20:00:52 <david-lyle> #startmeeting Horizon
20:00:53 <openstack> Meeting started Wed Nov 18 20:00:52 2015 UTC and is due to finish in 60 minutes.  The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:54 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:55 <hurgleburgler> (◠‿◠✿)ノ
20:00:56 <openstack> The meeting name has been set to 'horizon'
20:01:03 <TravT> o/
20:02:11 <mrunge> \o/
20:02:28 <krotscheck> \/
20:02:30 <esp_> o/
20:03:29 <david-lyle> Hi all
20:03:38 <doug-fish> hi
20:03:44 <peristeri> o/
20:04:04 <ying_zuo> o/
20:04:07 <david-lyle> The only general items I have today is a reminder that M-1 is Dec 3
20:04:41 <r1chardj0n3s> (also, lurker here, I'd like to push a little on getting midcycle nailed down ;-)
20:04:57 <matt-borland> r1chardj0n3s +1
20:05:15 <david-lyle> talked to possible hosts, need to pin down dates
20:05:34 <mrunge> yeah, good to know it in advance
20:06:06 <hurgleburgler> where are the possibilities?
20:06:28 <david-lyle> mostly west coast so far
20:06:53 <david-lyle> Texas and Colorado are also options, but twice to texas in 6 months is way too much
20:07:07 <hurgleburgler> david-lyle +1
20:07:15 <david-lyle> once is a bit much
20:07:27 * krotscheck likes seattle :)
20:07:36 <hurgleburgler> you are all welcome to come to Tucson
20:07:49 <hurgleburgler> It'll be warmer here :-D
20:07:54 <krotscheck> There could be snow on the volcano :)
20:08:04 <david-lyle> hurgleburgler's living room it is
20:08:16 <hurgleburgler> \o/
20:08:32 <david-lyle> Let me negotiate dates
20:08:39 <david-lyle> and narrow it down
20:09:21 <david-lyle> Seattle, Portland and San Jose are three easy locations on the west coast with offers to host
20:09:35 <hurgleburgler> San Diego will be warmer :-D
20:09:44 <hurgleburgler> and they have super impressive amounts of craft beer
20:09:46 <TravT> i'll email san diego
20:09:50 <TravT> anybody have their address?
20:09:59 <robcresswell> beer?
20:10:34 <krotscheck> I think at this point, impressive amounts of craft beer is not a distinguishing feature anymore.
20:10:43 * krotscheck points at both portland and seattle.
20:10:49 <TravT> maybe we could just reach out to these folks
20:10:50 <TravT> http://www.meetup.com/OpenStack-SD/
20:10:57 <hurgleburgler> san diego won't have snow though
20:11:16 <krotscheck> Snow, to me, is a bennefit.
20:11:46 <david-lyle> ski holiday in sunny colorado?
20:11:49 <krotscheck> Horizon-team ski outing :)
20:11:59 <mrunge> 2 hops to fly, about 22 hours in total :-/
20:12:10 <doug-fish> I can probably host in MN if we really need snow ...
20:12:20 <hurgleburgler> *shivers*
20:12:29 <TravT> if we want snow, then i vote Colorado
20:12:43 <hurgleburgler> snow: -1
20:12:46 <mrunge> we could ask tsufiev, if you want snow and cold temperatures
20:12:57 <krotscheck> Sure, let's do russia!
20:13:26 <krotscheck> On an unrelated note, I'm slowly trying to upgrade the eslint rules to the newer version of eslint-config-openstack, and wouldn't mind some help.
20:13:30 <r1chardj0n3s> I would also like to add a preference for date: week of 18th Jan please :-)
20:14:11 <david-lyle> let's table the midcycle for a bit
20:14:34 <david-lyle> and get a brief status on https://etherpad.openstack.org/p/mitaka-horizon-priorities
20:15:12 <david-lyle> documentation patches are in progress
20:15:39 <jasondotstar> o/
20:15:48 <david-lyle> the extensibility service merged
20:16:58 <david-lyle> Theming is one for two
20:17:26 <david-lyle> the doc patch for theming isn't on the etherpad
20:17:29 <david-lyle> hurgleburgler: status?
20:17:57 <david-lyle> pbr is still being spec'd
20:18:17 <r1chardj0n3s> xstatic/pbr is still is indeed still in spec hell, but I think we're making progress
20:18:30 <david-lyle> and nothing is linked for "Angular performance tuning"
20:18:41 <david-lyle> anyone with work on that please link it
20:18:50 <sqchen> just crealte BP
20:18:59 <sqchen> in progress
20:19:21 <TravT> nothing is up for upgrading angular or other libs right?
20:19:25 <TravT> i thought that was discussed
20:19:26 <david-lyle> I would appreciate folks linking their patches and bps to the ehterpad
20:19:28 <sqchen> david-lyle: ok
20:19:32 <TravT> we don't want those to come in at end of cycle
20:19:55 <david-lyle> TravT: that was blocked by pbr
20:19:58 <robcresswell> TravT: Releasing things is not easy without pbr
20:20:07 <TravT> right...
20:20:12 <hurgleburgler> yeah, i wanna update bootstrap, but blocked by pbr :(
20:20:17 <mrunge> r1chardj0n3s, would it be an option to drop the release as fourth component? or to use as last comporent version.release?
