20:01:00 #startmeeting Octavia 20:01:01 Meeting started Wed Jul 18 20:01:00 2018 UTC and is due to finish in 60 minutes. The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:01:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:01:05 The meeting name has been set to 'octavia' 20:01:11 Hi all 20:01:41 #topic Announcements 20:01:56 The Rocky priority bugs list is still going on.... 20:02:01 hi 20:02:03 #link https://etherpad.openstack.org/p/octavia-priority-reviews 20:02:10 please help with reviews! 20:02:36 Also, our python-octaviaclient release deadline is next week: July 26th 20:02:48 We have two patches we want in, backup members and UDP 20:03:12 we've made really good progress. only one patch in WIP. 38 tracked patches merged 20:03:18 I have used the UDP will success, just need to finish reviewing it 20:03:29 Yeah, we have 20:03:42 Some big stuff still open though 20:03:52 should we merge the UDP client patch before the server-side work? 20:04:02 Any other announcements? 20:04:49 Well, if we are confident we can get UDP in, I think so. The user will get a sane error if the API doesn't support UDP 20:05:24 ok. good enough 20:05:44 They should get a 400 with protocol not supported 20:06:05 #topic Brief progress reports / bugs needing review 20:06:25 I have been doing some reviews, but have been mostly focused on an internal fire drill. 20:06:52 O/ 20:07:19 If I can get something working today, I should be in good shape to focus upstream again. 20:07:54 I have an environment setup to address the issue cgoncalves noted on the dual down amp patch. Just need time to focus on it to fix it. 20:08:59 Any other updates today? 20:09:32 I hope we'll have this merged soon 20:09:36 #link https://review.openstack.org/#/c/580724/4 20:09:39 Feature freeze is also next week, so... 20:09:46 The back port for queens is already in place 20:09:58 patch "Automatically set Barbican ACLs" is ready for reviewing. this is something I'd like to have merged soon 20:10:00 #link https://review.openstack.org/552549 20:10:10 Timed out on functional... sigh 20:10:16 Yeah, me too 20:11:28 Thanks for helping track down the DIB/pylint issue BTW 20:11:50 sure, glad I could help 20:12:08 #topic Open Discussion 20:12:15 Other topics today? 20:12:29 It's a real bummer this internal thing landed early and right at the end of the cycle. 20:12:52 merge open queens backports and tag 20:13:30 Yes. Please ping me if you think we have in what we need. A second set of eyes tracking that would be helpful 20:13:45 I reviewed a bunch of backport stuff yesterday 20:13:54 oh, wait. most got merged a few hours ago 20:14:15 I think we could tag it as is now 20:14:35 Aren't we missing limits and the failover fix? 20:14:39 Let's also include https://review.openstack.org/#/c/583233/ :-) 20:15:58 if we can have them merged soon, yes :) 20:16:22 failover fix is super important, definitely 20:16:28 yes 20:17:31 Any other topics today? 20:18:04 I'm seeing some people kill -9 the controllers which leaves things in a bad state. 20:19:54 And story filed for this? To get some more details.. 20:20:00 I'm toying with the idea of having the controllers have Ids of some sort. Then when a controller takes ownership of an object, we save the owning contorller ID in the DB. Then if someone kill -9 the controller, it can search on startup and if there is an object marked as the controller owning it, it would automatically move it to ERROR. 20:20:31 Kind of an interim solution until someone takes on the flow recovery work 20:20:37 using taskflow 20:21:17 Thoughts? (other than file a story, grin) 20:23:00 Ha, ok then 20:23:47 In the interim interim I'm telling folks to stop doing that and use graceful shutdown 20:24:35 Haha 20:25:18 Any other topics today? 20:26:05 Otherwise I will close out the meeting and go back to fighting RPC issues in third party code.... 20:26:21 go for it :) 20:26:51 Ok, thanks folks! Have a good week 20:26:57 #endmeeting