12:00:34 <david-lyle> #startmeeting Horizon
12:00:35 <openstack> Meeting started Wed Sep 16 12:00:34 2015 UTC and is due to finish in 60 minutes.  The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:00:37 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
12:00:39 <openstack> The meeting name has been set to 'horizon'
12:01:09 <david-lyle> anyone around to talk horizon
12:01:11 <david-lyle> ?
12:01:25 <doug-fish> good morning
12:01:29 <tsufiev> o/
12:01:31 <mrunge> o/ good morning
12:02:43 <david-lyle> Let's talk Liberty RC-1
12:02:51 <david-lyle> #topic RC-1
12:03:21 <david-lyle> Talking to rel-mgmt, it seems we should have the FFEs landed by the end of the week
12:03:36 <david-lyle> and try to have an RC by the end of next week
12:04:06 <david-lyle> so that means code for the FFEs need to land quickly and the focus should move to bugs
12:04:43 <david-lyle> I need to start paring down the FFE list and bug list to be more tracking of what needs to be in vs what would be nice
12:05:29 <david-lyle> if fact, I think it would be nice to look at some of the bugs here to make a more informed decision on removing from RC-1
12:05:35 <robcresswell> I've been having a go at that. I think its closer to 25, down from 40+, bug wise. Curvature is ready, shelving is in, and I'm still going over databases bp.
12:06:00 <david-lyle> #link https://launchpad.net/horizon/+milestone/liberty-rc1
12:06:53 <david-lyle> robcresswell: ok, thanks
12:07:28 <david-lyle> for FFEs I think images won't make it, the other reasonably should
12:07:53 <david-lyle> now looking at bugs
12:07:57 <r1chardj0n3s> o/
12:10:13 <robcresswell> There's a few angular related ones which could be bumped from rc1 for sure.
12:10:40 <david-lyle> robcresswell: ok, sorry reading XSS one
12:11:02 <david-lyle> will finish comments later
12:11:37 <mrunge> david-lyle, masco tried to reproduce the first one xss, but failed
12:11:49 <masco_> yes
12:12:09 <david-lyle> yeah that needs further investigation
12:12:18 <masco_> i tried with the options what the reporter said
12:12:23 <david-lyle> if it's there, is been there forever
12:13:33 <masco_> he is hesitating to share the file to reproduce the issue :(
12:13:57 <david-lyle> ok, I'll leave it in RC-1 for now, but not sure I will block on it
12:14:15 <mrunge> nevertheless, while looking at container code, there's much room for improvement
12:14:40 <tsufiev> (Like throwing it away completely)
12:14:43 <tsufiev> :)
12:14:47 <david-lyle> mrunge: no doubt
12:14:53 <mrunge> hey, it's not sahara code
12:15:08 <mrunge> I mean sahara dashboard
12:15:09 <david-lyle> I've suggested containers as a work area to many people, no bites
12:16:37 <mrunge> about: https://bugs.launchpad.net/bugs/1494171
12:16:37 <openstack> Launchpad bug 1494171 in OpenStack Dashboard (Horizon) "Javascript error: Module 'horizon.auth' is not available!" [High,Confirmed]
12:16:44 <mrunge> I'fe had another report about this
12:16:56 <mrunge> kfox111 was able to reproduce it
12:16:59 <robcresswell> Yeah, from kfox right? He was asking about it
12:17:02 <david-lyle> since the second XSS is an indirect vector, I'm moving to next
12:17:08 <robcresswell> It was in another package I thought mrunge
12:17:59 <mrunge> robcresswell, we now have 2 packages reporting the same issue with missing horizon.auth
12:18:08 <mrunge> one debian, one is mine
12:18:29 <ducttape_> I have seen the issue too, this collect static stuff etc needs a test case
12:18:45 <ducttape_> something to automatically flag it
12:18:52 <mrunge> ducttape_, collectstatic was executed
12:19:11 <ducttape_> I have no doubt, but something to test we have all the js in the correct order and locations
12:19:14 <mrunge> and the environment was rebuilt before creating the package
12:19:27 <mrunge> but I agree
12:19:44 <mrunge> would be great to test js in order etc.
12:19:56 * robcresswell points at tsufiev
12:20:31 <ducttape_> when I saw it, I was missing some js libraries that were supposed to be autocollected etc.   this is the result of having autocollection, in my experience
12:20:36 <tsufiev> robcresswell: huh?
