14:00:22 <lajoskatona> #startmeeting networking 14:00:22 <opendevmeet> Meeting started Tue Mar 15 14:00:22 2022 UTC and is due to finish in 60 minutes. The chair is lajoskatona. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:22 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:22 <opendevmeet> The meeting name has been set to 'networking' 14:00:24 <lajoskatona> Hi 14:00:24 <mlavalle> o/ 14:00:31 <ralonsoh> hi 14:00:35 <obondarev> hi 14:00:41 <amotoki> hi 14:00:45 <isabek> Hi 14:01:16 <slaweq> Hi 14:02:14 <lajoskatona> We can start as I see 14:02:21 <lajoskatona> #topic Announcements 14:02:47 <lajoskatona> For one of the last times Yoga cycle calendar https://releases.openstack.org/yoga/schedule.html 14:03:09 <lajoskatona> We have a release countdown mail from elod: R-2: http://lists.openstack.org/pipermail/openstack-discuss/2022-March/027679.html 14:03:33 <lajoskatona> what is interesting that RC1 is out (20.0.0.0rc1 for Neutron) 14:04:03 <lajoskatona> we have Yoga branch, and all the fancy things for the release, great work everybody :-) 14:05:00 <ralonsoh> congrats! 14:05:06 <mlavalle> ++ 14:05:32 <obondarev> cool! 14:06:04 <slaweq> Yay! 14:06:05 <lajoskatona> We can check if wee need an RC2 14:06:21 <slaweq> Btw do we have any potential patches for RC2? 14:07:39 <lajoskatona> I remember this one: https://review.opendev.org/c/openstack/neutron/+/832662 (Fix ingress bandwidth limit in the openvswitch agent ) 14:07:55 <lajoskatona> but not sure if we really need to have RC2 for it 14:08:23 <slaweq> I don't think so 14:08:35 <slaweq> We can backport it later probably 14:08:45 <lajoskatona> ok, 14:09:28 <lajoskatona> We have ~1 week for it 14:11:05 <slaweq> Ok 14:11:14 <lajoskatona> The Zed schedule: https://releases.openstack.org/zed/schedule.html 14:11:33 <lajoskatona> And the Zed PTG etherpad: https://etherpad.opendev.org/p/neutron-zed-ptg 14:12:05 <lajoskatona> We have the Grizzly room from 13:00 - 16:00 UTC Monday-thursday 14:12:18 <lajoskatona> and 14:00-16:00 UTC for Friday 14:13:25 <slaweq> Thursday and Friday neutron sessions overlaps with TC sessions so I will probably be not available 14:13:44 <slaweq> Or not all time at least 14:14:11 <slaweq> And on Monday there is TC and PTLs session too 14:14:14 <lajoskatona> slaweq: ok, 14:14:24 <lajoskatona> the TC - PTL session will be on Monday 14:14:41 <lajoskatona> slaweq: ok, than I have good note for it :-) 14:15:21 <lajoskatona> so please add your topics to the etherpad 14:17:09 <lajoskatona> If there's no more question forr announcements, we can jump to the next topic 14:17:36 <lajoskatona> #topic Bugs 14:17:42 <lajoskatona> Report from amotoki: http://lists.openstack.org/pipermail/openstack-discuss/2022-March/027708.html 14:18:14 <lajoskatona> I saw one bug which needs attention: 14:18:20 <lajoskatona> [OVN Octavia Provider] OVN provider tests using too old version of OVN (#link https://bugs.launchpad.net/neutron/+bug/1964339 ) 14:18:46 <lajoskatona> amotoki: do you have something to highlight 14:18:48 <lajoskatona> ? 14:19:11 <amotoki> regarding the bug lajoskatona mentioned, looking at some recent job results, local.conf seems not to be confiugred properly. 14:19:26 <amotoki> it might be related to a gate configuration. 14:20:05 <amotoki> hopefully ovn-octavia-folks can look into it. 14:20:12 <ralonsoh> amotoki, I'll ping Luis 14:20:20 <amotoki> ralonsoh: thanks 14:20:39 <lajoskatona> ralonsoh, amotoki: thanks 14:21:11 <amotoki> I also added "rfe" tags to two bugs. they needs discussions for enhancements. 