20:00:29 #startmeeting octavia 20:00:30 Meeting started Wed Aug 19 20:00:29 2015 UTC and is due to finish in 60 minutes. The chair is xgerman. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:31 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:00:33 The meeting name has been set to 'octavia' 20:00:41 #chair johnsom blogan 20:00:41 Current chairs: blogan johnsom xgerman 20:00:47 o/ 20:00:49 hi 20:00:49 o/ 20:00:50 Howdy folks! 20:00:51 o/ 20:00:51 hi 20:00:59 o/ 20:01:02 hi 20:01:13 I am plagued by plumbing problems so might need to step out and talk to the plumbers 20:01:44 #topic Announcements 20:02:11 hi 20:02:13 You need Donkey Kong to throw barrels at them... 20:02:22 lol 20:02:34 Heh! 20:02:39 with Neutron proliferating into so many projects we put a huge load on infra so 20:02:52 so dougwig proposed https://review.openstack.org/#/c/212622/6/doc/source/devref/infra.rst 20:03:00 o/ 20:03:05 sounds like infra's got a plumbing problem too 20:03:14 o/ sorry is late 20:03:17 yep 20:03:29 lol 20:03:33 surround your house with more carnivorous plants 20:03:35 so for us we need to get dougwig’s +1 before infa looks at it 20:04:03 if dougwig is here he can elaboaret 20:04:06 dougwig's approval is hard to come by 20:04:18 yep, he will be even more a bottleneck 20:04:37 i think he's at the ops meetup so probably not paying attention 20:05:00 Yeah, and ask dougwig the infra questions before bothering the folks in the infra channel 20:05:03 yeah, anyhow don’t ask in the infra channel for +2 until you have dougwig’s +! 20:05:14 johnsom +! 20:05:23 noted 20:05:24 Heh! 20:05:38 johnsom anything from GSLB worth mentioning/ 20:05:40 ? 20:06:00 They are working on some architecture diagrams 20:06:12 Did y'all pick a new name yet? ;) 20:06:27 we celebrated Cosmos last week :-) 20:06:32 Kosmos 20:06:36 They(we I should say) are also looking at some contributed code called "polaris" as a start to health monitoring for GSLB 20:07:12 xgerman: Yes, I know, and pointed out how you're definitely going to run into trademark issues with that. But eh... whatever. Not my project, eh. 20:07:39 yeah 20:07:39 * blogan_ votes sbalukoff official openstack trademark validater 20:07:45 Haha! 20:07:59 sbalukoff is discovering his lawyer side by working for IBM 20:08:10 xgerman: If only you knew. I 20:08:11 hahaha 20:08:24 heh 20:08:24 more announcements? 20:08:42 've spent a good portion of the last week finally going through new employee orientation stuff (couldn't put it off any longer) and otherwise navigating the internal vagaries of the company. 20:08:52 Not from me! 20:08:58 #topic Brief progress reports 20:09:06 Fyi, there are three infra liasions, not just me. :) 20:09:21 dougwig you're ours 20:09:23 Not anything substantial from me this week. I hope to correct that by next week. :P 20:09:32 L7? 20:09:33 Oh, but we want to give you the love dougwig 20:09:46 * dougwig blushes 20:09:47 Something about all of those "octavia bug" comments.... 20:09:56 xgerman: I should have a new patch for that later this week / early next week. Sorry! 20:10:06 nice! 20:10:08 Progress: With some help from johnsom and blogan_ I've got a near complete Failover. 20:10:12 Yeah, bana_k has been submitting a lot of those. That's great! 20:10:13 i'm working on addressing xgerman's comments on this review: https://review.openstack.org/#/c/209210 20:10:34 TrevorV: Rad! 20:10:34 unit tests done , 20:10:34 I have been working on getting the health manager setup with an oslo config file and wiring it up to startup, etc 20:10:42 I finished including DB Cleanup, in addition to Spare Amphora, under HouseKeeping.. Tested it on devstack and works fine... 20:10:42 testing the application 20:10:49 i got a +1 from dougwig for my octavia gate setting, but still need another +2 from infra team https://review.openstack.org/#/c/211319/ 20:10:53 need to do more 20:10:56 xgerman: but i think doing another review with only test to get the full coverage you want would be ideal instead of putting it in this one bc that makes sense as a separate review to shore up test coverage 20:11:11 Mostly a lot of testing for health manager now. Next up is aligning the status messages on both sides 20:11:22 blogan_ I am ok with that 20:11:32 but I like the spec fixed before I give +2 20:11:45 xgerman: was going to do that too in this review 20:11:48 poking around lbaas scenario tests, all of the patches got merged. Thanks to reviewers 20:11:55 awesome!! 