16:00:55 #startmeeting openstack_ansible_meeting 16:00:56 Meeting started Thu Nov 17 16:00:55 2016 UTC and is due to finish in 60 minutes. The chair is asettle. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:57 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:01:01 The meeting name has been set to 'openstack_ansible_meeting' 16:01:18 #rollcall 16:01:24 \o/ 16:01:25 o/ 16:01:34 o/ 16:01:36 o/ 16:01:40 o/ 16:01:54 Oh guys lovin' the attendance! Good vibes, feeling it! 16:01:58 o/ 16:01:58 o/ 16:02:04 o/ 16:02:17 hi automagically 16:02:19 https://admin.mashable.com/wp-content/uploads/2013/07/Jeremy-Renner.gif 16:02:41 o/ 16:02:44 I'll give it another minute for people to trickle in :) 16:02:46 o/ 16:02:48 oh dear, that gif is freaking me out 16:02:52 hahahaha 16:02:55 Okay let me find another happy gif 16:03:01 Hawkeye shouldn't be that happy 16:03:02 lol 16:03:07 odyssey4me: the incentive is for you to say more so the gif disappears from your feed! :P 16:03:09 https://media.giphy.com/media/iPTTjEt19igne/giphy.gif 16:03:24 Alrighty let's boogie 16:03:37 o/ 16:03:45 Please update the agenda with anything you'd like to talk about if you haven't done so already 16:03:47 oh man, am I at the right place? 16:03:49 https://wiki.openstack.org/wiki/Meetings/openstack-ansible#Agenda_for_next_meeting 16:03:50 Link is here 16:03:52 Okay, starting off 16:04:02 woot! 16:04:07 #topic Review action items from last week 16:04:12 andymccr: please take it away 16:04:39 ok there are 3 action items from last 16:05:00 sign up for the PTG! can't speak for everybody on that one, so i'm gonna add that as another action item (and cover off the first topic) 16:05:14 #action sign up for the PTG if you're attending! 16:05:23 also there are limited spaces 16:05:30 #link https://www.openstack.org/ptg/ 16:05:31 For the PTG 16:05:35 as a heads up for those of you who want to come but haven't registered yet 16:05:40 Also - refunds up until Feb 17 16:05:48 ahh good tip palendae 16:05:49 Good point, thank you palendae 16:05:50 So if you dunno, you can grab one and refund it later 16:05:50 good to know palendae 16:05:59 So, you can buy, and refund if your company says 'no' 16:05:59 the tickets are $100 16:06:04 USD 16:06:08 evrardjp: Internally we've been told to get them then refund if necessary 16:06:10 I registered. 16:06:19 wanted to pay with zim dollars, they said no :( 16:06:20 anyway! 16:06:25 palendae: I didn't know there was a deadline for refunds 16:06:35 andymccr: haha 16:06:43 action item number 2 was to file a blueprint for upgrade testing, which I have done! 16:06:53 #link https://blueprints.launchpad.net/openstack-ansible/+spec/upgrade-testing 16:06:59 good! 16:07:08 there are some POC's up, although the one for keystone which jmccrory has done is the best to look at now 16:07:08 * mhayden stumbles in 16:07:19 Heya mhayden :) 16:07:21 jmccrory: if you're about and could update your keystone PR to point to the blueprint that'd be appreciated 16:07:52 sure will do 16:07:55 thanks :) 16:08:06 Okay, and one more from last week, andymccr :) 16:08:10 #link https://review.openstack.org/#/c/384269/ 16:08:14 odyssey4me looking into adding a prototype experimental integrated build upgrade test (as a periodical) 16:08:16 last action item! 16:08:31 yeah, I haven't managed to do that just yet - been stuck in other work 16:08:35 sweet 16:08:37 lets just add it again 16:08:40 carry that please, I aim to push something up for next week 16:08:46 #action odyssey4me looking into adding a prototype experimental integrated build upgrade test (as a periodical) 16:08:51 excellent action items done! 16:08:52 Great! Thanks team 16:09:02 andymccr you can add an experimental job to the integrated repo for an 'upgrade' scenario 16:09:03 Thanks andymccr 16:09:09 odyssey4me: ahh yeah i'll update that PR 16:09:11 maybe just 'upg' to keep it short 16:09:13 First action item for this meeting.... 