16:04:08 #startmeeting openstack-ansible-meeting 16:04:09 Meeting started Thu Apr 20 16:04:08 2017 UTC and is due to finish in 60 minutes. The chair is mhayden. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:04:10 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:04:13 #chair spotz major 16:04:13 The meeting name has been set to 'openstack_ansible_meeting' 16:04:14 Warning: Nick not in channel: major 16:04:15 Current chairs: major mhayden spotz 16:04:20 hehe 16:04:27 #unchair major 16:04:27 Current chairs: mhayden spotz 16:04:32 #topic Roll Call 16:04:36 * mhayden hands it off to spotz ;) 16:04:42 hehe, this is why I have issues! 16:04:57 \o/ 16:05:04 Continue with roll call:) 16:05:12 o/ 16:05:19 o/ 16:05:20 o/ 16:05:22 * mhayden isn't running the meeting, but i can do typey typey as needed :P 16:05:35 o/ 16:05:43 o/ 16:05:46 o/ 16:06:04 #topic Old Action Items andymccr to create spec for uwsgi/nginx changes 16:06:20 ok so! i will have that done today :P im literally finishing it up right now 16:06:32 so i will ping in #openstack-ansible at some point after the meeting 16:06:43 Woot! 16:06:47 \o/ 16:06:55 #topic Old Action Items andymccr create list of OSA related talks/sessions at the Summit 16:06:59 ok this one i have done! 16:07:05 Cause it's the andymccr show:) 16:07:08 Link? 16:07:11 #link https://etherpad.openstack.org/p/osa-summit-talks 16:07:18 if you have a talk and its not on there please feel free to add it! 16:07:25 o/ 16:07:27 i did a search for ansible and OpenStack-Ansible and copied any over 16:07:41 Mine aren't OSA but I'll be there:) 16:07:51 spotz: feel free to add them if you'd like! 16:07:56 oh. as a side note 16:08:19 I think on the bottom of that ether, lets everyone add if you're going and contact info so we can try to hang at least once? 16:08:30 it would be super useful to have as many people as possible at the project onboarding session (tuesday lunch time) and the ops feedback (Wednesday just after lunch) 16:08:38 andymccr: I should be there, I think 16:09:00 I can be there Wednesday though might be late have to come from the lunch and learn 16:09:01 asettle: excellent :) im not sure i can answer all questions from operators by myself - and would be good to have a mix of everybody involved if possible 16:09:08 spotz: thanks! 16:09:12 Yeah should be able to try and help 16:09:59 I don't remember if we need emails or phone numbers? 16:10:12 good call cloudnull 16:10:32 ;) 16:10:48 Last action item 16:10:51 #topic Old Action Items https://review.openstack.org/#/c/454742/ reviews for logan- 16:11:09 Pretty sure we did that no? 16:11:15 so i think this has some reviews - logan- do you need more, or is this looking good? 16:11:20 do you have enough to work on? 16:11:49 yes good on it right now. i am working on some tests for it 16:11:53 sweet :) 16:11:55 great 16:11:56 thanks jmccrory for taking a stab last week 16:12:19 np 16:12:28 Sweet, see how much we rock!:) 16:12:34 Hahahha 16:12:35 \o/ 16:12:40 #topic OPNFV Cross Community CI (XCI), its relation to OSA and where we are now 16:12:47 fdegir you here? 16:12:54 or hwoarang 16:12:59 this might be a carry-over from last week - since we spoke about that i believe 16:13:11 but i didnt want to remove it incase we needed to talk more :) 16:13:22 but odyssey4me i know you did some work to help on this, you may be interested in the chat logs from last week! 16:13:25 i am around but i thought fdegir talked about that last week 16:13:33 ok 16:13:41 ah, thanks andymccr 16:13:42 he did but Andy left it on the agenda.... 16:13:43 if you need extra help or anything, just ping 16:13:52 haha yes yes i take responsibility :P 16:13:52 sure thank you evrardjp 16:13:54 Bad andymccr 16:13:59 andymccr did you update the agenda yet? 16:14:01 :p 16:14:09 he half did odyssey4me 16:14:14 I hear that the agenda should really get updated. :p 16:14:15 odyssey4me: see i left some tricks in to see if others were paying attention 16:14:22 well there could have been updates 16:14:23 apprently they AREN'T :P 16:14:24 :) 16:14:35 hehehe well moving on and back to andymccr 16:14:40 don't be too harsh on andymccr, he is our benevolant dictator 16:14:41 #topic Release Planning and Decisions 16:15:05 ok - next week is more releases - however it's mitaka EOL coming up 16:15:15 I've asked for our role repos to be EOL'd and to leave our main repo in place 16:15:28 the plan will be to SHA bump to point at mitaka-eol for roles and upstream projects and then mitaka-EOL the main openstack-ansible repo 16:15:42 +1 16:15:54 other than that business as usual :) if anybody has anything specific to go in. let me know 16:15:59 oh wait. Can't vote on intentions. 