17:05:07 <hogepodge> #startmeeting refstack
17:05:09 <openstack> Meeting started Tue Oct 10 17:05:07 2017 UTC and is due to finish in 60 minutes.  The chair is hogepodge. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:05:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:05:12 <openstack> The meeting name has been set to 'refstack'
17:06:34 <mguiney> o/
17:06:37 <tosky> o/
17:07:27 <hogepodge> #topic agenda
17:07:31 <hogepodge> #link https://etherpad.openstack.org/p/refstack-meeting-17-10-10 agenda
17:07:40 <pvaneck> o/
17:09:01 <hogepodge> Please update the agenda, copy and pasted from last week.
17:09:47 <hogepodge> #topic tempest autoconfig
17:10:15 <hogepodge> Still waiting on this to merge
17:10:23 <hogepodge> #link https://review.openstack.org/#/c/508502/ tempest autoconfig project
17:10:42 <hogepodge> We're put the autoconfig project from Red Hat and RDO under RefStack governance.
17:11:18 <hogepodge> haha I just realized that instead of putting +1, I put +!
17:11:23 <tosky> thanks!
17:11:44 <hogepodge> definitely excited about it :-)
17:11:52 <hogepodge> any other comments on that work?
17:13:38 <mguiney> super excited about it, tbh
17:14:13 <hogepodge> #topic Subunit Upload
17:14:45 <mguiney> hello all!
17:14:47 <hogepodge> hoping to get this patch into shape soon for merge https://review.openstack.org/#/c/498735/
17:14:59 <hogepodge> Looks like there was a regression?
17:15:13 <mguiney> yeah, that seems to have been broken by the change we tried yesterday
17:15:39 <mguiney> i have a syntax thing to try to fix it, planning on having that up by end of meeting, its very minor
17:15:58 <hogepodge> hmm, ok
17:16:14 <mguiney> iiiinteresting, it looks like that may not be the case at all
17:16:27 <mguiney> ok, i will fix
17:16:49 <mguiney> sorry this has taken so long to get ironed out, please check back later today and review
17:17:18 <hogepodge> Looks like some syntax changes. Let's get the code in shape to go and if we're still having coverage failure we can revisit it in a later patch
17:17:59 <hogepodge> the api spec
17:18:02 <hogepodge> #link https://review.openstack.org/#/c/506826/ spec
17:18:19 <hogepodge> luzC mad some great comments on it
17:18:26 <mguiney> its trying to grab the alembic config from a function that no longer exists where it thinks it does
17:18:31 <mguiney> on it
17:18:56 <hogepodge> pvaneck if you can take a look that would be great too
17:18:57 <mguiney> anyways, yes! i really appreciated those
17:19:41 <pvaneck> sure, i'll review it
17:20:01 <mguiney> i think that they really helped me to straighten out some unclear wordings, When I originally wrote this, I spent so much time dug into what I was working on that I think that things that make sense to me, in the context of the wordings, may not be entirely clear to anyone else
17:20:24 <mguiney> basically *i* know what i mean, independent of actual meaning
17:21:41 <hogepodge> mguiney: I'm glad we got some more eyes on it, it's a good spec
17:21:51 <hogepodge> any other items on the subunit work?
17:22:32 <mguiney> nope, gonna delve back into subunit2sql tests, as soon as i finish up the last of my first pass at 2018.01 scorings
17:23:00 <mguiney> that got shoved onto the backburner for a minute, just because other things are shorter term and i know how long specs take to merge :)
17:23:09 <mguiney> other than that, that is it
17:23:54 <hogepodge> #topic result verification
17:24:03 <hogepodge> main script here
17:24:10 <hogepodge> #link https://review.openstack.org/#/c/499956/ update script
17:24:38 <hogepodge> we still need to disable anonymous uploads so we don't have to keep running the script, but I haven't done any work on config and testing
17:25:14 <hogepodge> #link https://review.openstack.org/#/c/507695/ disable anonymous upload
17:25:29 <hogepodge> I need to write tests for that, and also add a configuration option to not break other users
17:25:51 <hogepodge> mguiney: https://review.openstack.org/#/c/504358/ should be abandoned?
