20:00:10 #startmeeting Octavia 20:00:11 Meeting started Wed Apr 19 20:00:10 2017 UTC and is due to finish in 60 minutes. The chair is johnsom. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:12 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:00:14 The meeting name has been set to 'octavia' 20:00:18 o/ 20:00:22 Hi folks 20:00:32 o/ 20:00:54 #topic Announcements 20:01:00 TC voting is open, check your email for the link 20:01:13 I think today might be the last day to vote. 20:01:22 xgerman, is running :-) 20:01:41 Yeah, crazy person... Grin 20:01:48 :) 20:02:07 He is still on vacation, so I can get away with that 20:02:31 Also, our third party sonar gate is no more 20:02:48 yup, it's not like the meeting is logged or anything :-) 20:03:08 It got accidentally deleted, so.... 20:03:22 will it get restored? 20:03:37 I am setting up a coverage gate with infra that should at least give us the coverage information. 20:03:48 #link https://review.openstack.org/457819 20:04:00 We needed it anyway, but sad to see the sonar gate go. 20:04:19 I still have the setup instructions so maybe sometime in the future we will build another one. 20:04:54 ack 20:05:12 yes, very sad 20:05:33 Also of note, Mitaka is officially EOL. ( think I mentioned this last week too) 20:05:42 probably partially my fault for not insisting the person who created the VM name it more appropriately >_< 20:06:04 You will see the stable/mitaka branch go away soon. There will be a tag "mitaka-eol" with the last "release" 20:06:31 A, well, that poor gate has had a rough life anyway..... 20:06:40 ok, hopefully our mitaka stable is fully working... i noticed some weirdness with some of the stable branch scenarios recently 20:06:41 but 20:06:44 * rm_work shrugs 20:07:25 I really liked sonar, it was a good canary T_T 20:07:26 Yeah, we have had to do a number of stable branch fixes recently due to projects that run without stable branches, namely devstack and diskimage-builder. 20:07:42 finished in like ... 2 minutes 20:08:03 ... send me those instructions again, maybe I can get it set up here 20:08:20 Rackspace HP Sonar Gate (by GoDaddy) 20:08:27 rm_work Sure, I can dig those out. I also still own the domain. 20:08:29 Ha 20:08:40 tempted to do it just for the name 20:08:43 Yeah, like I said, rough life 20:09:04 Ironically, the domain is registered through my personal godaddy account..... 20:09:42 lol 20:09:51 maybe I can get them to credit you for it 20:10:18 Ha, it's not much so probably not worth the effort. I have other domains so I think it was discounted. 20:10:42 It's not like the old .io domain which was surprisingly expensive 20:10:51 yeah O_o 20:10:57 those run like $60/year 20:11:14 It was $70 when I thought about registerting it 20:11:26 #link http://www.octavia.cloud 20:11:33 It still goes to our docs page... 20:12:01 .cloud ... lol 20:12:06 :P 20:12:10 Any other announcements today? 20:12:51 #topic Brief progress reports / bugs needing review 20:13:17 Ok, thanks to rm_work our py3 and scenario gates are fully functional now. 20:13:43 \o/ 20:13:48 kudos rm_work! 20:13:52 also, our v2 API patches are ALMOST done 20:13:59 just HM and then paging/sorting 20:14:06 and some work around single-create and cleanup 20:14:09 I released our milestone pike-1 packages last week and registered out community goals 20:14:13 but the majority is complete 20:14:19 HM could use review 20:14:25 Yeah, stats and status still too 20:14:30 ah yeah T_T 20:14:35 But the majority is done 20:14:49 all of the objects, so you can create a fully hydrated LB 20:14:51 I have started on the pools API docs. 20:15:14 Should be able to get those finished this week. 20:15:58 I also did the quotas for the v2 API, though it needs some minor updates after this weeks changes 20:16:56 My plan for the next week is to keep cranking through API docs. 20:17:30 nmagnezi how is the user_group patch going? 