14:00:39 #startmeeting networking 14:00:39 Meeting started Tue Nov 15 14:00:39 2022 UTC and is due to finish in 60 minutes. The chair is ralonsoh. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:39 The meeting name has been set to 'networking' 14:00:41 o/ 14:00:42 hello all 14:00:56 hi 14:01:00 o/ 14:01:42 o/ 14:02:04 slaweq, bcafarel hi! 14:02:31 o/ 14:02:39 o/ 14:02:42 ok, I think we can start 14:02:47 #topic announcements 14:02:54 Antelope / 2023.1 schedule: https://releases.openstack.org/antelope/schedule.html 14:03:02 we are in week R-18 14:03:15 that means m-1 14:03:15 https://releases.openstack.org/antelope/schedule.html#a-1 14:03:36 next one (important one) will be the first week of january 14:04:00 and as usual, the recommendation 14:04:01 OpenInfra Live event - https://openinfra.dev/live/#all-episodes 14:04:07 please check the videos 14:04:33 ok, let's move to the next topic 14:04:38 #topic bugs 14:04:52 last week the bug report was on me 14:04:55 #link https://lists.openstack.org/pipermail/openstack-discuss/2022-November/031179.html 14:05:05 I have some bugs to highlight 14:05:13 Bugs without assignee 14:05:25 #link https://bugs.launchpad.net/neutron/+bug/1995972 14:05:39 L3 router is doing schedule_routers when adding/removing external gateway 14:05:59 it is reported in Stein 14:06:13 but I think this could be reproducible in master too (that's what I think) 14:06:19 does anyone have time for this? 14:06:37 o/ 14:06:40 sorry for being late 14:06:43 np 14:07:08 ok, I'll try to take a look at this bug this week 14:07:30 next one 14:07:33 #link https://bugs.launchpad.net/neutron/+bug/1996241 14:07:39 ow-hanging-fruit 14:07:41 low-hanging-fruit 14:08:23 I also want to talk about this one 14:08:25 #link https://bugs.launchpad.net/neutron/+bug/1996421 14:08:38 I already talked about it with slaweq this morning 14:08:44 is this the low hanging fruit? 14:08:54 no, the previous one 14:09:08 ack 14:09:11 * haleyb can't get to launchpad at the moment to see either one 14:09:14 hmmmm sorry, wrong links 14:09:22 one sec 14:09:33 links are good IMO 14:09:55 both links are same 14:10:04 sorry mis read :) 14:10:06 now 14:10:07 no 14:10:13 241 and 412 14:10:16 very similar 14:10:18 so I'll repeat 14:10:30 low-hanging fruit one: https://bugs.launchpad.net/neutron/+bug/1996241 14:10:45 my old eyes saw them the same too 14:10:46 and the one I would like to talk about 14:10:50 #link https://bugs.launchpad.net/neutron/+bug/1996421 14:10:58 ^^ please open this one 14:11:16 I talked to slaweq about this one this morning 14:11:30 and in c#5 I'm confirming what slaweq stated 14:11:57 in a nutshell: neutron is working fine, the RBAC policies are working as expected 14:12:17 if the user wants to limit the "port list", he/she can use the rule 14:12:17 "get_port": "rule:admin_or_owner" 14:12:46 ^^ with that rule you'll skip the port belonging to the other project (that is what is triggering this issue) 14:12:58 makes sense to me 14:13:32 yeap, makes sense 14:13:38 ++ 14:13:40 +1 14:13:50 +1 14:13:51 thanks folks 14:14:00 just adapt the default policy to suit your needs 14:14:03 agree, messing in code with it can cause more trouble 14:14:35 exactly (although we can discuss with Nova folks enabling the possibility of using other project ports) 14:14:40 but for the next PTG 14:15:23 and this week bcafarel is the bug deputy, next week will be lajoskatona 14:15:37 well a user cannot change the policy, can they? 14:15:47 no, that should be the admin 14:16:29 ack 14:16:40 frickler: it's in the policy.yaml file so only operator can change that 14:16:44 so if a cloud deployment wants to cater for different user requirements, this might still be an issue? 