16:09:16 Drum roll please 16:09:21 #topic Upgrade testing! 16:09:23 (er, again) 16:09:25 haha ok - 16:09:30 so we covered this a bit but hte blueprint is up. 16:09:41 we need reviews - thanks odyssey4me for already reviewing. I'll be updating the glance/swift POCs 16:09:46 and the keystone one is already up 16:09:51 #link https://blueprints.launchpad.net/openstack-ansible/+spec/upgrade-testing 16:09:57 the blueprint is tracking the work 16:10:06 i'm adding the test infra for those upgrade tests too 16:10:18 Nice 16:10:22 Thanks andymccr 16:10:41 i think thats that. but yeah feedback wanted 16:10:44 Great! 16:10:45 Thanks 16:11:01 #action feedback wanted from OSA team on upgrade blueprint 16:11:09 sneaky :P 16:11:09 That'll keep you all on your toes ;) 16:11:15 #topic Trusty deprecation in master 16:11:19 andymccr: I'm shocked to see it is YOU AGAIN 16:11:24 Take it away, gingey 16:11:31 omg 16:11:38 OK! So as a result of ansible 2.2 xenial is now much quicker and well within the 1.5 hour test time 16:11:48 so we are not seeing any failures for timeouts (as far as I'm aware) 16:11:56 I would like to propse that we now deprecate trusty in master 16:12:07 remove tests - remove the code paths 16:12:14 Poor trusty 16:12:17 oh also remove code path? 16:12:27 andymccr: That's in spite of mhayden's bug report on python perf? 16:12:29 like force break stuff? 16:12:34 sure - if you figure out the complex regex filter to ensure that trusty is voting for newton and not on master, but xenial is non-voting for newton and voting for master :) 16:12:50 and FYI - deprecate is not the right word 16:13:04 deprecate would mean we keep it for a cycle, then remove it - you're proposing removal 16:13:06 palendae: the issue is upstream projects are dropping trusty 16:13:15 andymccr: Sorry - I meant about the 2.2 performance 16:13:15 odyssey4me: nice catch. yes removal! 16:13:21 agreed with odyssey4me's wording here... 16:13:28 Not about dropping trusty 16:13:32 it would also means deprecation in newton 16:13:38 the performance for both master and newton is now acceptable for us to do this 16:13:46 palendae: major's bug relates to xenial itself and not ansible really. as far as i read at least 16:13:48 odyssey4me: Ok, question answered, thanks 16:13:49 we've done well to optimise 16:13:54 i mean im open to leaving it in if there is a massive need 16:13:58 yeah, the python performance bug w/canonical is still pending 16:14:09 but i think it'll optimize overall - and it also has to be done at some point 16:14:25 andymccr I would suggest waiting until the next SHA bump takes, as there're a lot of newton changes going in which are optimisations there 16:14:25 Are other projects keeping their trusty gating jobs around for older series? 16:14:38 Or have they all dropped series that supported Trusty? 16:14:41 mhayden: but you produced better performance by doing manual actions, right? 16:14:49 that could be automated in the bootstrap ansible? 16:14:49 palendae trusty was removed in newton for most upstream projects 16:14:59 odyssey4me: Ok 16:15:00 we're the only ones supporting newton on both trusty and xenial 16:15:07 odyssey4me: is the reason to wait so that xenial gating is quicker in newton branch? 16:15:27 (or more reliable at least) 16:15:31 evrardjp: no, those actions weren't effective 16:15:34 andymccr heh, fair point - no need to wait if xenial stays nv 16:15:42 someone has confirmed the performance impact as of today: https://bugs.launchpad.net/ubuntu/+source/python2.7/+bug/1638695/comments/8 16:15:42 Launchpad bug 1638695 in python2.7 (Ubuntu) "Python 2.7.12 performance regression" [Undecided,Confirmed] 16:15:43 yeah - i'll try figure that out :P 16:15:45 #link https://bugs.launchpad.net/ubuntu/+source/python2.7/+bug/1638695/comments/8 16:15:55 but it'd be nice if it comes under time often enough to make both trusty and xenial voting for newton 16:15:55 but are there any issues with removing trusty in current master, and promoting xenial to voting 16:16:00 "50% of the tests (32/64) have a significant variance in performance from which 19/32 are slower (in times ranging from 5-15%)." 