20:12:02 madhu_ak: thanks for getting that done adn your patience 20:12:09 and everyone else who worked on it 20:12:12 FYI, housekeeping is ready to merge in my opinion: https://review.openstack.org/#/c/202829/ 20:12:23 ready to merge eh? 20:12:25 blogan_ +! 20:12:27 bold statement sir 20:12:28 Nice! 20:12:38 we are bold here 20:12:47 johnsom: some pedantic people might see that as a challenge 20:12:58 * blogan_ looks at TrevorV 20:12:59 It's not my code, so go for it 20:13:03 XD 20:13:07 * blogan_ maybe himself 20:13:07 dougwig: I just submitted a patch: https://review.openstack.org/#/c/214767/ and need +1 from you before infra folks gives me +2. 20:13:29 ^^ the above patch is making scenario test api job as voting 20:13:35 also with L3 coming up on 8/31 we need to keep up the velocity 20:13:45 Yep. 20:13:57 madhu_ak: i have noticed it is much more stable, i haven't seen it fail in a while, have you? 20:14:19 blogan_ yep, it is much more stable now :) 20:14:25 madhu_ak: for random reasons 20:14:31 beyond our control 20:14:33 * TrevorV pedantic senses are tingling 20:14:53 things will be stable and from ym experience I dont see it any failures 20:15:00 *my 20:15:03 Very close to a functional and stable active/standby code 20:15:11 +1 20:15:12 awesome 20:16:12 ok, moving on 20:16:17 #topic Octavia reference implementation in Liberty 20:16:31 #link https://etherpad.openstack.org/p/YVR-neutron-octavia 20:16:42 let’s see where we are at 20:16:54 https://review.openstack.org/#/c/213597/ 20:17:20 yeah, ajmiller? 20:17:22 neutron-lbaas cores (ptoohill, ajmiller) need a +2 (or a -1 if so needed) 20:18:01 or -1 from anyone, but that's gets the octavia driver to a good state as far as reporting accurate provisioning statuses 20:18:05 xgerman here, looking at it 20:18:11 thanks 20:18:18 Had looked through it yesterday, but before the latest patchset. 20:18:19 TLS? 20:18:29 If I could draw some attention here in a second... I've noticed, through random testings in the flows, that we have a lot of un-implemented reverts in the tasks. Would anyone have a problem with me spending some significant time updating individual tasks accordingly, and then optimizing flows to utilize subflows much more effectively when duplication is necessary? 20:18:29 pothole was working on that 20:18:53 TrevorV: you're asking if anyone woul dhave a problem with making the code better? 20:19:11 yeah, we all love that - thanks for volunteering :-) 20:19:13 I'm asking if someone is currently interested in doing that or has progress on it, or if I should prioritize something else 20:19:41 Obviously someone making the code better isn't a problem, but it could be currently back-burnered, make sense? 20:19:43 Dooo eeet! 20:19:48 +a 20:19:50 +A 20:20:07 well, I think we need to get through our list first and then we can work on optimzing 20:20:08 It needs to be done. 20:20:12 Alright, then that's what I'll focus on after fixing the failover flow 20:20:21 TrevorV: plus some tasks dont need a revert, tasks that just retrieve values usually wouldn't 20:20:22 TrevorV +1 20:20:39 grr 20:20:41 grr 20:20:42 Yeah, blogan_ I'm more talking about the example in failover... It leaves ports and stuff floating around. 20:20:43 Yeah, I would prioritize failover flow over the reverts. It's dependency chain is getting long 20:20:43 double grrs 20:21:00 yeah agree with johnsom 20:21:04 +! 20:21:06 +1 20:21:11 finally 20:21:17 but i think TrevorV already had that planned 20:21:18 I just thought it'd be nice to have it all reverting appropriately in Liberty, and then optimization of flows will clean it up nice-like. 20:21:28 Yep, what TrevorV said.... 20:21:29 yep, agreed 20:21:32 it will be, but separate review unless necessary for the failover flows 20:21:38 Yeah, for sure 20:21:39 YES 20:21:45 I'm going to finish failover, and then make new review on master 20:21:49 off master*** 20:21:50 my b 20:21:51 excellente 20:21:52 sound sgood 20:21:56 so pothole? 20:22:03 any TLS news? 20:22:04 ptoohill? 20:22:12 i think he fell in a pothole 20:22:20 heh 20:22:25 lol 20:22:58 mwang2’s gate job is waiting infra +@ 20:23:10 blogan Deletes? 20:23:22 xgerman: in the driver? 20:23:26 yep 20:23:35 xgerman: TrevorV fixed that last week I believe 20:23:45 xgerman: its in 20:23:57 other than the issues about json.decode someone brought up earlier ir ead about 20:24:01 dont know much about it though 20:24:03 bana_k said he ran into trouble yesterday 20:24:17 bana_k: can you give more details on it? 20:24:49 I was just trying to create n delete the listener n pool 20:24:50 Wait, deletes are still failing? 