12:20:45 <robcresswell> tsufiev: They were asking for tests :p
12:20:57 <ducttape_> tsufiev - Rob is signing you up to test all the things ;)
12:21:09 <mrunge> awesome, thank you robcresswell
12:21:27 <tsufiev> robcresswell: the issue being discussed would have been caught by i9n tests for sure
12:21:37 <mrunge> is that due to autocollection feature?
12:21:49 <mrunge> if yes, can we revert that *feature*?
12:22:30 <ducttape_> i think it's quite core to having stuff work, as you are seeing right now
12:22:31 <tsufiev> robcresswell: I think it' s'more likely some unit-tests
12:22:54 <ducttape_> the other option is having a static html list of js includes
12:23:22 <david-lyle> the problem with disabling autodiscovery is then we have to specify all js files in plugin, or content loaded as a plugin, which is all
12:23:43 <robcresswell> With angular that list would be colossal
12:23:56 <ducttape_> its a few hundred lines, if I recall
12:24:01 <mrunge> yes, and it will be loaded, even if you're not using features
12:24:11 <robcresswell> ducttape_: For a couple of panels...
12:24:51 <robcresswell> I think one thing this does highlight is a big disconnect between developers and packagers that we need to fix.
12:24:57 <ducttape_> I think we still need to dig into this, the bug may or may not be autodiscovery
12:25:04 <robcresswell> ducttape_: agreed
12:25:12 <mrunge> right
12:25:25 <david-lyle> yes
12:25:32 <robcresswell> It would be good to get more info or docs etc on what packagers need us to be aware of
12:25:36 <robcresswell> pain points etc
12:25:45 <mrunge> robcresswell, still on my todo list
12:25:46 <ducttape_> I agree, this is also a side effect of xstatic.... which haunts me when I think about it
12:26:04 <tsufiev> robcresswell: on the second thought, I don't know whether auto discovery is enabled in i9n tests
12:26:06 <mrunge> uhm, not sure if this is an issue of xstatic
12:26:18 <tsufiev> Might be not
12:26:35 <ducttape_> right, let's get it understood first
12:26:52 <mrunge> I'm seeing myself hacking a something together about packaging
12:26:55 <r1chardj0n3s> it sounds like we've wandered into xstatic ;)
12:26:59 <robcresswell> mrunge: Sure, I know you're busy, wasn't supposed to be calling you out on it :)
12:27:00 <r1chardj0n3s> and probably shouldn't have
12:27:13 <mrunge> most probably to be done during the flight to tokyo
12:27:17 <ducttape_> r1chardj0n3s - we are wandering away.  don't go near there!
12:27:30 <david-lyle> yes, autodiscovery and xstatic are not related
12:27:37 <david-lyle> other than the word static
12:27:38 <r1chardj0n3s> I have a cleanup for xstatic in the works, just getting distracted by doa/tempest/devstack :/
12:28:10 <mrunge> r1chardj0n3s, doa could keep you busy for a longer time
12:28:22 <david-lyle> any bugs targeting RC-1 <= medium will be removed from the list
12:28:37 <david-lyle> the d-o-a issue is important because it blocks another release
12:28:40 <r1chardj0n3s> nooo, I just want my simple fix to the model in and a bunch of stuff in horizon is fixed by it and I can move on with my life please pretty please? :)
12:29:09 <david-lyle> which we have a minor FFE in around WebSSO and the login fix patch
12:29:34 <robcresswell> r1chardj0n3s: Which patch?
12:29:56 <r1chardj0n3s> robcresswell: https://review.openstack.org/#/c/222478/
12:30:10 <r1chardj0n3s> it ran smack into a devstack change that sdague checked in last week that broke a few gates :/
12:30:25 <david-lyle> and the webSSO is +A'd but got caught in the gate
12:30:59 <david-lyle> I'll push to see if we can get that reverted for now
12:31:06 <robcresswell> r1chardj0n3s: Ah, doa :p
12:31:14 <r1chardj0n3s> robcresswell: yes, hi :)
12:31:25 * robcresswell pretends he never saw it
12:31:50 <r1chardj0n3s> lol, nice try!
12:31:55 <mrunge> let's merge doa into horizon
12:31:57 <r1chardj0n3s> this meeting is recorded, you know :)
12:32:01 <mrunge> and then deprecate it
12:32:05 <robcresswell> dangit
12:32:17 <tsufiev> mrunge: seems reasonable to me as well
12:32:30 <robcresswell> Can we *please* merge curvature :)
12:32:41 <mrunge> (the first part, tsufiev)
12:33:00 <david-lyle> robcresswell: still trying to find time to code review it
12:33:35 <doug-fish> on curvature - I've missed many of the demos at-scale. Is it working well now?
