20:00:51 #startmeeting Octavia 20:00:51 Meeting started Wed Sep 30 20:00:51 2015 UTC and is due to finish in 60 minutes. The chair is xgerman. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:52 that is a lot of arms 20:00:52 Hi 20:00:53 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:00:55 o/ 20:00:56 The meeting name has been set to 'octavia' 20:00:57 o/ 20:00:58 o/ 20:00:59 #chair blogan 20:01:00 Current chairs: blogan xgerman 20:01:00 Howdy, folks! 20:01:01 o/ 20:01:10 o/ 20:01:17 Someone ate their wheaties this morning 20:01:18 #topic Announcements 20:01:28 o=/ 20:01:31 hi 20:01:33 chemically infused wheaties 20:01:35 hello 20:01:38 Mitaka design summit topics #link https://etherpad.openstack.org/p/neutron-mitaka-designsummit 20:02:10 please add, +1, etc. if something is interesting to you 20:02:28 Sweet. 20:02:46 hi 20:02:52 oh, and for those under a rock: amax is the new Neutron PTL 20:02:56 armax 20:03:12 all hail our new overlord 20:03:15 aye 20:03:18 heh 20:03:19 not sure if dougwig can cling on to be a lt in the new regime 20:03:20 Haha! Welcome! 20:03:21 Thanks... us rock people have trouble getting those updates 20:03:28 I did miss the official results, but cool 20:03:38 xgerman is my biggest fan. 20:03:54 if that’s all it took a lot of people would be lt. 20:04:11 He is already helping us out with the Octavia release 20:04:16 I am dougwig’s proxy 20:04:22 he’s the one really pulling the strings 20:04:41 welcome - all bow 20:04:45 \me bows 20:04:50 armax I am so sorry. He really does like to make people dance.... 20:04:58 * TrevorV laughs at xgerman and his failed bowings 20:05:00 You're Bush to dougwig's Cheney? ;) 20:05:07 johnsom: no need, that was the plan all along :) 20:05:16 sbalukoff lol 20:05:20 We've fallen right into their trap! 20:05:38 * dougwig cries in the corner. 20:05:44 * blogan watches 20:05:45 Seriously though-- great to have you on board, and thanks for attending this meeting today, eh! 20:05:57 +1 20:05:59 o/ 20:06:00 +! 20:06:02 +1 20:06:07 o/= 20:06:19 yes thanks for attending this meeting and the countless others 20:06:55 #topic Tagging our release + packaging 20:07:18 So we ran into some trouble getting the octavia release up on pypi 20:07:20 johnsom was working with mestery on getting us into pypi 20:07:40 trouble? 20:07:44 Evidently the octavia repository wasn't quite setup right (no stares in any direction) 20:08:02 * rm_work has no idea how pypi really works internally 20:08:11 But, we got some help from infra this morning and I have patches up to fix those. 20:08:15 * TrevorV shares rm_work 's lack of knowledge here 20:08:20 rm_work magic? 20:08:20 johnsom: patches in infra? 20:08:31 We are just waiting on the infra patch to go in 20:08:39 #link https://review.openstack.org/#/c/229518/ 20:08:39 Cool. 20:08:47 link? so we can all +1 to show our support and gartitude 20:08:51 gratitude 20:09:02 Octavia is up on pypi now, but it is just the initial push I did when setting it up. 20:09:32 well, does this mean we are still sort in a code freeze? 20:10:17 In theory yes, as I expect there will need to be another tag added to the repo to get it through the OpenStack release process 20:10:51 johnsom: Can you broadcast loudly in channel when that tag happens? 20:10:59 Sure can. 20:11:04 Danke! 20:11:10 also once we have Octavia in pypi who will test that it actually works? 20:11:12 fnaval? 20:11:17 mestery or armax will actually push the button, but I am trying to follow along 20:11:37 * xgerman thinks johnsom would make a great release manager 20:11:49 Oye 20:11:54 Let's vote on it! ;) 20:11:59 +1 20:12:01 the customary vote 20:12:10 +1 20:12:13 +1 20:12:15 Just for dougwig 20:12:24 That's 4 of us. Congratulations, johnsom. 