14:21:37 <lajoskatona> amotoki: thanks, I will add them to the drivers meeting agenda for Friday 14:22:10 <amotoki> that's all from me. 14:22:28 <lajoskatona> amotokig: thanks again 14:22:37 <lajoskatona> This week mlavalle is the deputy, and next week rubasov will be. 14:22:49 <mlavalle> on it, my fearless leader! 14:23:02 <lajoskatona> mlavalle: :-) 14:23:35 <lajoskatona> ok, next topic 14:24:09 <lajoskatona> liuyulong, liuyulong_: do you have something for L3 perhaps? 14:26:45 <mlavalle> he just dropped off 14:26:49 <lajoskatona> ok, it seems liuyulong has some network issues 14:27:01 <lajoskatona> We can move on then 14:27:17 <liuyulong_> We have L3 ndp merged. 14:27:48 <liuyulong_> #link https://review.opendev.org/q/topic:bug%252F1877301 14:28:21 <liuyulong_> The rest of the code can continue to move forward now. 14:28:38 <lajoskatona> liuyulong_: good news 14:29:49 <liuyulong_> #link https://review.opendev.org/c/openstack/neutron-specs/+/832660 I add this RA "speaker" spec for ovs agent, a simple solution, please take a look, if you guys interest. 14:30:46 <liuyulong_> an alternative for radvd. 14:31:19 <lajoskatona> liuyulong_: sure, I will check it 14:31:32 <liuyulong_> Last one, the feature of port range for port forwarding https://review.opendev.org/c/openstack/neutron/+/798961 14:32:07 <liuyulong_> A really big patch which mixed the server side and agent side works. 14:32:55 <opendevreview> Lucas Alvares Gomes proposed openstack/ovn-octavia-provider master: Fix zuul templates for functional tests https://review.opendev.org/c/openstack/ovn-octavia-provider/+/833798 14:33:22 <liuyulong_> This is very useful, since our customers are creating tons of port forwarding entries under one router, because it is 1-1 pair now. 14:33:58 <lajoskatona> liuyulong_: thanks for bringing it here, I will check it, but it is indeed huge patch 14:34:24 <liuyulong_> OK, then, no more from me now. 14:34:49 <lajoskatona> liuyulong_: thanks 14:35:06 <lajoskatona> #topic On Demand Agenda 14:35:23 <lajoskatona> I added one topic but I just chekced and I think it will be short dicussion 14:35:31 <lajoskatona> (lajoskatona): tick-tick upgrade / grenade-skip-level jobs: 14:35:43 <lajoskatona> mail from Dan: http://lists.openstack.org/pipermail/openstack-discuss/2022-March/027639.html 14:36:26 <lajoskatona> the base of this is the new release cadence resolution (#link https://governance.openstack.org/tc/resolutions/20220210-release-cadence-adjustment.html ) 14:37:14 <lajoskatona> which allows operators to upgrade to every second release (that is tick I think) 14:37:36 <slaweq> yes 14:37:47 <opendevreview> Fernando Royo proposed openstack/ovn-octavia-provider master: Fix deletion of members without subnet_id https://review.opendev.org/c/openstack/ovn-octavia-provider/+/833799 14:37:50 <ralonsoh> so everything must be compatible with version-2 ? 14:37:53 <lajoskatona> but we as community keep the middle releases also maintained (tock releases) and keep the upgrade working from tock to tick release also 14:37:58 <ralonsoh> not as is now, that should be version - 1? 14:38:18 <amotoki> yes for "tick" releases 14:38:27 <ralonsoh> understood 14:38:28 <lajoskatona> ralonsoh: yes, so for example there questions to keep RPC compatibility for N-2 14:38:38 <opendevreview> Lucas Alvares Gomes proposed openstack/ovn-octavia-provider master: Fix zuul templates for functional tests https://review.opendev.org/c/openstack/ovn-octavia-provider/+/833798 14:38:41 <opendevreview> Luis Tomas Bolivar proposed openstack/ovn-octavia-provider master: Make release job use more up-to-date ovn/ovs branches https://review.opendev.org/c/openstack/ovn-octavia-provider/+/833800 14:38:45 <slaweq> ralonsoh: if You are asking e.g. about RPC compatibility, then no 14:38:50 <slaweq> it still can be just N to N-1 14:39:08 <slaweq> tick-tick don't need to be live upgrade, according to the TC resolution 14:39:13 <lajoskatona> slaweq: ok, I thought that we go for N-2 RPC compatiblity 14:39:37 <amotoki> the resolution does not necessarily cover live upgrade from N-2 to N but perhaps it is nice to have. 14:39:39 <ralonsoh> ok, I think we'll get used to this 14:39:48 <slaweq> please check https://governance.openstack.org/tc/resolutions/20220210-release-cadence-adjustment.html#details and point 6 14:39:53 <frickler> it would be nice to support that still, even if TC doesn't require it, yes 14:40:08 <slaweq> yes, but that's not hard requirement 14:40:33 <lajoskatona> The AA release will be the first tick if I understand well 14:40:47 <frickler> it's the only thing that would make the whole effort worthwhile for a project like kolla without having to invent a new deployment type 14:41:46 <lajoskatona> so the whole thing in hard requirements will have to work first between AA and CC (whatever these will be) am I right? 14:42:09 <slaweq> lajoskatona: yes, that's my understanding too 14:42:22 <amotoki> According to the resolution, Y -> AA upgrade will be supported as experimental 14:42:36 <lajoskatona> so if we start testing it now we can have few releases to decide on these and fix it if there is a community agreement on it 14:43:15 <lajoskatona> amotoki: yes, this is why we start the testing with grenade 14:43:34 <lajoskatona> but it is more a test time , am I understand it correctly? 14:44:54 <amotoki> perhaps we need one more grenade job. 14:45:10 <lajoskatona> slaweq and I started to add it: https://review.opendev.org/c/openstack/neutron/+/833080 14:45:22 <lajoskatona> one for ovs and one for OVN 14:45:58 <lajoskatona> In the patch i added ovs job to periodic and OVN to experimental (as current grenade ovn also in experimental) 14:46:59 <lajoskatona> the ovs upgrade works ( checked it as experimental: https://zuul.opendev.org/t/openstack/build/b71ebe5826f64492a47c2a6e7feed31a 14:47:16 <lajoskatona> but OVN fails: https://zuul.opendev.org/t/openstack/build/8314af7d9257424da71c8ddf3eba7a4d 14:48:03 <slaweq> I think that OVN grenade jobs are broken for long time 14:48:21 <slaweq> and it's not related to the tick-tock upgrades probably 14:48:53 <lajoskatona> slaweq: true, the "simple" grenade also failed 14:49:42 <lajoskatona> I am not sure as OVN is the default in devstack, isn't that something that cross this community goal? 14:50:15 <ralonsoh> it should, I think 14:51:08 <lajoskatona> ralonsoh: ok, so it is something that we have to focus on, perhaps even discuss during the PTG 14:51:20 <slaweq> I think that when we were switching devstack's default we changed grenade jobs to explicitly use ml2/ovs 14:51:28 <slaweq> and we never came back to update it 14:51:36 <slaweq> but I may be wrong here, it should be checked 14:52:00 <lajoskatona> slaweq: ok, thanks, I will check history :-) 14:52:05 <slaweq> ++ 14:52:20 <frickler> having a job testing ovs => ovn migration would also be nice, btw 14:52:28 <lajoskatona> ok, thanks, that's it from me 14:52:56 <ralonsoh> frickler, that's not trivial 14:52:58 <slaweq> frickler: but currently such migration is only supported for the tripleo based deployment 14:52:59 <lajoskatona> jlibosva presented it during the last PTG 14:53:03 <ralonsoh> we have templates for tripleo 14:53:04 <slaweq> not in devstack 14:53:54 <lajoskatona> I will add this topic to the PTG etherpad 14:53:58 <frickler> would be really nice to also support the rest of the world 14:54:00 <frickler> ack 14:54:07 <mlavalle1> Good idea 14:54:46 <lajoskatona> If there is nothing more to discuss, we can close the meeting 14:56:01 <lajoskatona> #endmeeting