16:16:02 :p 16:16:09 also the CVE fixes for Horizon & Nova are in the latest release for OSA 16:16:31 * mhayden pours one out for mitaka 16:16:36 hahaha yeah 16:16:38 hahaha 16:16:51 alas, mitaka, it served us well 16:16:54 that was the first release where i broke less stuff 16:16:56 it did indeed 16:16:56 * mhayden sniffles 16:17:02 lol 16:17:08 hehehe 16:17:22 #topic Blueprint work 16:17:51 so aside from the uWSGI bits i dont think we have much 16:17:59 asettle: we could get an updateon the docs ops guide? 16:18:00 monitorstack is moving along 16:18:24 andymccr: sure. So we have a few more contributors rolling in from the Rackspace side. 16:18:34 mhayden: excellent, you may need ot ping for reviews since i'm not sure most ppl will be monitoring monitorstack (get it?) 16:18:37 I'll be having a meeting with the ops team next week about what they're up to 16:18:45 andymccr: what you did there -- i see it 16:18:45 I'll also be doing my best to plug it at the summit. 16:18:51 andymccr: ha! 16:18:58 as an aside, the latest OpenStack user survey shows Ansible on top 16:18:59 asettle: excellent! 16:19:01 Got a few new links here 16:19:05 links - ugh, names 16:19:06 although it is broken into 3 pieces 16:19:08 #link https://wiki.openstack.org/wiki/OpenStackAnsible#Documentation_planning 16:19:11 sweet ! 16:19:12 moving along 16:19:24 nicely - i'd like to get that done for Pike 16:19:35 or at least moved to a point where its not "draft" 16:19:44 Yeah, that should be on target. 16:19:54 :) 16:20:06 Any more blueprint talk? 16:20:17 not from me :) 16:20:18 so we'll need to start talking about the py35 bits 16:20:33 and id really appreciate somebody trying to take that one on 16:21:30 i know you're all trying so hard to be the first to type "me" that you're falling off your chairs - but let me know if you'd be ok with helping out on that :) 16:21:52 Have you seen my python code?:) Yeah so !me:) 16:22:02 I wouldn't know python if it hit me in the face. 16:22:07 Unless it's a snakey 16:22:10 Then I got that! 16:22:24 Who's not here.... Jimmy?:) 16:22:32 haha its ok we'll figure it out :) 16:22:36 but for now just to put it on the radar! 16:22:36 jmccrory mgariepy cloudnull 16:22:40 I'm super good at tagging people 16:22:42 logan-: 16:22:45 andymccr we have very little python of our own, and last I asked and did a cursory look most of our stuff was already py35 compatible 16:22:53 it'd be good to have some sort of test to validate 16:22:57 I can add translation code really well to python:) 16:23:02 odyssey4me: exactly - i dont think it'd be that hard 16:23:06 woo hoo , i was tagged. 16:23:08 i'm not good at python 16:23:10 odyssey4me: its pretty much testing and ensuring that the infra works properly. 16:23:25 odyssey4me: I don't know I think you volunteered:) 16:23:31 I think odyssey4me did 16:23:38 #action odyssey4me to investigate py35 requirements 16:23:39 is there a place to see how far along each openstack project is for py35? 16:23:41 Phew, you're welcome andymccr 16:23:41 I seem to recall cloudnull going down the road of trying to make ansible run on py3 16:23:52 andymccr: I did a py35 thing with the pip install role 16:23:54 but yeah, I'll happily take a look at adding some sort of py3 tests 16:23:57 jmccrory: im trying to find that out - for the uwsgi work too because its hard to keep track of that otherwise 16:23:58 voluntold! 16:24:06 thats a first step 16:24:15 hehehe, we're so bad:) Hi-5 to asettle 16:24:18 but i've not circled back on it 16:24:25 spotz: NICE :D 16:24:37 Well that leads us to.... 16:24:45 #topic Open Discussion 16:25:03 2.3. Great. 16:25:17 That's all I had to say for today. 16:25:43 so there's been a topic of consolidating containers 16:25:45 evrardjp: hahaha yeah how are we with that? i know jmccrory was fixing it up a lot 16:25:49 as in, don't have heat-api and heat-engine 16:25:51 just put them into one 16:25:59 also we merged the machinectl things 16:26:00 since if you lose the engine, the api is worthless 16:26:10 which would be great for folks to beat on 16:26:41 mhayden: yeah i saw that - im concerned about upgrade paths 16:26:41 andymccr patches are up for all integrated roles, most have merged by now 16:26:44 andymccr: I didn't start patching. I think the discussion with jmccrory was pending. Then went to do different things. 16:26:58 mhayden: i think its a nice approach tbh. 16:26:59 jmccrory: great. thanks. 