14:17:28 why? if you want to allow RBACs and you want to skip this issue using horizon, you'll need this policy change 14:17:30 but I also don't think the api should be changed 14:17:46 that's the point, the API is working as expected 14:17:54 the interaction with other projects should be updated 14:18:01 some users may want to see all ports, some may only want to see those in their project 14:18:32 from the other side I think that our default behavior is good as network's owner should IMO know about ports created in his/her network 14:18:34 I would limit this issue to what is described: the horizon por list 14:18:58 anything else could be filtered using the CLI 14:19:16 then an option in horizon filter by project might be needed. ack. or check whether skyline does this better ;) 14:19:56 fine for me, then, feel free to go on 14:20:01 thanks 14:20:12 let's move to the next topic 14:20:16 #topic os-ken 14:20:21 Story board to track ryu backports: https://storyboard.openstack.org/#!/story/2009283 14:20:33 I've detected some new patches not included in os-ken 14:20:51 the first 4 are msgpack related 14:20:52 https://github.com/faucetsdn/ryu/commit/8990ed47edc82fb2b7600bf37029d6f770ef1a41 14:20:52 https://github.com/faucetsdn/ryu/commit/aa10cac1db026c8c77354f257300440b55266c9c 14:20:52 https://github.com/faucetsdn/ryu/commit/dc7aa0abef38f3974d0e6fd7d80de5f5a6a03bf0 14:20:52 https://github.com/faucetsdn/ryu/commit/045eca0592680d99b3b03c20b965b88ff126bff9 14:21:08 and there are other 2 missing too 14:21:10 https://github.com/faucetsdn/ryu/commit/5b1343d7bc13026b85514515c60dd1aeb0a711a4 14:21:14 https://github.com/faucetsdn/ryu/commit/fe1ca30fd40b215ff7e1a256b2f15757456a2229 14:21:36 today I'll open the corresponding tasks in storyboard in order to track the backports 14:21:45 cool 14:22:11 and that's all for now 14:22:13 I missed these 14:22:13 next topic 14:22:28 nah, I don't think those are affecting Neutron 14:22:31 but just in case 14:22:50 #topic specs 14:22:52 #link https://review.opendev.org/q/project:openstack%252Fneutron-specs+status:open 14:22:59 First one 14:23:01 #link https://review.opendev.org/c/openstack/neutron-specs/+/862133 14:23:12 I'm finishing the review, IMO it's +2 14:23:21 please check it, is almost done 14:23:49 next one 14:23:52 #link https://review.opendev.org/c/openstack/neutron-specs/+/857858 14:23:58 slaweq, addressed the latest comments 14:23:58 * mlavalle1 will review after the meetings 14:24:13 IMO this one is ready too 14:24:15 +1 14:24:39 the last active one is 14:24:43 #link https://review.opendev.org/c/openstack/neutron-specs/+/860859 14:24:53 I'll address today Bence's comments 14:25:05 yestarday I had a 1:1 with Sean to talk about i 14:25:12 he left the feedback on the spec 14:25:26 last patch set I only had nits and left a +1 14:25:31 so, would it be better to wait for an updated patch? 14:25:37 I'll push a new PS today to address those nits 14:25:42 rubasov, thanks 14:25:47 mlavalle1, yes 14:25:52 ack 14:26:09 and that's all in this topic 14:26:24 I would say that we have just a few specs but very well attended 14:26:28 thank you all for this 14:27:09 #topic comminity_goals 14:27:12 Consistent and Secure Default RBAC 14:27:20 2 hours ago https://review.opendev.org/c/openstack/devstack/+/861930 was merged 14:27:28 so I think we can recheck https://review.opendev.org/c/openstack/tempest/+/614484 14:27:38 all dependencies are done 14:28:03 slaweq, any update on this topic? 14:28:09 yeah, short one 14:28:25 gmann recently found an issue with those new RBAC policies in neutron 14:29:00 there is bug reported and patch https://review.opendev.org/c/openstack/neutron-lib/+/864213 14:29:14 we will need to release new neutron-lib once this will be merged 14:29:26 and that's all from me 14:29:40 I didn't yet started any other work related to this topic 14:29:55 do we need that in Zed? 14:30:51 if someone will want to use new policies in Zed then it should be there 14:31:04 but actually there is possible workaround 14:31:04 ok, so we need it in master and Zed too 14:31:12 is that now a community goal to pass tempest with new roles? 