20:12:25 * rm_work wonders what johnsom votes 20:12:56 Is that even a role? 20:13:00 Is now 20:13:04 sounds official 20:13:23 doesn’t neutron has a release team? 20:13:35 yea its called blogan upstream 20:13:39 Haha 20:13:46 * johnsom bows 20:14:12 * blogan is lost 20:14:21 * sballe congratulates johnsom in his new role 20:14:28 this meeting is a tad wandering. 20:14:30 #topic Brief progress reports 20:14:32 yes it has 20:14:34 So, as release manager, what is going to be in next weeks release? hmmm??? 20:14:59 has rc2 been cut? 20:15:04 or deadline past for that? 20:15:35 Release stuff above and any little bug fixes we have needed for liberty. I also put up google docs to get us started on our Tokyo presentations. 20:15:52 I have a first draft of the blueprint the IBM research team wants to use to add active+active and auto-scale to Octavia, hopefully in Mitaka. I want to do a small revision of it before putting it in front of the group, and that team is on vacation right now anyway... so expect to see it around mid next-week. 20:15:52 i think this one is worth getting into liberty 20:15:55 https://review.openstack.org/#/c/210968 20:15:59 yeah!! we have an extra Tokyo topic 20:16:06 ajmiller: i added tests for it ^ hope you dont mind 20:16:16 And yes, I'll add that to the list for the design session. 20:16:18 blogan: not at all! 20:16:20 heat hooks BP is up but still wip, needs to cross check some parameter with the api doc. 20:16:42 great bana_k 20:16:46 blogan I tagged that as an RC candidate. I agree it would be good to get in 20:16:51 bana_k: Awesome! 20:16:51 johnsom: oh great 20:17:06 api doc is not entirely correct so reviewing it n raising bugs. 20:17:15 thanks 20:17:20 thanks :) 20:17:47 Apologies it's taking me so long to get the stupid docs updated. And yes, johnsom, if it's not in gerrit it doesn't exist. I'll try to correct that this week. 20:18:20 Grin 20:18:32 I need to fix https://review.openstack.org/#/c/227912/ — maybe paying some voodoo master 20:19:02 Otherwise, I'm just waiting for the go-ahead on release tagging to start guilting you into reviewing the Neutron-LBaaS shared pool code and L7 stuff that Evgeny has been working on. ;) 20:19:27 yeah, we want L7 in Mitaka 20:19:39 I think blogan and I need to do some work in LP for milestones, etc. 20:19:50 xgerman: sure 20:19:52 Cool beans! 20:20:03 Also, I put a global requirements change in to exclude coverage 4.0 as it breaks our unit tests in bad/strange ways. 20:20:13 #link https://review.openstack.org/#/c/228581/ 20:20:16 also noop driver fixes that we might use for some tests, but even if we dont at least this allow them to work: https://review.openstack.org/#/c/229262/ 20:20:30 NICE 20:21:14 should we aim to get that in Octavia RC2 (are we doing an RC?) 20:21:16 Upstream bug against coverage is here: 20:21:23 #link https://bitbucket.org/ned/coveragepy/issues/416/coverage-40-is-causing-existing-unit-tests 20:22:07 i still have one unit test for cert rotation #link https://review.openstack.org/#/c/215359/ 20:22:25 xgerman: that one's isn't too important, so i wouldn't fight for it 20:22:29 if its easy enough, then sure 20:22:42 well, we are still cutting so... 20:23:28 any other progress? 20:23:32 also need to address some update for octavia doc patches #link https://review.openstack.org/#/c/226599/ #link https://review.openstack.org/#/c/224929/ 20:23:47 yes, docs are real important — please review 20:23:54 +1 20:24:04 #action review #link https://review.openstack.org/#/c/224929/ 20:24:26 minwang2, links need to be on separate lines for the bot to pick them up 20:24:33 #link https://review.openstack.org/#/c/226599/ 20:24:38 #link https://review.openstack.org/#/c/224929/ 20:24:43 I gotchu though :) 20:24:48 Trevor thanks! 