20:24:58 I tested all the pieces I thought... May have missed something. 20:25:11 first time listener delete failed 20:25:27 but used to succeed in second time 20:25:41 took a short break n tested that again n it started working fine 20:25:53 but delete pool started failing 20:25:53 bana_k: hmm one of those wonderful hard to reproduce bugs 20:25:54 I haven't reloaded neutron-lbaas in my devstack in a while, so can't tell you if it's fixed. Will do that next restack 20:26:12 blogan_, I'm not sure the delete fix is merged yet, is it? 20:26:18 TrevorV: yes 20:26:21 If not, then bana_k are you sure you have the fix in code? 20:26:23 but delete pool was stuck in that state for all the time 20:26:38 TrevorV: i combined yours and the tls one and bana_k into the same review last week or this weekend and it merged 20:26:42 First time fail, second complete is mis-match between neutron db and octavia db. It's probably a false success 20:27:17 have yall updated your localrcs to not pull from that review's patchse? 20:27:18 Okay blogan_ 20:27:21 patchset? 20:27:24 just need to pull master now 20:27:27 No I need to take a look at the code 20:27:35 and if you want good status management, pull down https://review.openstack.org/#/c/213597/ 20:28:24 bana_k, let me know if you ARE up-to-date with my delete fix, because if there is still a bug there then I'll jump on it as quick as I can 20:28:31 \me stepping away for a bit 20:28:54 I have to step away, sorry 20:28:55 brb 20:28:57 bana_k: verify your localrc settings first, that neutron-lbaas isn't pulling down an old review 20:29:31 it was pointing to master last time I checked 20:29:36 bana_k: damn 20:30:26 so ll see if those changes are there in the code n let u know 20:30:52 bana_k: you able to do a git log right now and tell me the hash you have? 20:31:13 bana_k remove /opt/stack/neutron-lbaas as well, just to make sure it re-clones 20:31:18 yeah 20:31:29 or just git pull 20:31:42 ok 20:31:56 back sorry 20:31:57 anyway sounds like thats everything 20:32:18 Short meeting today, eh? 20:32:20 xgerman will have to come back to an empty room 20:32:26 git log shows the latest checkins 20:32:34 fec0535a85945ed43471fd0007487864d4d7e775 20:32:41 commit id 20:32:41 What about the stats stuff? 20:32:56 does this count as "open discussion"? 20:33:30 TrevorV: yes 20:33:32 so basically it has master code base 20:33:38 i dont have chair under this alias so i cant change topic 20:33:51 bana_k: i dont see that commit! i'm going crazy 20:34:05 bana_k: octavia or neutron-lbaas? 20:34:11 #chair blogan_ 20:34:12 neutron-lbaas 20:34:12 Current chairs: blogan blogan_ johnsom xgerman 20:34:20 #topic Open Discussion 20:34:27 German added me as chair in case he had to go 20:34:28 johnsom: thanks 20:34:40 hello. 20:34:59 So I saw a remark at the top of the meeting about infra plumbing problems, and see in zuul that no new instances are building. Gate is way backed up 20:35:10 Anybody know what's up? 20:35:36 ajmiller: i do not but i suppose our infra liasion dougwig would know 20:35:41 It's getting close to the end of liberty 3 is my guess 20:35:53 sorry about that 20:36:30 bana_k: whats the title of that commit? 20:37:17 xgerman: bana_k and i can take this discussion to the main channel 20:37:28 xgerman: not really something to do in a meeting 20:37:30 blogan_ with the gate in this state, even if it started building jobs now, we'll probaby be lucky if that patch merges this evening sometime. 20:37:37 k 20:37:45 ajmiller: yeah i know unfortunately 20:38:56 We need to raise the importance of https://bugs.launchpad.net/taskflow/+bug/1479466 20:38:56 Launchpad bug 1479466 in taskflow "The needs to treat subflows as standalone entity" [Undecided,New] 20:39:45 There is a review in progress to handle subflows as a single entity in taskflow : https://review.openstack.org/#/c/136171/ 20:40:43 mmh, wonder if harlow can comment? 20:40:48 jharlow 20:41:04 He has a bunch of comments on the bug 20:42:01 ok, abdelwas make sure to reach put to the other Min 20:42:09 Sure 20:42:45 also #topic Open Discussion 20:42:53 As long as I am not writing a spec, and implement it myself as the comment suggested :) 20:43:00 #topic Open Discussion 20:43:08 tep 20:43:09 yep 20:43:29 abdelwas: you want it you implement it! 20:43:40 abdelwas: you want a teleportation device? you invent it! 20:43:53 :) 20:44:02 :-) 20:45:32 anything else? 20:45:48 Nope! 20:45:50 #endmeeting