16:16:02 odyssey4me: true 16:16:05 ^^ between GCC 4.8 and 5.3.x 16:16:09 we can handle that separately though i guess. 16:16:23 andymccr yep 16:16:27 no issues that I can see 16:16:40 it'll be a nice optimisation to remove the systemd vs upstart stuff 16:16:44 agree 16:17:09 we can also use the package module instead of apt/yum, and that's a backportable optimisation - not sure if someone wants to pick that up? 16:17:13 it does make backporting harder by effectively stopping cherry-picks 16:17:22 so what's the proposal? Deprecate trusty in N, and remove it in master? 16:17:23 So, guys, it looks like we need to call a vote on the removing of trusty from the master branch. 16:17:34 stevelle yep, but the backports are slowing down now as we've hit M1 16:17:50 and that's a price we'll have to pay - the code base has to diverge at some point 16:18:02 evrardjp remove trusty in master entirely 16:18:11 #startvote Remove trusty from the master branch 16:18:12 Unable to parse vote topic and options. 16:18:14 agreed, just one more SHA would be helpful before removals happen 16:18:15 odyssey4me: we can register a bug for that -- it's simple enough to move from one to the other 16:18:19 Argh IRC is hard. 16:18:25 mhayden - can you run the vote commands? 16:18:42 haha i can -- i assume we're allowing yes/no? 16:18:59 yes, please :Pp 16:19:05 agreed with stevelle 16:19:06 andymccr I guess evrardjp is right - we should add a release note to newton making it clear that trusty is deprecated as a platform and will no longer be supported in Ocata. 16:19:13 #startvote Should Trusty be removed in the master branch? Yes, No 16:19:13 Only the meeting chair may start a vote. 16:19:22 asettle: type #chair asettle mhayden 16:19:29 #startvote Should Trusty be removed in the master branch? Yes, No 16:19:29 Begin voting on: Should Trusty be removed in the master branch? Valid vote options are Yes, No. 16:19:31 Vote using '#vote OPTION'. Only your last vote counts. 16:19:33 LAWL 16:19:34 yes 16:19:35 #vote yes 16:19:36 Thanks mhayden :P 16:19:37 #vote Yes 16:19:38 #vote yes 16:19:41 #vote Yes 16:19:42 #vote Yes 16:19:43 #vote Yes 16:19:44 #vote Yes 16:19:44 #vote yes 16:19:49 ok cool 16:19:57 how did the results look? :P 16:20:00 #vote yes 16:20:01 #vote yes 16:20:01 Well, that solved that problem and I learnt how to run a vote in a meeting. 16:20:03 ok the release note is a good idea 16:20:06 heheh 16:20:11 you need to end the vote ;) 16:20:12 Still waiting on the electoral college ;) 16:20:18 OMG 16:20:18 I should have said #vote yes if we do it properly 16:20:19 PANIC 16:20:26 evrardjp: Yeah ;) 16:20:27 #endvote 16:20:27 Voted on "Should Trusty be removed in the master branch?" Results are 16:20:28 automagically: haha! 16:20:28 Yes (9): mgariepy, evrardjp, odyssey4me, andymccr, spotz, logan-, automagically, jmccrory, mhayden 16:20:29 evrardjp: i can leave it to you to do it properly ;) 16:20:34 #vote hellyeah 16:20:36 I can do the release note if someone points me in the right direction:) 16:20:38 Right, well, that was a bit of a landslide. 16:20:46 the land, it slid 16:20:57 #action spotz to work alongside team to write release note for the removal of trusty from the master branch 16:20:58 ZING 16:21:01 Good job team :) 16:21:12 #action andymccr to create blueprint for trusty removal 16:21:16 sweet 16:21:22 so inspiring. 16:21:25 Anything further on this topic? 