20:00:02 #startmeeting Octavia 20:00:03 Meeting started Wed Sep 23 20:00:02 2015 UTC and is due to finish in 60 minutes. The chair is xgerman. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:04 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:00:07 The meeting name has been set to 'octavia' 20:00:13 #chair blogan, johnsom 20:00:13 Current chairs: blogan johnsom xgerman 20:00:17 o/ 20:00:20 Here 20:00:22 hi 20:00:24 hi 20:00:24 hi 20:00:35 #topic Announcements 20:00:35 o/ 20:00:43 o/ 20:00:44 Hi 20:00:59 oh #link https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda#Agenda 20:01:00 Octavia is the reference driver for neutron-lbaas! 20:01:06 yeah!!! 20:01:09 How is that for an announcement.... 20:01:12 #link https://blueprints.launchpad.net/neutron/liberty 20:01:27 (awesome) 20:01:28 ^^^ yeah 20:01:41 love the green color 20:02:14 Mitaka design summit ether pad: #link https://etherpad.openstack.org/p/neutron-mitaka-designsummit 20:02:29 good stuff… add any topics which you like to see discussed 20:02:30 I want to congratulate the whole team. We worked together well and pulled off a huge accomplishment in taking Octavia from an empty repository to the reference driver in about a year. 20:02:38 +100 20:02:49 o/= 20:02:53 rightly said 20:03:37 yeah, thanks everybody — this is HUGE!! 20:03:55 also thanks to our lt dougwig and fearless leader mesentery for support 20:03:59 mestery 20:04:12 +100 20:04:14 +1000 20:04:23 dow we get a 10,000? 20:04:36 mestery: i think he thinks you're like dysentery. 20:04:42 Howdy, howdy! 20:05:05 huge huge kudos to the team. lbaasv2 + octavia. 20:05:14 hh 20:05:16 gg 20:05:23 And yes, that's friggen awesome, y'all! 20:05:59 What are people leaving teams? This sounds like a retirement party. 20:06:33 the namespace server retired; labs v1 retired — thank you for your service you two 20:06:42 labs=lbaas 20:06:44 crc32: Eh... taking a moment to appreciate how far we've come is totally appropriate, eh. But it doesn't mean we're stopping by any stretch of the imagination. 20:06:52 No, but we should celebrate the achievement. No worries crc32, I'm sure we will be handing out new work by the end of the meeting... grin 20:07:02 ok. Just making sure. 20:07:28 It sounded like. "Congradulations this company has come really far. But... Times are getting hard and ... 20:07:56 nah, things are only tough if you are doing LBaaS v1 (not looking at a certain LB vendor) 20:08:02 On that note-- has RC1 officially been cut? Are we now open to getting other patches reviewed / merged that we know weren't going to make it into RC1? (I'm thinking mostly of the pool sharing patch that I keep on having to babysit. ;) ) 20:08:16 yes mitaka is open 20:08:22 Sweet! 20:08:28 once a certain patch merges 20:08:33 I shall start harassing y'all for reviews forthwith! 20:08:36 We are going to talk about Octavia release a bit later in the meeting 20:08:38 D'oh! 20:08:45 ok 20:08:58 yeah, today I even posted the agenda 20:09:03 dougwig xgerman: o_O 20:10:45 ok, Locking servers by service token” - #link https://review.openstack.org/#/c/203880/ 20:11:07 from our friends who run their service vms in the tenants project 20:12:29 interesting argument that the boundary between tenant spaces is stronger than between vms in a service tenant 20:13:14 #topic Tagging our release + packaging 20:13:48 I think we should tag a release for Octavia. 20:14:12 johnsom: +1 20:14:21 +1 20:14:30 We need to decide on the release number if you all agree we are ready for a release. 20:14:42 I feel like we've hit a milestone and therefore should tag it. 20:14:55 well, we should release an RC 20:14:58 like Neutron 20:15:00 Let's call it 0.5 ? 20:15:06 0.5.0 and 1.0.0 are on the table 20:15:32 i vaguely recall mark saying that openstack services had to be based on >1.0 20:15:33 Im in agreement with blogans comment earlier and feel its not quite poslished enough to be a 1.0 20:15:34 0.5.0 corresponds with the roadmap we've had (with minor adjustments) for a very long time. 20:15:40 o 20:16:02 I mean were there 20:16:09 dougwig: Aah... so we go with 1.0 for political reasons? 