20:01:20 #startmeeting Octavia 20:01:21 Meeting started Wed Jan 23 20:01:20 2019 UTC and is due to finish in 60 minutes. The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:01:22 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:01:24 The meeting name has been set to 'octavia' 20:01:24 Hi all 20:01:43 Sorry for being a minute late. Lots of announcements today 20:02:02 hi 20:02:04 o/ 20:02:11 #topic Announcements 20:02:20 Denver Summit call-for-papers is now open, deadline Today! 20:02:28 #link https://www.openstack.org/summit/denver-2019/call-for-presentations/ 20:02:39 Today is the last day to submit presentations for the summit. 20:02:56 We have new foundation directors 20:03:04 #link http://lists.openstack.org/pipermail/openstack-discuss/2019-January/001962.html 20:03:27 The bylaw change passed as well. 20:03:41 The OpenStack Diversity Survey is open 20:03:48 #link http://lists.openstack.org/pipermail/openstack-discuss/2019-January/001925.html 20:04:01 And it is time for Train TC Elections 20:04:08 #link http://lists.openstack.org/pipermail/openstack-discuss/2019-January/001829.html 20:04:53 Ok, that is all I have. 20:04:59 Any other announcements today? 20:05:47 Ok, onward! 20:05:57 #topic Brief progress reports / bugs needing review 20:06:16 Well, I have been a bit busy...... 20:06:22 #link https://etherpad.openstack.org/p/octavia-priority-reviews 20:06:38 o/ 20:06:44 It's been nice to be able to work on code. 20:07:07 * rm_work quietly watches johnsom have a meeting with himself 20:07:36 The flavors chain is basically done. I'm just wrapping up the final patch that addresses and issue raise by Yang JianFeng 20:08:05 Net split or just the only one giving an update???? 20:08:43 At least three people said they joined, so I'm not totally alone here. 20:09:07 You're not. Just no updates on my end 20:09:29 And boy do I have a deal for you.... Come on down to Gerrit land, review some flavor patches, and when you get to the end of the chain you get a BONUS! 20:09:45 I'm adding the ability to live update the amphora-agent configuration. 20:10:00 :P 20:10:07 oh, nice 20:10:11 The following amphora agent configuration options can be updated: 20:10:11 heartbeat_key 20:10:11 controller_ip_port_list 20:10:11 heartbeat_interval 20:10:11 loadbalancer_topology 20:10:40 live update would be great! it would allow to update the controller ip list, for example 20:11:05 That last one doesn't change the topology, just the topology the agent thinks it is part of. So, not useful from an admin perspective, but useful for internal spares pool. 20:11:26 Yep, super handy. 20:12:15 I'm just finishing testing the amphora side of that, then I will add a patch that creates an admin API for it. 20:13:24 I hadn't planned to work on that, but I needed it for the lb topology flavor support when getting an amp from the spares pool. So I decided it was time to do it. 20:14:00 I have also been spending some time on getting the gates cleaned up again. I think we made good progress there, so should be in better shape going forward. 20:14:14 There are still some coverage gate patches open for review though. 20:14:40 #link https://review.openstack.org/629955 20:14:49 #link https://review.openstack.org/629953 20:15:07 low hanging reviews.... 20:15:43 Any other updates today? 20:16:22 My plan going forward is to finish up this agent config update work, then dig into reviewing the TLS patches. 20:16:42 Also look at putting a heartbeat key rotation system in place 20:16:49 I worked a bit on a fedora 28 job. some issues in DIB and RPM. it is important for us, it's the closest we have now to RHEL8 that we can test against 20:17:00 Also want to look at adding TLS protocol/cipher section to the API. 20:17:50 I also spent some time with Ian looking at the centos slowness. It sounds like Ian may have found the cause in some systemd patches. 