14:00:08 <ralonsoh> #startmeeting networking
14:00:08 <opendevmeet> Meeting started Tue Jun  6 14:00:08 2023 UTC and is due to finish in 60 minutes.  The chair is ralonsoh. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:08 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:08 <opendevmeet> The meeting name has been set to 'networking'
14:00:11 <ralonsoh> hello all
14:00:13 <obondarev> hi
14:00:13 <haleyb> o/
14:00:21 <isabek> Hi
14:00:26 <frickler> \o
14:00:29 <ralonsoh> slaweq and lajoskatona won't attend today
14:00:57 <ralonsoh> ok, let's start, we have quorum
14:01:08 <ralonsoh> #topic announcements
14:01:16 <ralonsoh> #link https://releases.openstack.org/bobcat/schedule.html
14:01:24 <ralonsoh> next week is the Vancouver PTG
14:01:28 <ralonsoh> this is the schedule
14:01:35 <ralonsoh> #link Ping list: bcafarel, elvira, frickler, mlavalle, mtomaska, obondarev, slawek, tobias-urdin, ykarel, lajoskatona, jlibosva, averdagu, amotoki
14:01:37 <ralonsoh> sorry
14:01:44 <mlavalle> o/
14:01:49 <mtomaska1> o/ :)
14:01:51 <ralonsoh> #link https://vancouver2023.openinfra.dev/a/schedule
14:01:54 <bcafarel> this is to see if we react automatically :)
14:02:00 <ykarel> :)
14:02:07 <ralonsoh> yeah... sorry, bad copy/paste
14:02:23 <ralonsoh> in this agenda you can filter by day and topic
14:02:40 <ralonsoh> btw, I don't think we'll be able to have video meetings
14:02:48 <ralonsoh> but we'll try using out own laptops
14:03:12 <ralonsoh> ah, because of the PTG, there will be no meetings (team, CI and drivers)
14:03:30 <ralonsoh> and that's all I have in this topic
14:03:35 <ralonsoh> something I'm missing?
14:03:37 <opendevreview> Arnaud Morin proposed openstack/neutron master: Do not dispose local_vlan_hints  https://review.opendev.org/c/openstack/neutron/+/880334
14:04:00 <mlavalle> no CI meeting today either
14:04:14 <ralonsoh> yes, I had this for the last minute topic
14:04:39 <ralonsoh> ok, let's move on
14:04:43 <ralonsoh> #topic bugs
14:04:58 <ralonsoh> last week report is from mtomaska1
14:05:02 <ralonsoh> #link https://lists.openstack.org/pipermail/openstack-discuss/2023-June/033962.html
14:05:10 <ralonsoh> all bugs are assigned (that's good)
14:05:14 <ralonsoh> there are two pending ones
14:05:22 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/2022360
14:05:27 <ralonsoh> this one was already assigned
14:05:31 <ralonsoh> (but not in the LP bug)
14:05:48 <ralonsoh> and there is a patch: https://review.opendev.org/c/openstack/neutron/+/883235 (reviews are welcome)
14:06:05 <ralonsoh> the last one is
14:06:07 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/2022058
14:06:45 <ralonsoh> if I'm not wrong, what is expected from this bug is to put the general ML2/OVN config option for FIP distributed in the router dict
14:06:50 <ralonsoh> so n-d-r can read it
14:07:13 <ralonsoh> in this case the router resource will be different to the ML2/OVS one, because this will be a read-only parameter
14:07:25 <dmitriis> o/
14:07:29 <dmitriis> I guess I can comment
14:07:41 <ralonsoh> dmitriis, but ^^ is this what you expected?
14:08:03 <ralonsoh> at least from Neutron
14:08:12 <dmitriis> That's one way of doing it, I thought of an alternative as well
14:08:20 <ralonsoh> please, go on
14:08:36 <dmitriis> where we'd allow the distributed attribute to turn the distributed fips for OVN routers on a per-router basis
14:09:06 <dmitriis> i.e. the addition of `external_mac` to OVN would depend on whether the global config option is enabled and whether `distributed == True`
14:09:45 <ralonsoh> yeah, this is actually the "proper" way to implement this
14:09:51 <ralonsoh> but of course this is much more complex
14:10:19 <ralonsoh> and, IMO, this second alternative could be considered as a RFE
14:10:35 <dmitriis> the other part of the bug is the NDR part
14:11:01 <dmitriis> because currently /32 routes for `distributed == True` routers are created on the basis of floatingip_agent_gateway ports
14:11:14 <dmitriis> but those aren't created with ML2/OVN
14:11:49 <dmitriis> I'm not sure if this was a design decision to conserve IP space or not (compared to ML2/ovs there is an IP address per node per external network consumed for those)
14:12:17 <dmitriis> but they also allow for southbound routing to be done directly to a compute node with /32 routes
14:12:36 <dmitriis> so there are 2 things to consider:
14:13:13 <dmitriis> 1. if there are no floatingip_agent_gateway ports, we should probably have a fallback to /32 routes with the centralized gateway port IP as a next-hop
14:13:39 <dmitriis> that way we fix the issue with no routes being advertised for OVN routers that have `distributed == True`
14:14:09 <dmitriis> 2. consider if we need to optionally enable the creation of floatingip_agent_gateway ports with ML2/OVN akin to ML2/OVS for routing purposes
14:14:30 <dmitriis> (1) is probably a bug-fix type of a change. (2) is more of an RFE
14:15:09 <dmitriis> I hope it makes sense. There are some links to the code in the bug which show what I mean (just in case)
14:15:13 <ralonsoh> ok, I'll take this conversation out of this meeting but I would like to understand why these "floatingip_agent_gateway" ports are needed
14:15:27 <ralonsoh> I'll ping you after this meeting just to retrieve more info
14:15:33 <dmitriis> ok, sounds good
14:15:45 <ralonsoh> and it could be useful to have these thoughs in the LP
14:16:24 <ralonsoh> ok, any other bug you want to discuss?