12:33:46 <ducttape_> it is, seems pretty ready to go
12:33:50 <david-lyle> doug-fish: yes has been the consensus
12:33:55 <doug-fish> cool
12:34:00 <tsufiev> robcresswell: I'm afraid if we start reading curvature code there will be another couple of weeks to polish it
12:34:06 <david-lyle> there was a red-herring the past couple of days, but...
12:34:25 <robcresswell> I'm reluctant to approve it again, because I'm starting to look like I'm forcing it through. IMO its been pretty feature ready for the past 2/3 patchsets, but we keep adding little tweaks to it.
12:34:47 <ducttape_> if we don't read the code, thats a pretty scary thing too
12:35:14 <robcresswell> One would hope code review involves looking at the code...
12:35:20 <ducttape_> exactly
12:35:29 <doug-fish> you mean _all_ of the code?
12:35:33 <doug-fish> *kidding*
12:35:37 <david-lyle> ok
12:35:46 <tsufiev> ducttape_: nope, I meant some tiny refactoring a that I already did while angularizing it
12:35:59 <david-lyle> so bugs and FFEs are the priority
12:36:37 <david-lyle> timeline again, FFEs by the end of the week, or moved to M
12:36:43 <doug-fish> sorry - before we move on - does curvature have/need FFE?
12:36:47 <david-lyle> and RC hopefully next week
12:36:51 <robcresswell> doug-fish: It has FFE
12:36:51 <david-lyle> doug-fish: has
12:36:57 <doug-fish> ok thx
12:37:26 <david-lyle> #topic summit planning
12:37:50 <david-lyle> it's the time where we set up session planning, in the past we've done that on etherpad
12:37:56 <david-lyle> that's a bit messy
12:38:07 <david-lyle> but manageable
12:38:25 <david-lyle> another option is google forms
12:38:36 <david-lyle> but access is no global to that
12:38:55 <david-lyle> any preferences?
12:38:58 <r1chardj0n3s> I propose we lock ourselves in a room and don't come out until we have a solid plan for domain-scoped tokens and federation and get past the endless bickering...
12:39:09 <ducttape_> -1
12:39:14 <ducttape_> there is no hope
12:39:14 <r1chardj0n3s> of course, we'll all starve
12:39:28 <ducttape_> I'd go with etherpad
12:39:29 <robcresswell> etherpad has worked, stick with that I suppose. Bit messy, but seems to function okay.
12:39:37 <r1chardj0n3s> etherpad is fine :)
12:39:41 <mrunge> as long as we have beer, or shiraz r1chardj0n3s ...
12:39:54 <r1chardj0n3s> currently sipping whiskey ;)
12:40:08 <tsufiev> +1 for etherpad & beer
12:40:11 <mrunge> let's do an etherpad and try to keep total lines low
12:40:21 <david-lyle> domain scoped token will go in early post Liberty in d-o-a
12:40:29 <ducttape_> or a really small font :P
12:40:32 <robcresswell> Haha I still have this amazing memory of r1chardj0n3s and sqchen having this deep technical discussion, turning me to and saying "did you get all that down?" and I just looked up from my keyboard and groaned.
12:40:42 <r1chardj0n3s> :D
12:40:45 <robcresswell> Less hangover on Friday for me this time.
12:40:53 <david-lyle> ok, I will set up an etherpad and post to the mailing list
12:41:03 <robcresswell> Sounds good
12:41:10 <mrunge> thank you david-lyle
12:41:13 <david-lyle> #action david-lyle create etherpad for session planning email to dev list
12:41:20 <amotoki> +1 for etherpad. it is important to ensure we  put your name when you put an idea to etherpad
12:41:32 <mrunge> +1 amotoki
12:41:38 <mrunge> good suggestion
12:42:00 <david-lyle> #topic PTL elections
12:43:32 <david-lyle> So the nomination season is open. This is the last day to submit your self-nominations. Voting if necessary will be held the next week, IIRC
12:43:45 <tsufiev> david-lyle: do we have a choice :)?