20:16:10 i guess i slightly prefer 0.5 because 1.0 just sounds more polished and feature rich to me, but its not a hard stance 20:16:17 just a lot of rushing happened past month or so and feel like we may be a bit off in some areas 20:16:34 sbalukoff: i don't have a strong opinion. if there's no rule, 0.5 as planned is fine by me. 20:16:35 well I think we would release x.x-RC1 20:16:35 pc-pothole: So, what... 0.9? 20:16:43 sbalukoff: perfect! 20:16:50 RC1 indicating the lack of polish 20:17:18 If the release number has a lot of meaning outside this group, then I think we go with whatever makes us seem the most "legitimate" 20:17:25 Since we are, in fact, legitmate. :) 20:17:31 legitimate. 20:17:40 2015 20:17:42 ? 20:17:50 I don't really have a strong opinion on it. 20:17:56 The release number will likely end up on the vendor packages 20:18:07 but I have a lot of niggling annoyances with how I see others doing release numbers wrong. 20:18:15 Like naming releases after dates. 20:18:20 * sbalukoff stares at xgerman 20:18:21 well there is an argument to sue the same as neutron-lbaas to indicate we are compatible 20:18:43 Yeah, but we're also independent. 20:18:53 So, I don't really like that either. XD 20:19:13 ok, dougwig any advice 20:19:24 Like... I don't think I've ever heard anyone using corosync and pacemaker separately... but these packages are distinct and have separate versioning. 20:19:26 #link https://review.openstack.org/226933 20:19:40 Independent is changing... 20:19:51 if it were me, i'd go with 0.5 as planned. mestery, are there any requirements for us to use 1.0 or sync versions? 20:20:06 once independent changes, we don't get to choose. it's synced, fyi. 20:20:34 dougwig: I'm fine either way to be honest 20:20:39 dougwig: Right 20:20:39 johnsom: Are you saying that we'll be back in the same boat we were in a year ago where only Neutron core devs can approve Octavia patch merges? 20:20:49 dougwig: Once part of the release (in Mitaka), it will become 1.0.0 then 20:20:51 no 20:20:54 Because that's really, really not right. 20:20:55 So, 0.5 makes sense now 20:20:57 sbalukoff: no, still a separate core team. this is just release management. 20:20:58 sbalukoff no 20:21:02 Ok. 20:21:03 sbalukoff: no 20:21:15 ah, this sounds good then 20:21:19 So... we *are* independent. Even if we work together closely (as we probably should try to do). :) 20:21:38 sbalukoff: You are independant, but not release independant once the patch merges 20:21:45 you're part of the neutron borg from a release perspective at that point 20:21:45 :D 20:21:49 Right. 20:21:55 * mestery plugs sbalukoff in 20:21:57 Ok, so it sounds like 0.5 is ok. Do we need to vote, or should we just make it so? 20:22:03 johnsom: It is done 20:22:04 ;) 20:22:10 I just need to know when to release it to pypi 20:22:11 done + done? 20:22:35 it's not an octavia meeting without a vote. 20:22:39 let's vote on whether to vote. 20:22:40 Haha! 20:22:52 hahaha, ah dougwig is feeling better.... 20:22:55 #vote you cray 20:22:56 with belong gone I can use an iron fist :-) 20:23:06 blogan 20:23:08 lol 20:23:26 Ok, so let's make it so. 0.5 and cut the release! 20:23:30 blowgun 20:23:32 woot! 20:23:38 Yay? 20:23:41 yep - go, go, gop 20:23:42 mestery: are you going to put 0.5 on pypi? what's that process? 20:23:51 dougwig: yes 20:23:55 I will update launchpad 20:23:56 and how do we communicate to packagers that they need this with neutron-lbaas now? 20:23:59 dougwig: http://docs.openstack.org/developer/neutron/devref/sub_project_guidelines.html#sub-project-release-process 20:23:59 also we probably need to branch/tag — so we can back port patches 20:24:00 #link http://docs.openstack.org/developer/neutron/devref/sub_project_guidelines.html#sub-project-release-process 20:24:27 Cool. 20:24:46 :) 20:24:54 sub project owner = blogan? 20:24:55 I did some useful things as PTL after all! 20:24:56 Imagine that! 20:25:02 Haha! 20:25:09 *cough* lameduck *cough* 20:25:10 ;) 20:25:24 Actually, we love you mestery. You've been good to work with, eh! 20:25:29 rofl 20:25:37 sbalukoff: I feel the same way my friend :) 20:25:49 I thought what mestery was doing included the tag/branch. No? 20:25:50 Never say a little smile and some kindness doesn't go a long way, because it does. 20:25:51 mestery for president!! 20:26:02 johnsom: Yes, I do the release itself johnsom 20:26:52 Well. This is awesome, folks! 