20:18:09 johnsom, pretty good. tested okay with Ubuntu amp. tomorrow I'm going to test agains a centos amp 20:18:32 Ok, so should be good for review/testing tomorrow then? 20:18:33 johnsom, and if that passes it should be the final version of the patch (pending reviews) 20:18:40 yup 20:18:48 Excellent, looking forward to it. 20:19:19 I'll post detailed test results (manual tests I ran) in the patch as soon as it is ready :) 20:19:46 I know Jude has done some good work on getting the HM/HK processes to exit faster (not stuck sleeping). 20:19:51 That is good too 20:20:15 Ok, moving on 20:20:20 Still needs a review though 20:20:29 #topic Plan to make scenario gates voting 20:20:56 Now that we have a break in the OpenStack storm and our scenario gates are functioning I plan to make them voting. 20:21:05 This is part of our py3 goal for Pike. 20:21:24 which ones? 20:21:41 The new infra images have resolved our OVH issues, so things should be in good shape 20:21:42 I assume *not* anything n-lbaas (please T_T) 20:22:31 I plan to do all except centos (pending user_group patch) and lxd 20:22:56 Why not the n-lbaas scenarios? 20:23:23 because n-lbaas needs to die 20:23:30 what were the ovh issues? (I'm not aware of any changes to specifically address anything in ovh) 20:23:41 and also, having a voting dependency on n-lbaas could be problematic 20:23:45 clarkb libvirt was crashing 20:23:47 because we could run into a circular merge issue 20:23:57 johnsom: oh,thats actually a change to devstack, but good to know 20:24:25 I posted a "thank you" in the infra channel yesterday. 20:25:10 rm_work We are still going to make sure we function behind neutron-lbaas for Pike.... 20:25:57 yes, but 20:26:01 non-voting is fine >_> 20:27:09 I think we need to make it voting, we kind of have a contract for our users that neutron-lbaas will still work with the octavia driver 20:28:12 note that the change to use UCA doesn't currently affect stable branches. Its master and going forward 20:28:20 and we can make sure we watch for issues, but if we have a gate problem in octavia because of neutron-lbaas I am going to start throwing things 20:28:30 Ha 20:29:06 clarkb Yeah, we will live the recheck life in stable like we had been 20:29:11 if it's a vote, i'm going to -1, if it's your decision as PTL, that's fine 20:29:36 Well, I wanted discussion (thus the agenda item) 20:30:00 just yesterday we had an issue where we needed to merge something in both repos because both had distinct gate issues 20:30:23 it SEEMS like that would have been more of a PITA if we have a voting gate between them 20:30:53 Until we officially deprecate neutron-lbaas we need to make sure we have a fully functional stack 20:31:42 sure, so procedurally, let's make sure we look and see the check is green before we +A :) 20:32:36 Agreed, these outside changes breaking us are a huge pain, but that shouldn't stop us from making sure we have functional and quality code merged. 20:33:00 and having voting *octavia* scenario gates should do that 20:33:27 I will compromise. I will go voting on octavia and leave neutron-lbaas non-voting a while longer to see it's stability. 20:34:00 that's all I want :) 20:34:10 Yeah, but octavia gates go direct to the octavia API which is not how most of our deployments are setup. 20:34:47 Ok, next up. 20:34:52 #topic Boston Summit planning 20:35:13 I just booked travel... I *can* make it on Monday in time for the lab 20:35:16 and I can help on talks 20:35:21 Which unfortunately German is still on vacation so we have the opposite half of the conversation this week 20:35:24 but I don't have time to do much work preparing for either 20:35:42 so I am happy to show up and help 20:35:43 and review slides or something 20:35:45 That was what I wanted to discuss. Should we cancel some of the talks? 