18:14:59 <mmichelson> #startmeeting ovn_community_development_meeting 18:15:00 <openstack> Meeting started Thu Feb 11 18:14:59 2021 UTC and is due to finish in 60 minutes. The chair is mmichelson. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:15:01 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:15:03 <openstack> The meeting name has been set to 'ovn_community_development_meeting' 18:15:21 <numans> Hello 18:15:27 <mmichelson> I can lead things off 18:15:44 <mmichelson> I've mostly been focused on reviews this week. We are in the soft freeze period until the end of next week. 18:15:46 <_lore_> hi all 18:15:58 <mmichelson> I also got some feedback from Han on the ovn-controller refactor doc. 18:16:07 <mmichelson> And so there's been some good conversation there. 18:16:36 <mmichelson> I also merged the submodule patch and weekly OVS master CI run patch. 18:16:50 <mmichelson> numans found a problem in the submodule patch. I have another patch up that fixes it. 18:16:54 <mmichelson> And that's all from me. 18:17:04 <numans> mmichelson, thanks for the fix. It works for me. 18:17:25 <_lore_> can I go next? quite fast 18:17:41 <mmichelson> _lore_, go right ahead 18:17:50 <_lore_> this week I worked on bug fixing, posted some patches 18:18:02 <_lore_> thx mmichelson and numans for reviews 18:18:21 <_lore_> moreover I posted a RFC patch and I would like to get your opinion about it 18:18:53 <_lore_> discussion with openshift folks, they rised the concern automatic routes must have highest prio 18:19:08 <_lore_> automatic --> routes added assigning IP to the interfaces 18:19:31 <_lore_> so it would be not possible to overwirte this with static routes 18:19:45 <_lore_> I remember we discussed with zhouhan about it 18:20:28 <_lore_> so just wondering if the rfc patch is fine or we should keep the capability to overwrite them 18:20:40 <numans> I guess we calculate the priority for the flow in lr_in_ip_routing based on the network/prefix of the route. 18:20:50 <_lore_> http://patchwork.ozlabs.org/project/ovn/patch/e590661b77d0b57d2d115b0595d8ab7167dbf129.1612995252.git.lorenzo.bianconi@redhat.com/ 18:21:04 <_lore_> numans: yes, for longest prefix match 18:21:26 <numans> _lore_, ok. I need to check the patch to comment further. 18:21:36 <mmichelson> Yeah I haven't seen the patch yet either 18:21:44 <_lore_> what osh folks are suggesting is for e.g IPv4 to use like 33 for automatic IPv4 routes 18:21:51 <numans> I think it should be fine if we give preference to the automatic routes over static routes. 18:22:16 <zhouhan> _lore_: I don't remember discussing the same thing before (probably there was something related but different?). But I will look at this if needed. 18:22:51 <_lore_> zhouhan: numans: mmichelson: it seems to me a hack to overwirte automatic routes, but nowdays is possible IIUC 18:23:04 <_lore_> so I was just wondering if the idea in the patch is fine 18:23:20 <_lore_> ovn tests are not broken by it 18:23:34 <_lore_> anyway we can disucss on the ml 18:23:41 <numans> sounds good. 18:23:47 <_lore_> that's all from my side, thx 18:24:41 <_lore_> ah, I forgot.. 18:24:51 <_lore_> now I am working on IP to add some coverage counters 18:25:10 <_lore_> now that's all :) 18:25:12 <numans> IP as in IPv4/IPv6 ? 18:25:16 <_lore_> I-P 18:25:21 <_lore_> ;) 18:25:22 <mmichelson> incremental processing :) 18:25:25 <_lore_> yes 18:26:10 <numans> I can go real quick. 18:26:24 <numans> I spent much of the time in the reviews. 18:26:42 <numans> Also submitted a patch series to add mac learning support 18:27:15 <numans> and a small patch related to load balancer src nat ip 18:27:33 <numans> I saw some comments from mmichelson for the 1st patch of the mac learning series. 