20:27:04 indeed 20:27:12 You folks just need to say jump and 0.5 will be on pypi within an hour maybe :) 20:27:26 jump 20:27:28 jump! 20:27:38 lol 20:27:38 :) 20:27:43 sbalukoff: this is CS. "jump!" != "jump". 20:27:44 #startvote jump? jump 20:27:45 jump, jump, yeah 20:27:45 Begin voting on: jump? Valid vote options are jump. 20:27:46 Vote using '#vote OPTION'. Only your last vote counts. 20:27:48 OK, I'll start doing that now and ping folks if I hit questions :) 20:27:59 #vote jump 20:28:01 #vote jump HA! 20:28:02 mestery: jump HA! is not a valid option. Valid options are jump. 20:28:03 #vote please not a vote 20:28:04 dougwig: please not a vote is not a valid option. Valid options are jump. 20:28:04 :P 20:28:10 #endvote 20:28:11 damn openstack bot 20:28:11 Voted on "jump?" Results are 20:28:12 jump (1): sbalukoff 20:28:12 Cool, I will watch and do the launchpad steps. 20:28:13 * dougwig dies inside. 20:28:14 #vote how high 20:28:21 * mestery falls over 20:28:22 HAHA 20:28:31 Mine was the only vote that counted. 20:28:34 lol 20:28:35 As it should be. ;) 20:28:43 well, the vote was rigged to begin with :- 20:28:45 ) 20:28:55 lol 20:29:05 xgerman: Oh, you're just now catching on about that? ;) 20:29:09 folks... settle down. :) 20:29:33 At least no one linked Kris Kross or Van Halen youtube videos... 20:29:45 Heh! Ok... so... what's next? 20:29:49 hahahahhahaha 20:29:51 where is rm_work when you need him? 20:30:20 yea where is he? 20:30:55 #topic Brief progress reports 20:30:55 Next up on the agenda is either progress reports (kind of covered) or Octavia talk 20:31:16 I think mwang2 did some work on docs 20:31:25 yep 20:31:26 I have been focused on the release. I have more work to do on VRRP. 20:31:43 I started working on the heat 20:31:53 yeah!! 20:31:54 need more review #link https://review.openstack.org/#/c/224929/2 20:31:59 So, I've compiled a lot of notes about different ways to get Octavia working in devstack for a potential lab situation. I think I like a small variation on rm_work's script best for this. I intend to document this both in tree to get people started, and potentially separately specifically for the lab session if people think we need it. 20:32:05 at the same time waiting for inputs from heat guys 20:32:19 Part of the reason I bring this up is because I'm not sure what we have planned for the lab. 20:32:29 sbalukoff we will need it 20:32:37 And writing a good how-to doc here will depend somewhat on what we intend to do in the lab. 20:32:41 and we haven’t planned anything yet — it’s still like 30 days 20:32:53 I usually prepare talks the night before :-) 20:32:55 sbalukoff I was going to start a vm image. 20:33:04 johnsom: Oh, good! 20:33:20 I heard something about someone distributing USB sticks with said image on it as well? Is that happening? 20:33:33 well, we still need a sponsor for the usb sticks 20:33:45 (see #link https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda#Agenda) 20:33:45 It seems previous labs have used vmware images. I assume with player. So that is the direction I am heading 20:33:54 johnsom: Also, do you have any specifics on what we might want to do in the lab with that image? 20:34:08 install it on people’s laptops 20:34:12 I can contribute a baremetak server from a forgotten test account. 20:34:12 make them create lbs 20:34:16 look at logs 20:34:19 troubleshoot 20:34:21 sbalukoff I have got as far as "setup Octavia" 20:34:28 So, we need to hammer that out. 20:35:04 Heh! Ok, let's touch base outside the meeting then to come up with a plan. I figure a lab is a lot like a live demo-- you want to have a backup presentation in case things go sideways as a fall back in any case. 20:35:14 I'm going to put the lbaas talk up on google docs. If you want I can setup a document for the lab too 20:35:27 johnsom: that would be great! 20:35:39 most important for a lab is step by step instructions people can do and then we walk around and help if they bet stuck 20:35:44 Ok, I'll put some links in the channel later today 20:35:58 xgerman: Agreed. 20:36:16 yeah, I did my fair share of labs in the day ;-) 20:36:33 johnsom: I appreciate it. 20:37:31 ok, trevorv any progress on containers? 20:37:54 Beyond that I do want to report that we've got some stuff in the works which is likely to make it onto the Mitaka development cycle. I'm working internally with teams here to refine that before we present it to y'all... but you can probably anticipate hearing about it in the next couple weeks. 