16:21:36 party to say bye to trusty?:) 16:21:37 if anybody has concerns/things they want to make sure gets done let me know 16:21:46 #action party to say goodbye to trusty 16:22:01 alas, poor trusty, I knew him well 16:22:06 the trust is going to be broken,alas 16:22:12 * asettle wipes away tear 16:22:13 OKAY 16:22:15 Next item 16:22:20 #topic Documentation 16:22:24 Omg *gasp* it's me. 16:22:48 So, some of you may or may not know there was some discussion at the summit and discussion in the channel about working further on the OSA documentation. 16:23:11 There's three topics here, good news first 16:23:28 1. For those who do not know, OSA will be the first deployment guide available from the docs.o.o website. 16:23:33 I'm currently working with the infra team to get this happening. 16:23:36 Exciting times. 16:23:43 This changes nothing for any contributions to the documentation 16:23:45 So no panic there. 16:23:53 I'll keep you all up to date when we're published 16:24:01 2. We are working on a Configuration Guide for OSA 16:24:13 After the install guide was overhauled, we had a lot of config information just dumped in the dev docs. 16:24:20 Planning for the new guide has begun here: 16:24:21 #link https://docs.google.com/document/d/1xeJ_lep7P2e7HLbRFG57Dx4W9s8brkuNIqJmOvheWKI/edit 16:24:30 s/config guide/operations guide/ ? 16:24:42 #link https://blueprints.launchpad.net/openstack-ansible/+spec/create-ops-guide 16:24:43 Blueprint here 16:24:46 Whoops, yes, sorry, oeprations guide 16:24:52 Out of habit I keep writing configuration for some bizarre reason. 16:24:54 Apologies. 16:25:17 We're looking to get some support and help with the Operations Guide - those that know operators using OSA and can help test and contribute knowledge - please contact me 16:25:24 IRC or alexandra.settle@rackspace.com 16:25:40 Please also take the time to review the google doc 16:25:51 Our unofficial planning will happen there before we start moving mountains in the repo 16:26:02 3. We will be working on the Upgrade Guide to work alongside upgrade work happening 16:26:19 If anyone can help out there, that'll be good too. We're looking to work ahead of schedule and rearchitect the guide to fit more into the format of rolling upgrades 16:26:30 It's a hard task, but we have further planning to do to make sure we execute that properly. 16:26:34 If anyone can help out there too, that'd be great. 16:26:37 Alright 16:26:39 *takes breath* 16:26:41 Questions? 16:27:00 none 16:27:16 Hahhaa good good :) 16:27:31 Just be aware there will be some more docs movement out and about, so keep an eye on some changes coming around 16:27:42 We'll be needing reviewers, and operators galore 16:27:53 admin0 has been helping out a bit, but it would be cool to have more eyes 16:27:58 Feel free to suggest anything, we're open minded! 16:28:00 Okay 16:28:02 Cool. 16:28:18 #action everyone to review the Operations Guide draft proposal 16:28:23 Sneaky me again 16:28:29 Okay, shall we move to the next topic? 16:28:46 fine for me 16:28:52 #topic Proposing new Core Reviewer 16:28:56 andymccr: all you home dog 16:29:10 I'd like to propose logan- to become a core reviewer on OSA. 16:29:23 +1 16:29:24 * stevelle claps 16:29:26 +1 16:29:30 Yayyyy logan- 16:29:43 +1 16:29:47 :) thanks! 16:29:50 +over9k 16:29:56 definite yes :) 16:30:11 Wonderful! 16:30:12 * logan- honored 16:30:27 Is there like, a success command? is that a thing? 16:30:32 #success logan- is new OSA core 16:30:35 grats 16:30:37 *shrug* tried it out 16:30:37 sweet, that looks pretty unanimous - thanks for the hardwork so far logan- and look forward to the help/contributions moving forward! 16:30:49 i'll put that one in motion 16:30:56 #action logan- to do core things with vigor 16:30:57 There we go 16:31:02 lol 16:31:03 lol 16:31:06 lol 16:31:12 We expect great things 16:31:15 No pressure 16:31:18 no pressure 16:31:20 haha! 16:31:25 HAHA ZING 16:31:28 Okay, moving on :) 16:31:30 +3.14 16:31:41 #topic Release Planning and Decisions 16:31:46 Everyone, let me introduce to youuu 16:31:47 andymccr: 16:31:50 release day tomorrow. so the release things will happen tomorrow 16:31:52 Your friendly neighbour PTL 16:31:56 Wow.. 16:31:58 What a summary. 