14:16:41 <dmitriis> no bugs, just a review at the end of the agenda
14:16:59 <ralonsoh> sure, please add it in the on demand section
14:17:22 <ralonsoh> the week bcafarel is the bug deputy, next week is lajoslatona
14:17:43 <ralonsoh> because there is nothing related to os-ken, let's jump to the next section
14:17:47 <dmitriis> ralonsoh: ack, it's the last one here (https://wiki.openstack.org/wiki/Network/Meetings#On_Demand_Agenda)
14:17:47 <ralonsoh> #topic specs
14:18:04 <ralonsoh> dmitriis, don't worry, I'll ask for this at the end
14:18:10 <dmitriis> ack
14:18:17 <ralonsoh> we have one active spec: https://review.opendev.org/q/project:openstack%252Fneutron-specs+status:open
14:18:25 <ralonsoh> #link https://review.opendev.org/c/openstack/neutron-specs/+/885324
14:18:29 <ralonsoh> Multiple backend support for L3 routers in ML2/OVN
14:18:39 <ralonsoh> this RFE was discussed (and approved) last Friday
14:18:52 <ralonsoh> so, as usual, reviews are more than welcome
14:19:07 <ralonsoh> (when there is something in the spec)
14:19:12 <mlavalle> I'm writing the spec, that is WIP in gerrit for now
14:19:19 <ralonsoh> sure
14:19:58 <ralonsoh> something else in this topic?
14:20:20 <ralonsoh> ok, I'll skip next one
14:20:24 <ralonsoh> because
14:20:37 <ralonsoh> 1) slaweq is not here and there are no bugs related to sRBAC
14:21:00 <ralonsoh> 2) lajos is not here and there are no new patches related to the client migration
14:21:13 <ralonsoh> so this is going to be a very very fast meeting
14:21:23 <ralonsoh> #topic on_demand
14:21:25 <ralonsoh> I have one
14:21:49 <ralonsoh> as commented, we don't have the CI meeting during the next 2 weeks
14:22:13 <ralonsoh> but, of course, if there is something urgent/blocker (tox5, sqlalchemy3.0 or similar)
14:22:25 <ralonsoh> please ping any core in this channel and open a critical LP bug
14:22:36 <ralonsoh> dmitriis, please
14:22:43 <frickler> don't make jokes like that, we don't even have sqla2.0 running
14:22:53 <ralonsoh> the CI is now passing (for Neutron)
14:23:09 <ralonsoh> so at least we are good (I think so)
14:23:20 <frickler> neutron yes, openstack ... meh
14:23:26 <ralonsoh> yeah...
14:23:49 <dmitriis> ralonsoh: are you asking about my item?
14:23:53 <ralonsoh> sure
14:23:58 <dmitriis> ah, ok
14:24:23 <dmitriis> so I have a review for the multiple gateway support spec implementation
14:24:32 <dmitriis> https://review.opendev.org/c/openstack/neutron/+/873593
14:24:49 <dmitriis> just looking for someone to look at it to have 1 more core review on it
14:25:19 <dmitriis> I understand it's quite a large one to review
14:25:31 <dmitriis> but it's quite self-contained as far as the functionality goes
14:26:01 <ralonsoh> for sure, I'll check it tomorrow morning, you have my word
14:26:14 <fnordahl> \o/
14:26:20 <fnordahl> there is of course a series of reviews stacked on top of it, and they are smaller
14:26:29 <dmitriis> ralonsoh: tyvm, just trying to get it in early to avoid rushing close to the change freeze
14:26:52 <dmitriis> that's all I had for this
14:26:58 <ralonsoh> yes and you have worked hard in this one, for sure
14:27:01 <ralonsoh> dmitriis, thanks
14:27:21 <frickler> are there some docs for it somewhere? the reno is pretty terse
14:28:11 <dmitriis> We'll add some as a follow-up, just want to make sure we have the functionality set in place
14:28:40 <dmitriis> Some of the functionality is documented in the API ref change
14:28:53 <dmitriis> but definitely deserves a separate doc entry, I agree
14:29:03 <ralonsoh> thank you
14:29:35 <frickler> might help reviews to see some example scenario to compare against
14:29:50 <frickler> but I'll check the API ref change
14:30:09 <dmitriis> frickler: ack
14:31:24 <ralonsoh> any other topic?
14:31:45 <dmitriis> not from me
14:32:01 <ralonsoh> thank you all for attending
14:32:04 <ralonsoh> #endmeeting