16:07:52 <spotz> #startmeeting openstack_ansible_meeting
16:07:52 <spotz> #link https://wiki.openstack.org/wiki/Meetings/openstack-ansible#Agenda_for_next_meeting
16:07:52 <spotz> #topic Roll Call
16:07:53 <openstack> Meeting started Thu Aug 17 16:07:52 2017 UTC and is due to finish in 60 minutes.  The chair is spotz. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:07:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:07:57 <openstack> The meeting name has been set to 'openstack_ansible_meeting'
16:08:07 <asettle> Did you have anything, andymccr ?
16:08:09 <evrardjp> o/
16:08:30 <spotz> We have no action items from last week so...
16:08:37 <andymccr> spotz: mhayden suggested we just do bug smash :)
16:08:37 <spotz> #topic Topics for Discussion
16:08:39 <andymccr> and focus on that!
16:08:47 <spotz> #topic PTG - 11-15 September - Denver, Colorado
16:09:01 <andymccr> i'll give a quick update - again, update the etherpad :)
16:09:10 <spotz> Well PTG is really the only thing we have, cause we're between releases, etc
16:09:25 <spotz> Ther eis one open discussion item
16:09:40 <andymccr> yeah.
16:09:43 <andymccr> also for releases this week
16:09:45 <spotz> vakuznet You here?
16:09:45 <andymccr> we're pushing out an rc1
16:09:48 <andymccr> now that everything has worked!
16:10:05 <vakuznet> yes, I'm here
16:10:31 <spotz> heheh our one open discussion isn't here, so we're done?:)
16:10:38 <spotz> Ahh he is!
16:10:43 <spotz> #topic open discussion
16:10:51 <spotz> #topic backport Use root folder group_vars into ocata
16:11:05 <spotz> #link  https://review.openstack.org/#/c/493216/
16:11:10 <spotz> floor is yours vakuznet
16:11:59 <andymccr> i know evrardjp put that up on the board
16:11:59 <vakuznet> well, I'm told it's up to community to decide. I'd like to see this merged into stable/ocata.
16:12:19 <evrardjp> yeah I think it's worth discussing
16:12:25 <andymccr> vakuznet:  ok - cloudnull mhayden you about?
16:12:32 <evrardjp> I think there are workarounds to do the same thing
16:12:36 <evrardjp> just less convenient
16:12:43 <cloudnull> o/
16:12:44 <vakuznet> much less
16:12:46 * mhayden woots
16:12:54 <spotz> and we need odyssey4me
16:13:04 <asettle> spotz: he just returned to his desk
16:13:08 <spotz> cool
16:13:27 <mhayden> i'm not opposed to this one -- would make future backports a little easier
16:13:41 <andymccr> my concern is upgrade path and existing deployments
16:13:44 <andymccr> logan-: are you about?
16:13:48 <andymccr> jmccrory:
16:13:51 <andymccr> would be good to get your input
16:13:52 <logan-> o/
16:13:53 <odyssey4me> vakuznet that's a major change with a lot of risk - why is it needed in ocata?
16:14:31 <logan-> sorry forgot to meeting... :) reading backscroll
16:14:32 <odyssey4me> especially with a pike release imminent, personally I'd rather just recommend upgrading if you need it
16:14:45 <andymccr> logan-: sure its relaly just on. https://review.openstack.org/#/c/493216/ its a backport to stable/ocata
16:15:00 <odyssey4me> evrardjp also did a good blog post outlining other ways to do much the same using ocata: https://evrard.me/group-and-host-variables-overriding-in-openstack-ansible.html
16:15:13 <vakuznet> ocata is last stable
16:15:27 <evrardjp> this patch would also need a openstack-ansible.rc change.
16:15:29 <odyssey4me> yes, and pike stable is hitting RC today
16:15:38 <logan-> how is our test coverage on the vars plugins
16:15:46 <logan-> -s
16:15:46 <odyssey4me> logan- non existant, I think
16:15:50 <evrardjp> yes :/
16:16:44 <spotz> vakuznet: Is there a reason you couldn't go pike once it's released which would probably be before we could do the backport
16:16:46 <odyssey4me> I'm pretty firm on this one. While it'd be nice to have it in Ocata, it's too much risk to implement. We've had enough code churn, so I'm very definitely a -2 on that backport.
16:17:23 <odyssey4me> Anyone who *really* wants it can happily do their own fork and maintain their fork at their own risk. Anyone who doesn't want to do that can use Pike instead.
16:17:37 <logan-> personally i have not tested this but i remember all of the discussion we had around merging it to master and iirc there is no precedence difference with how we changed it, it should function equivalently to regular group_vars and host_vars
16:17:39 <evrardjp> vakuznet: if there really a good reason we should do it. BUt right now, I don't see one
16:17:49 <evrardjp> logan-: yes it should
16:17:58 <odyssey4me> emphasis on *should*
16:18:02 <evrardjp> we just need to change the openstack-ansible.rc
16:18:09 <evrardjp> and then seen how it goes.
16:18:20 <odyssey4me> we're talking different ansible versions, and many other changes which have been done since implementing that patch originally to solve issues which arose
16:18:23 <evrardjp> However, why change stable?
16:18:27 <vakuznet> you want to release pike, then backport is more likely?
16:18:48 <spotz> We want to know why you need this in Ocata and then determine if it's worth the risk
16:18:49 <odyssey4me> vakuznet Pike already has it. No backport necessary.
16:19:13 <vakuznet> i meant to ocata
16:20:02 <spotz> vakuznet: Is there a reason you need this on Ocata vs using pike which has it?
16:20:45 <vakuznet> we use last stable. ocata is the one right now.
16:21:25 <asettle> Well, Pike will be the last stable soon. Do you have urgent requirements?
16:21:27 <spotz> vakuznet: Understood but pike is about to be stable
16:21:39 <odyssey4me> vakuznet you realise that by backporting that you are destabilising ocata, completely defeating the purpose of using the stable branch
16:22:03 <andymccr> so it'd be ok if it was a security/bug fix but im not too sure what bug its fixing.
16:22:24 <andymccr> my understanding is you can do all those things already (maybe slightly less conveniently) but once you upgrade to pike it'll be that way so from then on its all good
16:22:47 <jmccrory> this did have an upgrade impact for pike (https://review.openstack.org/#/c/474787/), think it's better not to backport. ocata is under phase 2 right now already
16:23:00 <odyssey4me> see the 'For Newton, and Ocata' heading in https://evrard.me/group-and-host-variables-overriding-in-openstack-ansible.html for how you can do things in ocata
16:23:00 <andymccr> jmccrory: yeah agreed
16:23:08 <logan-> yeah and as evrardjp (nice blog post btw) points out there's ways to do scoped vars on ocata already too
16:23:13 <andymccr> jmccrory: yeah agreed
16:23:16 <odyssey4me> I'd suggest that everyone involved add their vote to the review
16:23:48 <evrardjp> sounds a legit approach
16:23:50 <spotz> odyssey4me: Sounds like a good plan, though it sounds like the cores might be in agreement
16:24:02 <evrardjp> thanks for all the kind words on the blog post :p
16:25:18 <spotz> Anyone else have anything?
16:26:14 <spotz> Ok then back to the bug smash!
16:26:17 <spotz> #endmeeting