17:27:15 <hogepodge> ah, nevermind, it's already abandoned. My bad
17:27:17 <mguiney> ah yes, just merged it into the main patch
17:27:27 <mguiney> the content of it, i mean
17:28:16 <hogepodge> mguiney: are you still working on https://review.openstack.org/#/c/499956/ ?
17:28:23 <hogepodge> will you let us know when it's ready for review?
17:28:34 <mguiney> it is!
17:28:38 <hogepodge> ok, great!
17:28:47 <mguiney> i just realized that i hadnt removed the misplaced tox.ini mods
17:29:24 <mguiney> so since it was a 10 second fix, i reset it to master and pushed it. should pass tests, given that it was just a reset to master state
17:29:39 <hogepodge> ok, I'll get a review in once jenkins/zuul has posted
17:29:48 <hogepodge> anything else on this work?
17:29:50 <mguiney> thank you!
17:29:53 <mguiney> kind of
17:30:26 <luzC> o/
17:30:26 <mguiney> I have a patch (i know i have a lot up, and we dont have a ton of review bandwith right now, but this is actually pretty handy, for the refstack db update process)
17:30:49 <mguiney> (hello luz! thank you for the excellent reviews!)
17:31:18 <mguiney> #link https://review.openstack.org/#/c/503495/
17:32:05 <mguiney> this patch, which was a feature originally suggested by CatherineD at the boston summit, was something that I didnt have time to get into the original script, but that makes life a lot easier in using it
17:33:21 <mguiney> it basically just builds in the option to generate a new token if you dont already have a currently valid one, along with letting you know if updated failed because of an invalid or nonexistent or expired token
17:33:56 <luzC> mguiney: any time ;)
17:34:07 <hogepodge> this builds on the previous patch?
17:34:10 <hogepodge> hi luzC !
17:34:22 <mguiney> to be honest, i would love to add a more generalized version of this somewhere into refstack-client, because it does the token auth testing for you, but given how much is going on right now, it
17:34:35 <mguiney> seems like a good idea to drop it, for the time being
17:34:47 <mguiney> hogepodge: yes, this is built into the update script
17:34:49 <hogepodge> mguiney: can you add it as a topic to the agenda so it will carry forward?
17:35:01 <mguiney> oh absolutely, apologies
17:35:14 <hogepodge> no need to apologize, I think it's a good idea and I want to capture it
17:35:31 <mguiney> should it be under the refstack db update, given that its actually not a critical thing to running the script on prod?
17:35:49 <hogepodge> yeah, we can stage the merges
17:35:59 <mguiney> (more of a thing to use for users who  arent as familiar with the landscape)
17:36:03 <mguiney> cool, will add
17:36:14 <mguiney> i'd just forgotten to bring it up, last meeting
17:36:48 <mguiney> ah its already there. is that new?
17:37:06 <hogepodge> I don't think so
17:37:07 <mguiney> probably not. anyways, thank you!
17:37:43 <hogepodge> Ok, that's a good lead into the next topic
17:37:55 <hogepodge> #topic replace ostestr with tempest run
17:38:16 <hogepodge> Tempest moved off of ostestr and it's been causing some problems for users.
17:38:28 <hogepodge> These patches merging will require bump of tempest version
17:38:42 <hogepodge> #link https://review.openstack.org/#/c/503209/ ostestr to tempest run
17:39:06 <hogepodge> #link https://review.openstack.org/#/c/509934/ parallel test running
17:39:12 <hogepodge> the first being more critical than the second
17:39:38 <hogepodge> luzC: I don't understand why dynamic credentials won't work, but I'm probably missing something
17:39:51 <hogepodge> isn't refstack client just a front end for tempest, so any valid config would work?
17:41:51 <luzC> for the credentials in particular... it support only pre-provisioned which comes with using the accounts.yaml
17:42:17 <hogepodge> luzC: why?
17:42:19 <luzC> I think the reason is that we want tempest to be run with a non-admin user
17:42:46 <hogepodge> it's a policy choice and not a technical barrier though?
