18:16:11 <mmichelson> #startmeeting ovn_community_development_meeting
18:16:12 <openstack> Meeting started Thu Feb  4 18:16:11 2021 UTC and is due to finish in 60 minutes.  The chair is mmichelson. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:16:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:16:15 <openstack> The meeting name has been set to 'ovn_community_development_meeting'
18:16:17 <blp> OK, I'll ask again then.
18:16:18 <mmichelson> OK, now :)
18:16:20 <blp> I'm trying to port new features to the ddlog northd. There's a big new feature for bfd. It has basically no documentation, and none that says what it's for. What's it for?
18:16:41 <blp> "Contains BFD parameter for ovn-controller bfd configuration." is not good documentation.
18:16:54 <_lore_> blp: sorry for that
18:17:14 <blp> If you can help me understand it, I'll add some documentation.
18:17:27 <_lore_> sure, let me open the schema, one sec
18:18:12 <_lore_> ok, do you have specific questions or do you want me to describe it?
18:18:26 <blp> As a programmer, it's really easy to document the individual trees, and the branches on the trees, and the leaves on the branches, and forget about the forest.
18:18:36 <blp> _lore_: Well, I understand BFD itself.
18:18:40 <mmichelson> blp, https://bugzilla.redhat.com/show_bug.cgi?id=1847570 was where the feature was initially requested
18:18:40 <openstack> bugzilla.redhat.com bug 1847570 in OVN "[RFE] Add support for BFD from OVN routers to other directly connected L3 devices" [Urgent,Modified] - Assigned to lorenzo.bianconi
18:19:05 <blp> _lore_: I think my question is, what parts of a network is it used to monitor? And what happens when a connection fails?
18:19:08 <mmichelson> That at least can give some background on where the feature came from.
18:19:37 <blp> Ah, the bug report will help.
18:19:53 <blp> I didn't even realize yet that it was connected to ecmp.
18:20:11 <blp> Is it only for ecmp?
18:20:12 <_lore_> ovn bfd is used to check the state of the connection with an external enitity
18:20:19 <_lore_> like external router
18:20:38 <_lore_> blp: yes, it is used for ecmp at the moment
18:20:51 <blp> OK, I understand better now.
18:20:53 <_lore_> but it can be link even to over features
18:20:59 <_lore_> sorry for that
18:21:15 <blp> OK, I'll continue reading code and try to document a little.
18:21:41 <mmichelson> Yeah we need to stop leaving documentation to the person trying to adapt the code :(
18:21:44 <blp> My report is that I'm forward porting features to ddlog northd. I also pushed a few patches that improved tests, etc.
18:21:54 <blp> and one bug report :-)
18:22:10 <_lore_> I can do it if you point me out which part to improve documentation (e.g. schema, northd code, ..)
18:22:31 <_lore_> just le me know
18:22:56 <blp> _lore_: My first plan is to add a paragraph to the schema BFD table explaining the use case and how it fits in with everything else.
18:23:07 <_lore_> ok
18:23:17 <blp> _lore_: There's also a couple of other references to bfd in ovn-nb.xml. I don't know whether they need updates.
18:23:32 <_lore_> ack
18:23:47 <blp> _lore_: NB_Global has some BFD options. I don't know whether they apply to the new bfd as well, or just to the previous use of bfd.
18:24:14 <_lore_> blp: I think just to previous implementation
18:24:17 <blp> _lore_:  And there's a statement that "OVN uses BFD to monitor gateway connectivity" later on. I think this is a separate use of BFD, and it might be worth clarifying.
18:24:39 <blp> Those are the bits I was planning to figure out and update, but if you want to do it that's even better!
18:24:59 <_lore_> yes, they are from the ovs bfd stuff
18:25:06 <_lore_> sure, I can look at them
18:25:10 <blp> Thank you!
18:25:13 <_lore_> yw :)
18:25:28 <blp> I am done with my report. Thanks everyone.
18:25:31 <dceara> hi
18:25:34 <dceara> blp: For the bug you reported, sorry about that, should be fixed now.
18:26:32 <blp> Bugs happen! Thanks for fixing it.
18:27:46 <mmichelson> OK, who'd like to go next?
18:27:55 <numans> I could go.
18:28:19 <numans> I have been working on adding mac learning support in OVN.
18:28:38 <numans> Basically OVN will learn the macs seen from the ports whose port security is disabled and 'unknown
18:28:42 <numans> address is set.
18:29:05 <numans> so that ovn doesn't flood to all unknown ports once it has learnt.
18:29:32 <numans> I'm almost done and I'll post the patches tomorrow. It would be great if they can be considered for the next release.
18:29:43 <numans> Apart from that, I have been doing some code reviews.
18:29:51 <numans> That's it from me.
18:30:33 <numans> blp, I had added some load balancer hairpin patches which would require northd-ddlog work. Have you addressed that gap ?
18:30:41 <numans> blp, Actually I had volunteered that I
18:30:46 <numans> I'll work on it.
18:31:39 <numans> Let me know. I can work on them.
18:32:04 <mmichelson> Cool. numans are you going to target 21.03 with the MAC learning patches?
18:32:10 <numans> mmichelson, Yes.
18:33:46 <mmichelson> Anyone else?
18:34:33 <dceara> I have a one liner: I posted a v2 of the series to make the flow cache size in ovn-controller configurable.  The series also fixes a couple of already existing bugs.
18:35:24 <dceara> I'll try to submit another patch to further optimize the number of openflows for hairpin tomorrow before soft freeze but I'm not sure yet if I'll make it.
18:35:29 <blp> numans: At least for now, I'm porting everything that's on master to ddlog.
18:36:08 <dceara> That's it on my side (sorry turned out to be more than one line :) )
18:36:18 <_lore_> blp: anyway for bfd most of the code is in ovn-controller
18:36:29 <mmichelson> I also have a one-liner. I put up a patch last week to create a weekly test run against OVS master. I also put up a quick patch that removes the 2020 calendar from the release documentation and adds a 2022 calendar.
18:36:36 <_lore_> northd is almost db management
18:36:46 <_lore_> iirc :)
18:36:58 <zhouhan> numans: I saw your patch for "Split logical flow and phsyical flow processing". I will review it next week (sorry for the delay)
18:37:24 <numans> zhouhan, Sure. Thanks.
18:37:34 <numans> I'll check if it requires a rebase and post v2 in case.
18:40:01 <mmichelson> Does anyone else have an update?
18:41:49 <mmichelson> All right, I guess that's all for this week. Bye everyone!
18:41:53 <blp> Bye!
18:41:54 <numans> Bye
18:41:56 <dceara> Bye!
18:42:00 <mmichelson> #endmeeting