20:38:12 sbalukoff cool!! 20:38:26 This is mostly around heat integration and moving forward with an active-active design... and getting more people to help review existing code and work toward greater stability. 20:38:41 all good stuff!! 20:38:44 Yep! 20:38:47 stability - that is awesome 20:39:03 That's it for me for now. 20:39:04 sbalukoff Until we see code, it doesn't exist. Grin, just had to throw that back at ya 20:39:28 I think we covered the Octavia talk 20:39:32 so next up 20:39:38 johnsom: Oh yes! and I've been throwing that in the face of these IBM devs who have never done anything public OpenStack related before. 20:39:47 #topic Open Discussion 20:39:48 johnsom: You know me, I flip shit in all directions. ;) 20:39:54 the following patches need to be reviwed #link https://review.openstack.org/#/c/226599/ #link https://wiki.openstack.org/wiki/Neutron/LBaaS/HowToRun, #link https://review.openstack.org/#/c/215359/ 20:39:55 hahaha 20:39:56 I have following query : Is L7 capability is targeted for Liberty (https://review.openstack.org/#/c/148232/)? If it is targeted for Liberty corresponding changes could be done for Vendor(NetScaler) Driver. 20:40:17 Santos should be M 20:40:21 Seems it is targeted for Mitaka? 20:40:27 Thanks 20:40:33 yep, but early M 20:41:02 and will backport to liberty? 20:41:10 not sure yet 20:41:14 Yep. Work is close to compete on that for neutron-lbaas. And it shouldn't be hard to replicate in Octavia. (If nobody else tackles that before me, I'll probably get to it once the documentation and stuff for the talk is sorted.) 20:41:43 Yes, no decision on a back-port just yet. 20:42:00 I think we will decide once the code is in 20:42:16 dougwig would that be ok? 20:42:33 or is their a rule we can’t back port since it will affect vendors 20:43:08 i don't think it'll backport to liberty no matter what (l7, right?) i've never seen a feature backported to stable. 20:43:30 dougwig: L7, yes. 20:43:30 er... correct 20:43:40 so that’s settled 20:44:13 Santoshs: I think most of the operators here are likely to go with a pre-release of the "M" version of neutron-lbaas and Octavia if they decide they need L7 badly enough. (We do!) 20:45:06 Pre-release of M version?? 20:45:17 well, master 20:45:23 or M-1 20:45:26 k 20:45:41 the packagers will cut releases at the Milstones 20:45:58 so if we get it in early it will be in M1 20:46:02 M2... 20:46:03 And operators may choose to temporarily fork to get certain features. 20:46:18 Right. I'm hoping to get it in early. :) 20:46:42 It could be assumed as some L + master version 20:46:52 Santoshs: Right. 20:47:01 But not L 20:47:20 That is almost certainly the case. 20:47:33 In any case, the code has to land first. :) 20:47:39 +1 20:48:23 Lastly, sbalukoff / blallau — you said we had more EMEA based people for whom that time is to late 20:48:35 any input if we need to move the meeting time 20:48:43 also SantoshS 20:49:09 trying to gather if there is a need 20:49:17 It would be great if it is earlier 20:49:21 The IBM teams I'm working with have set up an internal meeting on Thursdays at 16:00UTC. So if we were to move the meeting time, going earlier in the day would be better. 20:49:48 Though, I have not yet gotten a solid commitment to attend these IRC meetings from that team yet. 20:49:51 o/ 20:49:56 So, I'm not sure what to tell you. 20:50:06 ok, blallau? 20:50:39 I would like to be regular attendee if it is earlier 20:50:42 sorry, worked 17 hours yesterday, just woke up <_< 20:50:53 rm_work: no problem 20:51:00 rm_work: Don't worry, I voted for you. 20:51:08 sbalukoff: glad you like my script; xgerman: sorry i couldn't link contextually relevant youtube videos earlier :P 20:51:21 heh 20:51:22 no worries 20:51:46 ok 20:51:59 earlier better for you? 20:52:11 yes thanks 20:52:31 ok, I will propose some new time to the ML 20:52:39 For my own sake: please no earlier than 16:00 UTC! 20:52:50 I am the polar opposite of a morning person. 20:52:59 :-) 20:53:17 ... same 20:53:27 if that weren't obvious 20:54:34 yeah, we need to find a balance + the meeting channels need to be available (need to sync with dougwig how to figure that one out) 20:54:46 Ok. 20:55:20 anything else? 20:56:11 #endmeeting