17:23:35 <imaximets> #startmeeting ovn-community-development-discussion
17:23:35 <openstack> Meeting started Thu Jun  4 17:23:35 2020 UTC and is due to finish in 60 minutes.  The chair is imaximets. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:23:36 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:23:38 <openstack> The meeting name has been set to 'ovn_community_development_discussion'
17:24:04 <imaximets> Hi, everyone.  There is no much participants today.
17:24:25 <imaximets> But it's time for the OVN meeting, so let's start.
17:24:35 <panda> I count more as a lurker than a participant
17:25:03 <dceara> Hi! Me too, nothing much I have to share this week
17:25:40 <_lore_> hi all
17:25:45 <_lore_> me too as well
17:27:03 <imaximets> Who wants to go first anyway?
17:27:57 <imaximets> zhouhan_, are you here today?
17:28:17 <imaximets> :)
17:28:27 <panda> I have a one liner. I'm looking at something to review in the ML, and I think I completed my onboarding. I can work on OVN only one day per week right now, but tomorro I'll try to move forward the patch on router-preference on non-periodic RA.
17:28:46 <numans> Hello everyone.
17:28:59 <numans> (Joining late)
17:28:59 <panda> I think I'll post the onboarding experience somewhere. Whoops two lines.
17:29:08 <numans> panda, hope my reply made sense ?
17:29:19 <numans> panda, and welcome to OVN :)
17:29:33 <zhouhan> imaximets: sorry, in parallel meeting. just one moment
17:29:34 <panda> numans: yes, it did, but I still have other question, I'll look more again tomorrow
17:29:41 <numans> ack.
17:29:49 <imaximets> zhouhan, sure. np.
17:30:23 <imaximets> panda, thanks for update.
17:30:33 <imaximets> who wants to go next?
17:30:59 <zhouhan> I can go quickly
17:30:59 <numans> I can go real quick.
17:31:03 <numans> zhouhan, go ahead
17:31:04 <numans> :)
17:31:08 <zhouhan> numans: thx
17:31:33 <zhouhan> I was mainly on code reviews, and I think I finished reviewing the I-P v5
17:32:00 <zhouhan> I am also working on the flow explosion solution
17:32:37 <zhouhan> There is another problem regarding flow explosion brought up by Girish, and I proposed a solution, but no time to work on that yet.
17:33:30 <zhouhan> imaximets: did you get some time to review the dp_hash patch?
17:33:55 <numans> zhouhan, you mean v9 ?
17:34:15 <numans> because tht was the last version posted.
17:34:16 <zhouhan> numans: Ah, sorry, typo. Yes, v9
17:34:33 <numans> no worries. Just checking in case you were reviewing old one.
17:34:51 <zhouhan> v9 was in my brain but I typed just v5, I don't know why :D
17:34:55 <imaximets> zhouhan, I only had a quick glance.  I'm a bit concerned about passing hash computed in userspace to kernel.  blp said same couple of weeks ago.  I'll reply to the patch by the end of this week.
17:37:25 <zhouhan> imaximets: thx for the review. Yes there were 2 options discussed. I explained in the email why I choose this option in the end.
17:38:07 <zhouhan> imaximets: I will wait for your reply :)
17:38:19 <imaximets> zhouhan, ack. :)
17:38:27 <zhouhan> That's my update. numans please go ahead
17:38:42 <numans> Thanks.
17:39:23 <numans> I've been working on the v10 of I-P. It's almost ready. But before posting the series, I'll be doing some scale testing internally to be sure that there are no regressions.
17:39:33 <numans> So please expect some delay.
17:39:49 <numans> zhouhan and dceara , Thanks a ton for the reviews so far
17:40:04 <zhouhan> numans: np
17:40:14 <numans> I also posted another patch for ovsdb IDL.
17:40:27 <dceara> numans, np
17:40:39 <numans> thanks for the review zhouhan. I'll submit v2 soon addressing your comment.
17:40:46 <numans> That's it from me.
17:40:57 <zhouhan> v10? :)
17:41:26 <zhouhan> oh, I see, you mean the IDL v2
17:42:34 <numans> zhouhan, Yes. the IDL v2.
17:42:41 <zhouhan> p.s. the new flow explosion problem brought up by Girish is: #link https://mail.openvswitch.org/pipermail/ovs-discuss/2020-June/050187.html
17:43:15 <imaximets> I have a quick question.
17:43:26 <imaximets> I'd like to ask about backporting of the patch "raft: Avoid sending equal snapshots.".  What everyone thinks about it.  It might be good to have it at least in 2.13.
17:43:45 <numans> zhouhan, I saw that email. Is it a regression ? Any idea ?
17:44:14 <numans> imaximets, I'm fine. But probably zhouhan can answer better.
17:45:32 <imaximets> It's always unclear what is something a bug fix or an enhancement while talking about scale issues.
17:45:47 <imaximets> zhouhan, what do you think?
17:45:52 <zhouhan> numans: it's not a regression. It is from day one that the rules were added per router port, which is unnecessary
17:46:52 <zhouhan> imaximets: I think it is better to be backported if it is a big impact on scale
17:46:57 <numans> zhouhan, Ok. thanks. I was worried if I'm the culprit. Because I had a patch to address one issue.
17:47:19 <imaximets> zhouhan, ok. thanks.
17:47:41 <panda> numans: quick noob question. Are the scale tests in the repo ?
17:48:06 <imaximets> In this case I'm going to backport this patch to 2.13.
17:48:21 <numans> panda, yes. https://github.com/ovn-org/ovn-scale-test
17:48:31 <panda> numans: ah in the other :) thanks.
17:48:51 <imaximets> that's it from me.
17:49:05 <numans> panda, there's another repo which internally uses ovn-scale-test. That's still not in official ovn-org repo. For now its here - https://github.com/dceara/ovn-heater
17:49:12 <dceara> panda, there's also https://github.com/dceara/ovn-heater which runs ovn-scale-test on an ovn-fake-multinode virtual cluster
17:49:25 <dceara> panda, numans beat me to it :)
17:49:33 <panda> wow :) ok.
17:50:40 <imaximets> OK.  Who wants to go next?
17:52:23 <imaximets> Last call for updates. :)
17:52:55 <numans> Looks like we are done.
17:53:07 <imaximets> ok.  Let's call it a meeting.
17:53:15 <imaximets> Thanks, everyone.
17:53:31 <panda> bye
17:53:34 <imaximets> #endmeeting