20:24:49 dougwig wonder if there is some more step—by-step guide w need to write 20:25:05 also sbalukoff you said you were working on that 20:25:08 #link https://review.openstack.org/#/c/218735/ 20:25:15 ^^ can I get a review too? 20:25:19 sure 20:25:31 xgerman: If I need to, I'll make my patch dependent on Min's. 20:25:32 #action review #link https://review.openstack.org/#/c/218735/ 20:25:59 sbalukoff I saw Min’s docs but I also got asked in IRC for a more like deployment guide 20:26:10 xgerman: Yep, that's what I'm working on. 20:26:28 awesome 20:27:18 ok, any other progress? 20:27:38 Sounds like a "nope" to me. :) 20:27:39 #topic Octavia talk 20:27:42 I can't add here, honestly... Been doing internal stuffs :( 20:27:55 TrevorV: I hear you. 20:27:57 so we are still looking for a sponsor for the USB drive ... 20:28:02 Also I won't be able to go to Tokyo... so enjoy it without me guys :'( 20:28:25 too bad… they should have just chartered a plane 20:28:26 Oh! I suppose it's worth mentioning that I'm in the stages of making sure the legal paperwork is right with the OpenStack Foundation to assign the 'Octavia' trademark to them. So... yay! 20:28:44 sbalukoff: thats great news! 20:28:50 +1 20:28:52 So, sometime in 2017, the USPTO will likely have finished reviewing that same paperwork and the trademark will officially transfer! ;) 20:29:08 you used the IBM lawyers? 20:29:11 ha ha ha ha 20:29:14 xgerman: Yep. 20:29:35 Nice 20:29:37 heh 20:29:41 Doc for the hands on lab: 20:29:44 #link https://docs.google.com/document/d/12x6Xc6ktopg9mhxz7yqp_VsFTTqVRxJ_PtZlvzb0xPU/edit?usp=sharing 20:29:50 xgerman: Didn't have a choice. This would have been easier had the USPTO moved faster than IBM's acquisition team (ie. if this were still only under Blue Box), but... such is life. 20:30:08 yeah, but glad it’;s happening... 20:30:15 Yep, me too! 20:30:23 Slides for the presentation: 20:30:25 #link https://docs.google.com/presentation/d/1PYBLtNrkyg__FjyZ2ywynNv-22FW88hDz5j9q8W1ZOQ/edit?usp=sharing 20:30:30 so USB sticks? We can also put up a donation box at the session... 20:30:36 Not much progress on either yet. 20:30:37 sbalukoff: Or the project gets renamed. 20:30:57 hah 20:31:01 heh how much do those cost 20:31:06 for however many we need 20:31:14 did we figure out what size we'd need minimum? 20:31:20 and how big do they need to be? 20:31:29 crc32: That's what I had to demonstrate to the IBM lawyers in exquisite detail and my usual... er... coloful language. 20:31:33 ;) 20:31:34 I think we can fit in 8GB 20:31:35 johnsom: we need a new road image 20:31:54 cool. Ho many usb sticks? 20:31:58 8GB are $7 at staples 20:31:59 Let me know if you want edit rights. 20:32:17 rm_work accepting suggestions 20:32:22 xgerman: yeah but I assume we can get a bulk discount :P 20:32:26 johnsom: something with less clouds 20:32:34 johnsom: it looks like a stormy road ahead 20:32:45 that sends a scary message :P 20:32:48 I was hoping some corporate marketing department would see that as a good use for their money 20:32:50 rm_work hahahaha, I was thinking links.... 20:33:10 xgerman: how many do we need tho? 20:33:13 These images pre-date me 20:33:15 xgerman: I checked with my team. I think they're willing, but worried about the timeline to get it through red tape. 20:33:32 28 dollars for 10 Gigs 20:33:32 Nobody here has any idea how many we need, right? :P 20:33:45 28 dollars for 10 8Gig sticks 20:34:12 I think the web page knows 20:34:16 johnsom: http://www.ewallpapers.eu/w_show/open-road-1920-1080-5092.jpg 20:34:17 I don't know how many we need. I can look and see how many people have already signed up, but that doesn't tell us the limit 20:34:33 Do we know how many seats the room will have? 20:34:41 right 20:34:43 I mean, maybe we'll pack the place, maybe we won't. 20:34:47 rm_work nice pic 20:34:48 like how many could it possibly be 20:34:49 But it seems silly to bring more than that. 20:34:49 100? 