12:43:52 <david-lyle> tsufiev: always
12:43:59 <mrunge> voting will begin tomorrow
12:44:08 <david-lyle> thanks mrunge
12:44:51 <david-lyle> tsufiev: I encourage anyone who thinks they want or can improve on the job being done, or just wanting a change to run
12:45:20 * tsufiev steps back just in case
12:46:08 <david-lyle> #topic Choosing a final replacement name for an 'AVAILABLE_REGIONS' setting
12:46:41 <david-lyle> no author but I'm looking at tsufiev :)
12:47:09 <tsufiev> So lhcheng suggested to drop AVAILABLE_ prefix because it's excessive
12:47:55 <mrunge> shouldn't REGIONS be renamed to something else?
12:47:56 <tsufiev> Meaning that the new setting name is AVAILABLE_REGIONS -> KEYSTONE_ENDPOINTS
12:48:05 <mrunge> yes!
12:48:16 <r1chardj0n3s> yes, I think I already +1'ed that :)
12:48:22 <mrunge> make sense to me,
12:48:23 <masco_> +1
12:48:27 <david-lyle> ok, are we resolved on changing, I propose and indefinite deprecation cycle to accommodate existing installations
12:48:37 <mrunge> do you have a review handy?
12:48:39 <r1chardj0n3s> seems fair
12:48:39 <doug-fish> I think that's fine for the setting -- but is that how we'd label the field in the UI?
12:48:44 <mrunge> yesm david-lyle
12:49:07 <mrunge> uhm, do we have it at all in UI?
12:49:21 <david-lyle> mrunge: yes
12:49:23 <david-lyle> two places
12:49:28 <tsufiev> mrunge: https://review.openstack.org/#/c/222129/
12:49:40 <david-lyle> one in d-o-a and one optional header in Horizon
12:49:51 <mrunge> I thought we present it, yes, but without a name on it
12:50:16 <david-lyle> ah, trying to remember user visible string
12:50:55 <david-lyle> Region
12:51:00 <david-lyle> :)
12:51:12 <david-lyle> suggestions on that change?
12:51:13 <tsufiev> doug-fish: doesn't federation bp aim to somehow hide keystone ENDPOINTS in user facing UI?
12:51:24 <david-lyle> Cloud?
12:52:08 <doug-fish> tsufiev: maybe ... federated keystones wouldn't have to be explicitly chosen, their regions would just be available
12:52:14 <mrunge> sounds cloudy, but I can not think of a better term
12:52:26 <david-lyle> I'm good with the new setting name, for the record
12:52:45 <david-lyle> it's the most accurate, but I'll look for further input
12:52:55 <tsufiev> doug-fish: so I think we should have consistent UI for both federated and non-federated setups
12:53:05 <masco_> we can get list of name and give it for voting
12:53:58 <doug-fish> tsufiev: I'm not sure that will be possible: non-federated setups expose the keystone endpoints before login; federated setups discover the keystone endpoints after login
12:54:26 <doug-fish> tsufiev: I suppose some kind of hybrid of this would be possible, but that kind of makes my head hurt.
12:54:42 <david-lyle> I think that's a discussion for a different time :)
12:54:51 <doug-fish> agreed.
12:55:01 <tsufiev> doug-fish: do we mean the same thing here by keystone endpoint?
12:55:16 <tsufiev> david-lyle: okay
12:55:29 <david-lyle> perhaps Tokyo?
12:55:37 <doug-fish> indeed
12:55:40 <david-lyle> #topic Open Discussion
12:56:21 <tsufiev> Have you seen my recent mail about cinder quotas?
12:56:51 <david-lyle> tsufiev: I read through it, basically we have a bogus policy choice
12:57:13 <david-lyle> did cinder implement limits instead?
12:57:35 <amotoki> info: this mail: http://lists.openstack.org/pipermail/openstack-dev/2015-September/074351.html
12:57:48 <david-lyle> or are we just forcing worse UX on users?
12:57:51 <tsufiev> amotoki: thanks!
12:58:19 <tsufiev> david-lyle: no insight about cinder :(
12:58:48 <robcresswell> Did nobody from Cinder reply?
12:58:56 <tsufiev> robcresswell: not yet
12:59:01 <david-lyle> tsufiev: I don't believe they did, just recalling the nova fun
13:00:04 <amotoki> at least we should use a same policy for quota-show among nova, cinder, neutron and other projects.
13:01:04 <david-lyle> I'll work on some input/feedback from cinder
13:01:10 <david-lyle> we are out of time
13:01:44 <david-lyle> Thanks everyone and please focus your reviews on the remaining FFEs and bugs for RC-1
13:01:54 <david-lyle> #endmeeting