16:27:12 There is an issue when upgrading to newton from mitaka when having old stuff in /etc/openstack_deploy/env.d/ 16:27:14 mhayden: and I have a couple prs in flight for removing a few of our container groups and moving the services a combined container https://review.openstack.org/#/q/status:open+project:openstack/openstack-ansible+branch:master+topic:hyperconverged 16:27:25 it'd be great to get some thoughts on that too 16:27:39 oh. 16:27:40 haha 16:27:41 the point of the multiple containers is to allow scale - it's not really anything to do with anything else 16:27:43 what major said 16:27:43 also, on the nova side, we can combine scheduler/conductor/api together 16:28:04 cloudnull: mhayden: im personally in favour of it if we can fix the upgrade path up to work nicely 16:28:05 we could, as a default in OSA, have combined containers on a per project basis - as long as the functionality to split them out is still there 16:28:20 it might be nice to have env.d differences for hyperconverged, as andymccr mentioned 16:28:24 so someone could pick that option 16:28:30 yeah we could default to converged 16:28:34 per service 16:28:44 I like the option path though might be hard to code 16:29:12 well, doing this would require a dynamic inventory change 16:29:26 odyssey4me: sure it would just require an env.d change 16:29:31 *surely 16:29:47 andymccr well, for one option yes - to have multiple pre-set options I don't think so 16:30:00 unless we provide a set of override files, but that gets messy 16:30:06 odyssey4me: i dont think it gets that messy 16:30:19 you have an env.d which has optional different env.d files per service to not converge 16:30:26 so, well, what's the purpose of doing this - what do we gain? 16:30:32 so we have the scale groups already. 16:30:41 fewer containers really 16:30:42 but they're largely un documented. 16:31:06 I think the documentation is on the way into the ops guide? 16:31:09 so we could combine the defaults that we have already and then document how to scale each group 16:31:18 ah. 16:31:22 then we may already b eclose 16:32:04 limitations of the inventory, again on the table. 16:32:14 #link https://review.openstack.org/#/c/442091/ 16:32:18 ^RE py3 support 16:32:29 oh nice :) 16:32:31 well, to me there's limited value and it's a massive pain for upgrades - so unless we change the approach to how we work with containers I'd recommend against doing this 16:33:16 perhaps a spec would help illustrate the aims youre going for mhayden cloudnull? 16:33:16 RE: py3 -- i need rhel help :) 16:33:34 that may be helpful 16:33:35 mhayden: mgariepy ^ 16:33:43 ie if we no longer need to ssh into containers, and no longer need to identify them as hosts, then we can change them around at will 16:34:43 cloudnull: mhayden would you be up for adding in a spec for that? 16:34:51 i could possibly do that 16:34:53 I can try and crank out a spec 16:34:54 sweet :) 16:34:56 or mhayden 16:35:01 cloudnull and i can collab 16:35:04 #action add spec for hyper-converge - mhayden cloudnull 16:35:08 woot ! go team! 16:35:08 we are collabing a lot lately 16:35:11 hahah 16:35:13 YAY TEAM 16:35:17 ok so one other topic i want to bring up 16:35:24 the user survey came out. im following up on a few things. 16:35:28 Gartner is telling me that hyperconverged is where it's at 16:35:38 mhayden: is it bimodal though? 16:36:10 on the user survey - im following up around the way projects get listed as "projects" essentially, since ours isnt listed - i think that has a negative impact and i'd like to get that fixed 16:36:20 (for next time, since this one is done!) 16:36:26 andymccr: I saw ansible is +9000 but osa is not listed. :'( 16:36:32 cloudnull: yeah basically thats the problem 16:36:50 there is also a projects section that includes some deployment projects but not others. 16:36:59 ansible (LXC) - is, but i think that's a write in 16:37:07 odyssey4me: andymccr - I think they just call us Ansible, and the reason I say that is the video of the new mascots 16:37:27 spotz: link? 16:37:30 so im making some queries and hopefully asking the right ppl the right questions and hopefully we'll get something moving for next time! 16:37:45 cloudnull: saw it on facebook but it was from sueruser let me look 16:38:26 cloudnull: http://superuser.openstack.org/articles/openstack-project-mascots/ 16:38:58 Here we are listed as openstack-ansible though https://www.openstack.org/project-mascots 16:39:39 ah i see in the video, we're "ANSIBLE" 16:39:51 either way the question wouldnt be that specific in the survey i think 16:39:56 but i think it shows we have good traction 16:39:59 Yeah but on that second link we're openstack-ansible so who knows:( 16:40:03 either way - we just need to be in the right sections too :) 16:40:10 WE ARE ANSIBLE. WOOT. 16:40:24 thats all i have for this week afaik :) 16:40:58 Anyone have anything else? 16:41:23 Nothing. Thanks spotz and everyone! 16:41:29 #action Everyone going to summit put your contact info on https://etherpad.openstack.org/p/osa-summit-talks 16:41:47 #endmeeting