20:35:15 johnsom: that seems more "laxidaisical" to me 20:35:18 it says limited 20:35:19 or however you spell that 20:35:38 so far we have 10 signed up 20:35:46 johnsom: maybe http://alphatravelblog.com/wp-content/uploads/2013/07/French-Roadtrip-Open-Road.jpg 20:35:56 We have 11 signed up right now 20:36:02 Well... as long as we have the image available for download someplace, I don't feel guilty about not having enough USB sticks. If you really want to check out Octavia you can download a friggen image. 20:36:25 yeah that sounds like the default solution, and usb stick is a backp 20:36:25 johnsom: did we get the actual rights to the last one or was it just google-image-search? 20:36:26 backup 20:36:41 well, the internet AT the conference will not be great, it never is 20:36:42 rm_work again, pre-dates me 20:36:57 but just a couple to pass around would be fine for that 20:37:03 just post the link to the image on a keynote slide, let everyone at the summit download it all at once 20:37:13 Yeah, that is my worry. I don't want to count on good internet. 20:37:25 actually, maybe it Tokyo it won't be too bad 20:37:32 like everyone there has gigabit i think 20:37:40 Right, so $28 for 10 of these sticks... and we're covered, eh? (Also, who doesn't have the ability to download this stuff? Are we expecting to hand a bunch of these out to Chinese citizens?) 20:37:47 as long as we put ti in a JP data center 20:38:14 closest we can get is like... Hong Kong 20:38:40 I'm ordering 3 sets of 10. 22$ each 20:38:42 alright plan needs some work it sounds like, but a direction in place 20:38:55 Our CDN has a pop in Japan. I will load the image there 20:39:07 awesome — so we have it all covered 20:39:13 crc32: Thanks, eh! 20:39:22 crc32: erk, k -- hopefully they reimburse you 20:39:23 Sweet! 20:39:24 I can’t really see how we can’t expense that... 20:39:26 is 30 enough? 20:39:43 cause thats all I can afford now. 20:39:46 and/or I wonder if RAX has a vendor where we could get them cheaper and with RAX logo on them :P 20:39:55 crc32: It sounds like it probably will be, yes. 20:40:07 yeah, let’s work the corporations again 20:40:11 crc32: i might wait just a moment at least and ask someone at RAX 20:40:17 +1 20:40:21 +1 20:40:24 When they arrive I'll hand them to rm_work since I lose things. :( 20:40:28 lol 20:40:41 Heh! 20:40:42 you ask rm_work. its only 68$ 20:40:42 We need to get the image on them too 20:40:50 hopefully data rate from HK to JP is not far :P 20:40:58 yea but you said you would load them in Tokyo. 20:40:59 err, hopefully data rate is good 20:41:00 I thought 20:41:09 or we can load them here. 20:41:18 yeah we'd want to pre-image them 20:41:23 I'm guessing were going to load them at the last second though. 20:41:26 pretty easy to do en-mass 20:41:31 It's a long plane flight, so whoever has them could image on the way.... grin 20:41:32 crc32: +1 20:41:42 I think rm_work would do that. 20:41:45 incase some last second patches make it in. 20:41:46 He can't sleep regularly anyway 20:41:55 I am so going to be taking ambien on the plane >_> 20:42:12 ok, I think we are all set 20:42:29 #topic Open Discussion 20:42:40 what part of this meeting hasn't been open discussion? :) 20:42:52 lol 20:43:00 Two or three minutes... 20:43:17 so moving the meeting time didn’t garner much support on the ML... 20:43:53 dougwig: We are always on topic. The topic just likes to shift without a general notice to everyone. No our fault, eh. 20:44:07 If anyone knows Matt Riedemann at IBM you can ping him to remove his -1 from the global requirements update. He didn't ready the commit message. 