17:42:59 <luzC> correct
17:43:10 <hogepodge> ok
17:43:55 <luzC> when you use dynamic credetials in tempest it requires admin credentials because it create the resources on the fly (including additional users/ tenants/ etc)
17:44:02 <hogepodge> I can remove that to discourage admin use. Anyone else have comments on that?
17:44:21 <hogepodge> yeah, some clouds just send their full test results up, because they're testing anyway so just use existing jobs
17:45:04 <hogepodge> we don't reject results with admin credentials, we just don't require admin
17:46:12 <hogepodge> luzC: I can make the change to only talk about additional creds in accounts.yaml, to be consistent with the rest of our policy
17:46:34 <luzC> yes, that would be good
17:46:54 <hogepodge> pvaneck: I made the changes you suggested to the first path. thanks for catching that
17:47:15 <luzC> as you mentioned is just to discourage admin use
17:47:20 <hogepodge> any other comments on these patches?
17:47:47 <pvaneck> hogepodge: good, probably good for merge then
17:48:45 <hogepodge> pvaneck: cool. if luzC or catherine can take a look we can get it in.
17:48:53 <mguiney> hogepodge: one thought-
17:49:07 <hogepodge> yeah?
17:49:43 <mguiney> if we remove the admin credentials usage thing, would it perhaps make sense to add a refstack-client flag that would just autogenerate an accounts file, so the user doesnt even have to think about it?
17:50:17 <mguiney> there is already a script that does that, of course, but just add a flag that's like "please run this script ahead of time, i dont even want to think about user credentials"
17:50:22 <hogepodge> mguiney: one of the issues is creating a user account requires admin creds, and not all users would have them
17:50:30 <mguiney> ahhhhh fair
17:50:32 <hogepodge> there are other fixtures that need to be set up beforehand too
17:50:43 * mguiney nods
17:50:49 <hogepodge> like images, networks, and so on. So it's presumed that in public clouds these things already exist
17:51:33 <hogepodge> with the time winding down, need to move on to scheduling for the next month
17:51:43 <hogepodge> #topic upcoming meeting scheduling
17:52:26 <hogepodge> I'm not going to be available on October 17, I'll be at a workshop in Austin.
17:52:57 <hogepodge> I'll also be traveling on October 31 to the summit, and will be at the summit on the 7th
17:53:31 <hogepodge> So that leaves one meeting time left in the month, October 24. Are we ok with that? Does anyone want to run a the meeting next week or on the 31st?
17:54:36 <mguiney> i'd definitely like to try and keep reviews/progress moving forward, if possible
17:54:42 <luzC> I think we can skip next week and on the 24th assess again
17:55:10 <hogepodge> I lieu of a meeting next week, do you want to send out review reminders mguiney?
17:55:30 <mguiney> sounds good, what would be the best way to do that?
17:55:30 <hogepodge> to help keep moving things forward?
17:55:54 <mguiney> and apologies, /me hates productivity choke points
17:55:56 <hogepodge> just send an email out to the regular team members and cores with a list of reviews?
17:56:03 <mguiney> sounds good!
17:56:25 <hogepodge> I'm hoping we get a few of these reviews merged this week to start moving things forward some more
17:56:28 <ltosky[m]> or to openstack-dev@?
17:56:54 <hogepodge> ltosky[m]: sure, that works too. wider audience. use the [refstack] tag on it
17:56:59 <hogepodge> that's a good idea
17:57:21 <ltosky[m]> yep
17:57:27 <mguiney> can do
17:57:29 <hogepodge> I'll send a notice to the mailing list about the schedule for the next two weeks also
17:57:48 <mguiney> will make a note in my calendar to have that on deck and ready send
17:57:52 <luzC> sounds like a plan :)
17:58:04 <hogepodge> I'll leave the floor open for a few more minutes for discussion on any other topics.
17:58:06 <hogepodge> Thanks everyone!
17:58:41 <mguiney> good meeting!
18:00:02 <hogepodge> #endmeeting