18:27:52 <numans> thanks for it. I'll wait for further comments from you before addressing them. 18:27:53 <mmichelson> Confession: I had to look up what "FDB" stands for 18:28:27 <numans> mmichelson, Ok :). I used that term since its used by OVS to for mac learning. 18:28:40 <numans> I can change to a better name if its confusing or not clear. 18:28:59 <mmichelson> numans, no it's fine. I just is because of my lack of insider knowledge for switches :) 18:29:06 <mmichelson> s/I/It/ 18:29:29 <numans> ok :) 18:29:39 <_lore_> mmichelson: numans: IIRC even linux kernel relies on the same name for bridge table 18:29:51 <_lore_> so I guess it is ok :) 18:29:53 <blp> I have a short report when there's a gap. 18:30:00 <numans> I think that's it from me. 18:30:06 <numans> blp, go ahead plz. 18:30:20 <blp> There's also a fairly common term "RIB" for Routing Information Base that you might run into occasionally. 18:30:33 <blp> OK, my report is that I've now caught up again to the state of master 18:30:49 <blp> and I'm debugging what I think is one remaining bug in my ddlog implementation. 18:31:03 <blp> When that's fixed, I'll post v10. 18:31:07 <numans> cool. 18:31:19 <blp> Oh, I'm not sure whether I mentioned this here before 18:31:30 <blp> but some people were having linking problems depending on some kind of avx thing in OVS. 18:31:33 <blp> I found and fixed that problem. 18:31:49 <numans> blp, that's cool. 18:31:57 <blp> It was related to the libtool horror show. 18:32:00 <blp> ugh, libtooll 18:32:03 <numans> I saw your reply on it 18:32:29 <blp> I think that's my whole report, unless there are questions 18:32:35 <numans> blp, got a chance to try out the script which I shared sometime back ? 18:32:56 <blp> numans: I still have it flagged in my mail client, and I'm going to try it, but I haven't yet. 18:33:03 <numans> ok. 18:33:26 <numans> thanks. 18:33:53 <numans> blp, I'm fine if you want to test it after you post v10. 18:33:56 <blp> numans: thank you for writing it up 18:34:03 <mmichelson> The timing for v10 should be about perfect 18:34:14 <blp> mmichelson: great 18:34:15 <mmichelson> Because it'll happen right around when we branch 21.03 18:34:15 <numans> blp, welcome. 18:34:42 <mmichelson> All right, anyone else? 18:35:13 <imaximets> I have nothing OVN-related. :) 18:35:52 <imaximets> OVS 2.15 release is on Monday with column diff support and other cool stuff. 18:36:02 <imaximets> That's it. 18:36:05 <numans> cool. 18:36:07 <mmichelson> nice! 18:37:19 <mmichelson> OK, anybody else have a report? 18:38:06 <zhouhan> I spent sometime on the ovn-controller refactor doc written by Mark 18:38:32 <zhouhan> Maybe we can continue discussing in ML 18:38:57 <mmichelson> zhouhan, yes, that would be nice. I appreciate your input! 18:39:27 <zhouhan> It seems to be a big and important discussion. Thanks Mark for writing it up! Hope more folks join and share thoughts :) 18:41:13 <zhouhan> (I don't have anything else to report :) 18:41:34 <numans> One attention on a PR posted by Girish 18:41:42 <numans> on the debian packaging 18:41:55 <numans> https://github.com/ovn-org/ovn/pull/76 18:42:16 <numans> I'm planning to test that PR and if it works, I'm planning to apply it. 18:42:38 <numans> I'm not a debian user or developer. If some one wants to review, then please do so. 18:43:12 <numans> or if any objections on the PR, please comment on it. 18:43:26 <zhouhan> numans: thx, I can have a test as well. great to have debian back :) 18:43:26 <numans> (That's it again from me) 18:43:35 <numans> zhouhan++ 18:43:38 <numans> that would be great. 18:44:16 <mmichelson> Great to hear from everyone! Have a good day! 18:44:21 <mmichelson> #endmeeting