20:44:07 rm_work are you gonna ask rax or do I just order them. I'm like ready to click shit to address. 20:44:19 dougwig, you just upset that we young peeps keep up with the conversational flow and you have trouble with it :P 20:44:36 johnsom: I've no idea who that is... :/ 20:44:57 TrevorV dougwig - that’s how we roll 20:45:01 doesn't this seem like an important patch to get into rc2? 20:45:03 https://review.openstack.org/#/c/220654 20:45:21 yep 20:45:22 #link https://review.openstack.org/#/c/220654 20:45:28 Yeah. 20:45:53 does it need a bug report in lp to get into rc2? 20:46:04 Could be. Do you all want me to tag it? That means we have to get on the reviews and get it merged..... 20:46:08 Yes 20:46:20 TrevorV: get off my lawn. 20:46:45 forget this order then. Deliver estimate Oct 27 - Nov 13 20:47:08 yeah, it might be the JP CDN 20:47:17 dougwig, I found a snipe, and I followed it under your porch, but this snipe had a long tail, and looked more like a large mouse. 20:48:41 crc32: ah lol such slow 20:48:43 back to octavia stuff, some of us discussed having a lot of the long running tests run the noop drivers so the tests run faster 20:48:57 yeah, we said API tests 20:49:01 since they sometimes hit the timeout 20:49:06 Ah yes.... 20:49:36 The new lower and variable timeout no less 20:49:40 so is there an easy way to change what drivers the gate tests use? 20:49:44 indeed 20:49:53 easy = i can do it in octavia 20:49:57 or does it require a change in infra 20:50:05 It goes in neutron-lbaas 20:50:11 yeah tahts what i mean 20:50:12 sorry 20:50:13 well, can’t we change the octavia.con easily? 20:50:16 it's possible to do with iniset in the gate-hook script 20:50:18 octavia.conf 20:50:20 in gate_hook? 20:50:24 yeah tahts what i thought 20:50:25 rm_work +1 20:50:28 yep 20:50:33 so that'll be ab it more involved that i hoped but i'll start on it 20:50:47 It's really not that bad. 20:50:53 me + bash = bad 20:51:05 but i gotta get better at it 20:51:06 ask rm_work to mentor you :-) 20:51:07 I can do it if you would prefer 20:51:18 no, ill do it, im just complaining :) 20:51:35 ok, settled 20:51:44 meeting time? 20:51:50 ending time? 20:51:51 I think we keep the slot... 20:51:56 ohhh 20:52:06 well i guess i dont see a lot of people wanting the change in time 20:52:10 which is why i asked 20:52:11 in the thread 20:52:12 yep 20:52:42 but it’s probably good to check once in a while... 20:52:43 what originally brought it up? 20:52:58 sbalukoff and his EMEA IBM people + Evgeny wanted it earlier 20:53:00 xgerman: Yep, keep the slot for now. 20:53:29 xgerman: I will poke the IBM research team again on this, and I can get them motivated to start attending these meetings (as they should), we can revisit the meeting time again then. 20:53:38 ah okay well if we hit a larger amount of contributors we can do the split one that neutron does 20:54:00 or we do an ASCTIVE-ACTIVE spin-off 20:54:13 yeah though discussing meeting time during a meeting that people can't attend due to the time... is lulz 20:54:13 blogan: I'm hoping we can find a time in the morning that works well for most... and doesn't make me want to gouge out my eyes. ;) 20:54:19 but i guess if no one responded on the ML <_< 20:54:27 sbalukoff: 7am est 20:54:28 rm_work: Yep, but nobody much has said much on the ML abou tit. 20:54:33 blogan: I will destroy you. 20:54:44 sbalukoff: +2 20:54:46 blogan Enjoy talking to yourself.... 20:54:46 sbalukoff: no need, that time will do that for you 20:54:51 I am core on "destroying blogan" 20:54:52 Haha! 20:55:11 anything else (serious)? 20:55:21 nope 20:55:23 #endmeeting