14:31:20 if "enforce_scope" will not be enabled, it should works fine 14:31:42 in devstack there is one config knob to switch both "enforce_new_defaults" and "enforce_scope" 14:31:51 lajoskatona, yes, that was the goal of this release 14:31:53 but actually we are not using scopes other than project for now 14:31:55 if I'm not wrong 14:32:18 so if I think about it a bit longer, I don't think we need to backport it to Zed 14:32:26 we can but it's not critical IMO 14:32:26 right 14:32:37 better then, just for A 14:32:47 ok, thanks 14:33:11 the next topic we have is the migration to Ubuntu 22.04 14:33:15 #link https://review.opendev.org/c/openstack/neutron/+/862492 14:33:25 still some dependencies not merged 14:33:27 and this patch 14:33:36 #link https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/857031 14:34:06 (I need to review it today) 14:35:10 the next topic, that I didn't add to the agenda, is the devstack neutron legacy removal 14:35:22 ralonsoh: it's in my todo list 14:35:24 I'll open (or re-open) a LP bug to track it and add it to the agenda 14:35:32 do you have a link? 14:35:32 but I didn't had time to get to this yet 14:35:44 ralonsoh: no, I don't have any link for that 14:35:46 np, in any case I'll create a new topic 14:35:49 I will open LP today 14:35:50 and thanks! 14:35:56 thx 14:36:35 and the last one (neither added to the agenda yet) 14:36:42 is the zuul config errors 14:36:59 I didn't start digging into this problem 14:37:09 I'll add the corresponding links and progress in the agenda 14:37:25 once we have this, we'll be able to share the work between the Neutron community 14:37:32 I can work on that too, most of them I suppose from stadiums anyway :P 14:37:38 (there are many old CI jobs failing) 14:37:47 yeah, most of them stadiums 14:37:56 lajoskatona, do you mind adding this topic to the agenda? 14:37:59 https://wiki.openstack.org/wiki/Network/Meetings#Community_Goals 14:38:04 and the corresponding links? 14:38:08 let me know if I should force merge anything 14:38:16 frickler, thanks a lot 14:38:54 and that's all in this topic, let's move to the last one 14:38:55 yeah lets do it, and track 14:38:58 although mostly making failing jobs n-v should be the better option 14:39:09 lajoskatona, thanks! 14:39:36 #topic on-demand 14:39:39 one topic only 14:39:42 #link https://review.opendev.org/c/openstack/releases/+/862937 14:39:54 that's has been reviewed by 4 people 14:39:59 Autumn cleanup time 14:40:13 and we didn't receive any feedback (good or bad) related 14:40:18 i need one review on this to fix py 3.11 14:40:18 so i think we should merge it 14:40:19 https://review.opendev.org/c/openstack/neutron/+/864448 14:40:44 haleyb, thanks! 14:40:49 one comment related to the docs bugs you mentioned earlier 14:41:02 the "normal" install guide still uses linuxbridge 14:41:03 ah this is my patch hehehe 14:41:17 ralonsoh: well, it's a follow-on with same commit message 14:41:43 so you may want to consider to switch that to ovn directly. or switch to ovs and keep ovn guide separate 14:42:11 frickler, do you have the link? 14:42:19 that is referring to linuxbridge 14:42:29 https://docs.openstack.org/neutron/latest/install/ 14:42:48 "Install and configure for *" 14:42:57 where suse might also be obsolete 14:43:40 well, at least not mantained or tested 14:43:56 and then in there e.g. https://docs.openstack.org/neutron/latest/install/controller-install-option1-ubuntu.html#configure-the-modular-layer-2-ml2-plug-in 14:43:58 but I'm reluctant to remove a whole section 14:44:43 hmm this is indeed an old manual section 14:45:00 ok, I'll check it. We have a whole section for OVN 14:45:13 maybe we can just refer to OVS on these other sections 14:45:50 https://docs.openstack.org/install-guide/overview.html#example-architecture is also very old 14:46:04 networking for pike, queens or rocky ... 14:46:47 frickler: where the repo for this? 14:47:33 I'm not sure. somewhere in openstack-manuals probably 14:47:43 duplicated in Nova and Neutron 14:47:47 (almost duplicated) 14:48:06 https://opendev.org/openstack/openstack-manuals/src/branch/master/doc/install-guide/source/overview.rst 14:48:11 thanks 14:49:20 ok, we should probably remove this from our repo 14:49:33 and just refer to the openstack-manuals only 14:49:50 the manuals refer to the neutron docs afaict 14:49:56 just to outdated ones 14:50:05 this is also duplicated in Nova 14:50:14 so I'm not sure what is the source 14:51:42 I'll check the docs this week 14:51:46 any other topic? 14:52:18 thank you all and see you in some mins in the CI meeting 14:52:22 #endmeeting