16:32:13 if there are critical bugs that are in flight 16:32:16 now is the best time to let us know :D 16:32:28 poke for reviews too 16:32:55 andymccr: do you have a list of those critical bugs? ARe there any left over that people could be working on? 16:33:16 as far as im aware, no, so if anybody has any let us know and we can track those :) 16:33:25 asettle I think andymccr was asking if anyone is aware of any 16:33:34 Oh right, I need a translator. 16:33:41 I somehow missed the 'if' 16:33:42 i'm blind. 16:33:44 Pity me. 16:33:52 ok cool well, i'll get that going tomorrow 16:33:54 otherwise the bug list is there, with priorities for triage - it just takes actions for people to go and verify/triage/fix them 16:34:08 ^ that 16:34:12 asettle sounds like you need a scotch :p 16:34:14 agreed 16:34:15 all that triaging is for a purpose :D 16:34:17 odyssey4me: RIGHT?! 16:34:26 andymccr: really? 16:34:33 evrardjp: no i just like clicking buttons in launchpad :P 16:34:41 :p 16:34:48 On that note, a reminder to all watching the meeting here that our bug triage happens every Tuesday 1600UTC 16:34:59 #action everyone watching the meeting to attend bug triage on PAIN OF DEATH 16:35:04 I'll bring more gifs 16:35:07 Okay, next item 16:35:16 #topic Blueprint work 16:35:23 mhayden: got any updates for us homie? 16:35:35 security work is still moving along 16:35:39 just need some reviews there 16:35:53 i think we're at around 90-100 controls implemented out of 236 but i haven't checked today 16:35:53 mhayden: could you link us? :) 16:35:58 si - un momento 16:36:05 Oh wait 16:36:06 I can 16:36:06 asettle: sorry for the mess up with the time zones for the bug triage :p 16:36:07 #link https://blueprints.launchpad.net/openstack-ansible/+spec/security-rhel7-stig 16:36:18 evrardjp: nein problemo hombre :) 16:36:19 #link https://review.openstack.org/#/q/status:open+project:openstack/openstack-ansible-security+branch:master+topic:bp/security-rhel7-stig+status:open 16:36:29 Gracias mhayden :) 16:36:33 also, the octavia spec merged, so i'll try to get that kicked off a bit 16:36:38 i need to pester andymccr to make a repo ;) 16:36:53 Wahoo 16:37:01 mhayden: sure. openstack-ansible-os_ocativa? also do you have an existing one to copy over or is it fresh 16:37:06 OCATIVA! 16:37:07 #action andymccr to make a repo for mhayden 16:37:15 let's get johnsom to rename the project to ocativa 16:37:20 just kidding 16:37:22 Activia? 16:37:23 mhayden: ok you just got your repo banned 16:37:27 make it yourself :P 16:37:27 openstack-ansible-os_octavia would be fine, andymccr 16:37:33 Hahahhahaha 16:37:37 Okay well that went down in flames fast 16:37:41 palendae: activia would be a nice name 16:37:42 NEXT 16:37:45 Nobody disagree with PTL, Lord and Master, andymccr 16:37:51 Alright, next up 16:37:57 #topic Open discussion 16:37:59 don't get mccrazy on me 16:38:01 hahahahaha 16:38:04 oh one other blueprint 16:38:05 haha 16:38:06 mhayden: he's mccrazy about me damnit 16:38:09 the upgrade-testing one is up now 16:38:11 We are not integrating yogart 16:38:13 * mhayden has no comment 16:38:21 and the ops guide! 16:38:24 but we have talked about those already 16:38:37 We've got a few inventory refactor reviews floating about 16:38:41 Yep, add to teh agenda for next week andymccr :) 16:38:45 palendae: continue fine fellow 16:38:57 I'd link to them, but I'm on the hook to update my rebase chain. So I'll post them once htat's done 16:39:41 palendae: i'll review once you're ready with those 16:39:52 https://review.openstack.org/#/c/392417/ appears to be ready though 16:39:55 Which I need to rebase on 16:39:57 Thanks andymccr 16:40:23 Alright 16:40:26 Any other comments/topics? 16:40:40 Sorry for not knowing how to vote 16:40:41 LEARNING 16:41:15 Alright, you've all stopped paying attention 16:41:16 #endmeeting