20:35:57 I am not sure what we will have time to prepare for 20:36:02 we NEED the project update one 20:36:07 I will not be able to attend, but can help with slides and such. 20:36:09 and probably we should do the main LBaaS one 20:36:16 what was the third talk? 20:36:24 the lab is ... we should discuss the lab, but need xgerman 20:36:27 Load balancing 101 20:36:57 It seems like the main LBaaS talk and the project update could overlap content. 20:37:12 hmm 20:37:15 yeah mayhaps 20:37:25 I feel like... is LB101 the same thing we've done every year? 20:37:29 I had planned project update to talk about the new project and the roadmap, where the main one would be about the merge details 20:37:43 is that useful? 20:37:45 No, LB101 is like "A network load balancer is..." 20:38:06 Like basics and cookbook content. 20:38:17 right ... haven't we done that before? 20:38:18 or no 20:39:10 We have done small segments, but not in a while or to the depth I was going to go in this session. It's one we could cancel and just post a youtube for as well 20:39:41 k 20:39:42 I mean... I can try to give any amount of talks 20:39:48 I just *can't* plan for them 20:39:49 so 20:40:26 Ok, if you are up for them, I can help with slides. Most of those were my talks so I didn't want to dump the presentation job on other folks. 20:40:59 also I can't guarantee I can be as thorough as you could for description or questions 20:41:07 but I can disclaimer up front that I'm a backup :P 20:41:23 Yeah. Or phone-a-friend 20:41:24 HA 20:41:42 the lab is what I'm worried about the most 20:41:43 lol yeah I could Zoom/Skype you in on my Laptop 20:41:55 Giant head on a projector 20:41:57 wouldn't that be fun 20:42:01 Yeah, if the wifi is good enough 20:42:07 yeah we can test and see 20:42:40 Anyway, ok, so current plan, without German here, is to still do all four, but we need to create slides 20:42:48 right? 20:42:51 I guess 20:42:56 Ok 20:42:57 I am concerned about the lab 20:43:02 because with just two of us 20:43:07 it will be difficult as-is 20:43:11 and i don't know what his plan for it is 20:43:13 or if he has one 20:43:15 Yep, agreed 20:43:35 just play the recording of past years :D 20:43:42 it took a lot of prep last time that I did on the technical side 20:43:43 so I hope he isn't just planning to wing that the night before 20:44:00 in addition to the actual content of the talk 20:44:04 We can have it on next weeks agenda as well 20:44:08 ok 20:44:20 cutting it close though 20:44:24 #topic Discuss asynchronous API and response handling 20:45:05 This was discussed last week in depth. Do we want to try to discuss more in 15 minutes, or just give some open discussion time and put it on next weeks agenda? 20:45:46 crickets 20:45:59 hmm 20:46:07 i missed last week 20:46:10 maybe i should catch up 20:46:19 I think we need some time on this, so let's move it to next week but plan to decide a path forward then. 20:47:00 #link http://eavesdrop.openstack.org/meetings/octavia/2017/octavia.2017-04-12-20.00.log.html#l-69 20:47:19 #topic Open Discussion 20:47:29 Other topics for today? 20:47:49 I am taking over the python-octaviaclient 20:48:04 Ah, yes, Excellent! 20:48:09 just a general fyi 20:48:13 ^^ along with possibly another resource here 20:48:24 we'll see if he is able to come on 20:48:31 I should probably look at the state of the SDK work as well 20:48:56 Yeah, we dont have a fully fleshed out spec for it though but I am using https://docs.openstack.org/developer/python-octaviaclient/decoder.html#lbaas-v2-cli as aguide 20:49:28 Yeah, any time you have questions please ask in the channel. 20:49:46 Perfect ty 20:49:51 The old client was not as good as we can make the OSC plugin so 20:50:29 If you would like me to put tracking bugs in or something like that, let me know. 20:50:56 That would be useful IMO. 20:51:51 Ok, I will make some time to enter those. I will create a new tag 20:52:15 Probably "octavia-osc" 20:53:25 Any additional items today? 20:54:12 Ok, thanks folks! Chat with you in the normal channel 20:54:16 o/ 20:54:20 o/ 20:54:22 #endmeeting