17:15:27 <mmichelson> #startmeeting ovn_community_development_meeting
17:15:28 <openstack> Meeting started Thu Apr 29 17:15:27 2021 UTC and is due to finish in 60 minutes.  The chair is mmichelson. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:15:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:15:31 <openstack> The meeting name has been set to 'ovn_community_development_meeting'
17:16:35 <mmichelson> I can get things started. I have my latest version of the ARP/FIP fixes up for review. I believe numans had comments on it, and I'll need to check that out. Looking into an interesting issue where it appears that expression parsing in certain cases can result in repeated conjunction actions
17:17:01 <mmichelson> Other than that, it's been code reviews from me.
17:17:22 <mmichelson> Reminder: soft freeze is in 8 days. So if you have features you're working on for 21.06.0, please get those submitted before then.
17:17:25 <mmichelson> And that's all from me.
17:17:34 <_lore_> hi all
17:17:51 <zhouhan> "expression parsing in certain cases can result in repeated conjunction actions" -> interesting...
17:18:08 <zhouhan> mmichelson: could you share if you have the test case to reproduce?
17:19:02 <mmichelson> zhouhan, sure, just a sec. I'll link a bug report, and a script
17:19:31 <zhouhan> I can go next
17:20:07 <mmichelson> zhouhan, https://bugzilla.redhat.com/show_bug.cgi?id=1955161 and https://pastebin.com/fjxEN1JU
17:20:07 <openstack> bugzilla.redhat.com bug 1955161 in OVN "OVN Expression parser generates repeated conjunctions" [Unspecified,New] - Assigned to ovnteam
17:20:13 <zhouhan> Firstly, thanks Mark Gray's review to my "flow explosion" series. I applied the first 4 patches, but left the last one since he request another look from someone else
17:20:19 <mmichelson> zhouhan, the pastebin is a script I run in the sandbox
17:20:26 <zhouhan> thanks mmichelson
17:21:10 <zhouhan> numans please also take a look if you can: #link https://patchwork.ozlabs.org/project/ovn/list/?series=241037
17:21:32 <numans> zhouhan, ack.
17:22:07 <zhouhan> I also reviewed some patches including numans's revision of split I-P logical/physical. I will still need to look at the ct.inv patch
17:22:21 <zhouhan> I need more time for the ddlog part :)
17:23:03 <mmichelson> zhouhan, I think we all have the problem right now :)
17:23:25 <zhouhan> Then I submitted some minor fixes. Thanks folks for reviewing. Most of them were merged, but there was one I thought was small but caused CI broken - the warning log change. Sorry for that.
17:24:31 <zhouhan> The warning log change triggered a discussion regarding how should we monitor DPG changes. imaximets, dumitru and I are still discussing on it.
17:25:12 <zhouhan> Also, checking with numans on the assertion problem in incremental flow installation.
17:25:18 <zhouhan> That's all I have to report
17:25:25 <mmichelson> Yeah I was curious about that assertion problem too
17:25:27 <numans> zhouhan, I just replied with my latest findings
17:25:35 <numans> rather a while back
17:25:54 <zhouhan> thanks numans I will check
17:26:13 <imaximets> zhouhan, we probably need to apply revert for the sake of CI while we're discussing on a list. To unblock ovsrobot and other patches.
17:26:49 <mmichelson> zhouhan, I made an error in the first pastebin I put up earlier. Here's a corrected version: https://pastebin.com/YvU09Yuw
17:27:06 <zhouhan> imaximets I have reverted it (before sending the reply - also mentioned in the reply :)
17:27:27 <imaximets> zhouhan, ack.  thanks. :)  I didn't read t yet.
17:28:40 <zhouhan> imaximets in fact it's in the one you already replied (the last section :D)
17:28:57 <zhouhan> mmichelson: got it, thx
17:29:20 <mmichelson> Anyone else want to share today?
17:29:25 <imaximets> zhouhan, oh, I missed the second part of that email.  Oops. :)
17:29:45 <imaximets> mmichelson, fyi, we discussed last week that the name of the meeting suddenly changed few months ago from 'ovn_community_development_discussion' to 'ovn_community_development_meeting', so the link for logs from the channel topic doesn't lead to new logs.
17:30:22 <imaximets> I have a very small update
17:30:46 <mmichelson> imaximets, whoops. That probably coincides with when I switched laptops.
17:31:44 <imaximets> It was a busy week, but I finished all the code changes for ovsdb 2-tier deployment patch set.  Adding some documentation now.  Not sure if I will send it today, but it should be ready tomorrow.
17:32:18 <zhouhan> imaximets: awesome
17:32:27 <imaximets> I also sent a small patch today to fix permissions for system test logs in GHA.
17:32:44 <imaximets> #link https://patchwork.ozlabs.org/project/ovn/patch/20210429121143.2430530-1-i.maximets@ovn.org/
17:33:00 <imaximets> currently we have no logs for system tests in artifacts.
17:33:29 <imaximets> that's it from me.
17:34:02 <mmichelson> Thanks imaximets
17:34:05 <mmichelson> Anyone else?
17:34:11 <_lore_> can I go next?
17:34:33 <mmichelson> go for it
17:34:35 <_lore_> quite fast..this week I worked on the CoPP
17:35:08 <_lore_> as discussed with dumitru, I removed the per-port metering (we have just per-dp metering now)
17:35:29 <_lore_> I addressed some comments on rfc and I added system-tests and unit-tests
17:35:57 <_lore_> I added some leftover metering for new actions (e.g. bfd or acl/lb reject)
17:36:10 <_lore_> I posted v1 just few min ago
17:37:00 <_lore_> #link http://patchwork.ozlabs.org/project/ovn/list/?series=241400
17:37:16 <mmichelson> Thanks _lore_
17:37:18 <imaximets> _lore_, out of curiosity, are you configuring burst size for meters?
17:37:25 <_lore_> moreover I posted a trivial patch to add some debug logs for bfd in ovn-controller
17:37:49 <_lore_> #link http://patchwork.ozlabs.org/project/ovn/patch/3a9ca2d94478a3b2db51d719233c9a28cafcf75b.1619268349.git.lorenzo.bianconi@redhat.com/
17:38:05 <_lore_> imaximets: it is up to the CMS to configure the meter
17:38:16 <_lore_> do you mean in system-tests?
17:38:39 <imaximets> _lore_, ack.  no, in general.
17:39:09 <_lore_> in general the workflow is:
17:39:44 <_lore_> - CMS create a given meter (rate + w or w/o burst) and associate it to given protocol per dp
17:39:47 <_lore_> like:
17:40:11 <_lore_> I create a meter to limit arps on this logical router
17:40:32 <_lore_> during testing I do not set the burst for the meter
17:40:47 <_lore_> since it is easier for testing
17:41:04 <imaximets> _lore_, ok.  Thanks for explanation.
17:41:09 <_lore_> that's all from my sice
17:41:13 <_lore_> imaximets: yw :)
17:41:31 <_lore_> btw to mark g. for the reviews
17:41:36 <_lore_> *thx
17:42:04 <mmichelson> Anyone else?
17:42:30 <numans> I can go real quick
17:42:43 <numans> I did few reviews.
17:42:54 <numans> But mostly looking into the assert issue
17:43:27 <numans> Looks like I'm able to reproduce the assert issue with my split flow table patch.
17:43:43 <numans> I'll continue debugging on it.
17:43:51 <numans> That's it from me.
17:44:22 <mmichelson> Thanks numans
17:45:13 <mmichelson> All right, thanks everyone. Have a good day!
17:45:16 <mmichelson> #endmeeting