I think he is in other meeting at the same time) 14:34:22 how stable is the job? (or should be) 14:34:37 mostly thinking of CI stability these days :) 14:34:40 it is quite stable 14:34:54 a couple of failures in fwaas and bgp 14:35:02 but seems stable 14:35:03 I think instability is not coming from this now but from general issues in all other jobs 14:35:08 this job is pretty stable 14:35:10 the same as all other similar jobs 14:35:12 yes, I agree 14:35:12 lajoskatona, right 14:35:30 ok, then I'll change my vote on this one 14:35:31 those neutron-tempest-plugin-xxx jobs are IMHO one of the most stable once recently :) 14:35:37 yes 14:35:48 ++ 14:35:50 if it stable big +1 to have it running and seeing results already on these (before switching next cycle) 14:35:53 thx 14:36:00 perfect 14:36:16 will be good to have it now so we will also have it in stable/anthelope later :) 14:36:37 ok, I've increased the review priority 14:37:08 in next cycle we will change it to be only job running with "old default policies" as we will switch to use new ones by default 14:37:40 perfect so new ones will be the default ones 14:38:04 in 2023.2 yes 14:38:17 in 2023.1 we will have one job which will test them 14:39:02 just out of curiosity, what is the status of other projects? 14:39:35 gmann knows it the best but AFAIK nova, glance, cinder are all good to switch also 14:39:42 and keystone of course 14:39:53 I don't really know much about other projects 14:40:13 good to know that big projects are on track too 14:40:36 https://review.opendev.org/q/topic:secure-rbac 14:40:46 there is many patches in that topic :) 14:41:31 ok, in some weeks we'll swap to the new policies by default 14:41:37 yes 14:41:48 perfect, we'll check at this time then 14:41:50 thanks! 14:42:11 the next sub-topic is 14:42:13 2) Neutron client deprecation 14:42:21 lajoskatona, we have new releases, right? 14:42:24 Actually no news from my side 14:42:31 so we can continue with some patches now 14:42:49 perhaps one is htat I was waiting for SDK release, but as it is now out the req bump is still missing 14:42:56 ah right 14:43:02 (I always forget that second step) 14:43:08 never mind 14:43:26 ok, thanks lajoskatona! 14:43:27 So that is true for other libs also, we have releases finally but we still waiting for he u-c files to be upgraded 14:44:21 and the last topic for today is 14:44:25 #topic on_demand 14:44:28 lajoskatona, please 14:44:51 I just wanted to advertise the possibilty to propose forum topics for the summit 14:45:03 #link https://etherpad.opendev.org/p/neutron-bobcat-ptg 14:45:26 I think the forum schedule will be finalized mid April, and I suppose we have topropose before that 14:45:46 https://openinfra.dev/summit/vancouver-2023/forum/ 14:45:46 good point lajoskatona 14:45:50 Thanks for the reminder 14:45:53 yeah, aboyt this 14:45:55 one sec 14:46:06 maybe we can open another etherpad to collect some ideas there 14:46:08 I opened this second etherpad 14:46:09 https://etherpad.opendev.org/p/neutron-vancouver-2023 14:46:11 ^^^ 14:46:18 ralonsoh++ 14:46:33 thanks, +1 14:46:45 so that's the point: what kind of topics are we going to discuss in Vancouver (after the virtual PTG)? 14:47:23 the PTG is in March, Vancouver summit is in June (3 months) 14:47:54 ralonsoh I think that those are 2 different things: one is our "ptg" in vancouver and second are topics for the forum session, to e.g. collect feedback from operators, etc. 14:47:56 but we can use the same etherpad for now to collect ideas for both 14:48:31 right, we can open different topics in the agenda 14:48:52 agree, but I see the Vancouve mid-cycle meeting as a frre to use opportuninty for disccusion 14:49:03 in any case, I'll send a mail today with both links 14:49:13 just to make it public for everyone 14:49:13 but an agenda is good idea 14:50:01 to be honest (but this is my opinion) I don't see the relevance of the Vancouver PTG 14:51:14 anyway, as commented, I'll send a mail today with both agendas 14:51:26 +1, excluding the non-travelling part of the community isn't a good idea IMO 14:51:26 +1 14:52:19 yeah but the meeting opprotunity is good, and I hope there will be option to have it hybrid 14:52:36 I agree on this 14:53:54 ok folks, is there anything else to be discussed today? 14:54:07 (please remember the CI meeting in 6 minutes) 14:54:20 ++ 14:54:33 see You in few minutes on CI meeting 14:54:34 see you there! bye 14:54:39 #endmeeting