20:18:29 #link https://etherpad.openstack.org/p/dib-centos-slow 20:18:39 #link https://bugs.centos.org/view.php?id=15735 20:19:28 oh, neat 20:19:31 Yep 20:19:47 Any other updates / plans? 20:20:43 Sigh, ok.... 20:21:18 The review list is back up to 45 patches. A lot of them are flavors as I tried to break it down into smaller individual patches. 20:21:46 #topic Open Discussion 20:21:57 Any other topics today? 20:22:12 my next tasks are around F28, ovn driver, octavia-lib packaging, RHEL8 dib support 20:22:22 OSP 13 support for active/standby ????? grin 20:22:35 lol 20:22:41 Ah, yeah, I should get back on the octavia-lib patch... 20:22:49 It has some bugs 20:22:49 nmagnezi proposed backport to stable/queens a few hours ago 20:23:07 #link https://review.openstack.org/#/c/632763/ 20:23:14 I saw that. It looks like it needs some work 20:23:15 we might have to move some other stuff over to octavia-lib for providers 20:23:20 yup 20:23:41 cgoncalves Like? 20:23:43 Reedip asked me about octavia.common.utils 20:24:00 * johnsom notes there are four open octavia-lib related patches open 20:25:27 Hmm, maybe a sub section of that. Obviously wouldn't do "get_network_driver" 20:25:36 johnsom, speaking of act-stdby, it will not allow us to test against released versions 20:25:49 because it depends on amphora stats >= stein 20:26:09 Correct, the tempest test will not. 20:27:03 That test is still open too. (merge conflict it looks like) 20:27:09 sad... I wish there could be some other way of testing 20:27:25 I guess we could to a dumber test, that blindly kills things and show that traffic still flows. 20:27:45 +1 20:28:58 I would still keep this test though, as it proves act/stdby is working as expected. 20:29:21 agreed 20:30:05 Is that a gate to getting OSP 13 to support act/stdby? 20:30:16 Or could that happen with manual testing 20:30:43 tricky question ;) 20:31:30 let's focus in getting the backport in for stable/queens for now :P 20:31:44 +1 20:32:11 hmmm how long do we have until the next RC? 20:32:12 Ok. Can't say I didn't try to get it in writing 20:32:25 first one for this cycle, i guess 20:32:58 We are working towards MS3 March 4th. Followed by RC1 March 18th 20:33:04 i would maybe like to do the thing where we minimally review the flavors chain for anything glaringly obvious, then focus on improving testing and doing bugfixes until the end of the cycle to perfect it 20:33:14 rather than wasting a ton of time being super pedantic on the 10+ patch chain 20:33:15 lol 20:33:22 we have enough time then i think 20:33:24 for that to be feasible 20:33:28 johnsom, sneaky you 20:33:32 this wouldn't be a last minute merge 20:33:54 15+ chain... lol 20:34:05 as long as it doesn't break existing functionality, which our tests at this point pretty confidently prove it does not 20:34:17 super happy about the state of our testing now 20:34:42 Yeah, plus I have full test coverage for all of the flavors stuff. Tempest API and scenario included 20:35:02 sweet 20:35:15 soooo i'm gonna go scan them, and then +2 unless i see something really dumb <_< 20:35:28 i'm pretty stuck internally right now anyway on a dumb pip issue 20:36:31 Ok. 20:36:53 I'm hoping to pivot to reviewing the TLS patches so that they don't become a last minute push either. 20:40:11 There are great features in there that I want to get into Stein. 20:41:07 Any other topics this week? 20:41:31 BTW, there is OSC CLI support for flavors too which makes kicking the tires super easy 20:42:55 Either everyone went to sleep or it's a net split. 20:43:39 If there aren't other topics, I will get back to working on the amp-agent update patches. (pretty excited about that actually. We have wanted them for years) 20:43:40 sleep :P{ 20:43:48 ima go +2 those patches now 20:43:52 bbl 20:44:00 * cgoncalves is splitting is time between dinner and meeting 20:44:03 lol 20:44:08 s/is/his 20:44:13 Yeah, I just finished my lunch too 20:44:40 Ok. With that, thanks folks! Don't forget to come on down to Gerrit town for some sweet review deals! 20:45:00 #endmeeting