20:21:03 <matt-borland> david-lyle, just FYI on the Angular side alone (Users/Images/Launch Instance), I count ~30 outstanding patches, but I don't want to clutter up the etherpad
20:21:18 <matt-borland> (does not count Flavors or other deprioritized work)
20:21:40 <david-lyle> matt-borland: I'd rather see them on there for priority items or at very lest the link to the bp
20:21:44 <matt-borland> sounds good
20:21:48 <matt-borland> will do
20:21:56 <r1chardj0n3s> mrunge: I think we can get there, with 4 digits
20:22:18 <mrunge> ok, I was just reading apevec comment on https://blueprints.launchpad.net/pbr/+spec/pbr-fourth-component
20:22:40 <mrunge> but, that is still 3 dots in the version number
20:22:59 <mrunge> so, forget that please
20:23:00 <hurgleburgler> gotta leave early today o/
20:23:33 <david-lyle> that covers the criticals and would appreciate updates to the etherpad so we can track better, since launchpad is launchpad
20:24:04 <matt-borland> I'll update the angular work
20:24:09 <david-lyle> thanks
20:24:27 <david-lyle> the agenda for today is available at
20:24:36 <david-lyle> #link https://wiki.openstack.org/wiki/Meetings/Horizon#Agenda_for_November_18
20:24:50 <david-lyle> #topic reno for release notes
20:25:18 <david-lyle> The two patches lhcheng linked are now merged
20:25:31 <david-lyle> but the idea was to raise awareness
20:26:14 <david-lyle> release notes are now included in the tree when warranted, then the compilation becomes the final release notes at the end
20:26:26 <david-lyle> the process documentation for horizon was documented in https://review.openstack.org/#/c/246124/
20:27:23 <david-lyle> the idea is if it's release note worth as outlined in the doc content there, the release note component should be merged at the same time or in a follow on patch
20:27:34 <david-lyle> so reviewers help remind/enforce this
20:28:12 <david-lyle> the process is new for mitaka, so try to remember
20:28:20 <david-lyle> questions?
20:29:05 <david-lyle> perfect
20:29:17 <david-lyle> #topic Domain scoped tokens DOA patches bpokorny, esp
20:29:21 <mrunge> should we consider missing release notes as blocker?
20:29:27 <mrunge> sorry, too slow to type
20:29:43 <david-lyle> mrunge: I think yes
20:29:52 <mrunge> ack david-lyle
20:30:00 <david-lyle> can be a dependent patch, but should be in progress
20:30:25 <david-lyle> I think writing the migration part will also inform some patch work :)
20:30:43 <r1chardj0n3s> I'd also like to plug the phantomjs/webdriver patches for review please
20:30:55 <david-lyle> ok esp_, do you want to type?
20:31:03 <esp_> sure
20:31:08 <david-lyle> ack r1chardj0n3s
20:31:17 <r1chardj0n3s> https://review.openstack.org/#/c/242926/ and https://review.openstack.org/#/c/243946/
20:31:33 <esp_> brad is absent today but we’d like to make a bid for merging these DOA patches
20:31:39 <esp_> https://review.openstack.org/#/c/141153/
20:31:40 <esp_> https://review.openstack.org/#/c/196328/
20:31:51 <bpokorny> I just got back also, just in time :)
20:32:20 <esp_> ^ allows for keystone v3 integration in horizon by support a domain scoped token in the session
20:32:33 <esp_> cool bpokorny
20:33:10 <david-lyle> so, I'm on the fence and as one of the authors that's an odd place to be
20:33:22 <david-lyle> I think the code is sane
20:34:08 <esp_> aye, I believe doug-fish and lin have reviewed pretty well and gave their blessing
20:34:12 <david-lyle> the last lingering concern that was raised at the summit and I'm coming around to appreciating better is the log noise
20:34:37 <robcresswell> It has 2 +2s? Was there too much uncertainty?
20:34:42 <esp_> yeah, talked to lin about the extra logging this morning
20:34:46 <david-lyle> the will conceivably be many failures in the horizon and keystone logs
20:35:10 <doug-fish> robcresswell: it wasn't merged because it really isn't useful without the follow on policy patch; I'd expect them to get merged together
20:35:19 <robcresswell> doug-fish: Understood
20:35:20 <david-lyle> *there
20:35:50 <mrunge> yeah, we might want to change default settings for sessions then
20:36:06 <mrunge> since this will cause a session overflow otherwise
20:36:09 <david-lyle> there's that too
20:36:20 <doug-fish> mrunge: there is a check for that in the code IIRC
20:36:26 * ducttape_ can't believe we are still talking about session overflow
20:36:42 <bpokorny> Yeah, there's a check for the session backend being used.
20:36:47 <mrunge> doug-fish, that's not enough
20:36:57 <mrunge> ducttape_, that's reality
20:37:09 <david-lyle> ducttape_: signed cookies is a much simpler and HA ready session backend
20:37:14 <mrunge> but we're still talking about utf-8
20:37:22 <david-lyle> as a default it makes sense
20:37:28 <esp_> david-lyle: if we can, I’d like to tackle that as a bug to keystone to reduce logging or at least include better info regarding when a user doesn’t have a domain associated with it
20:38:52 <ducttape_> what if we added a cache call for the catalog, and didn't store that in session?  I thought that was like 75% of the cookie
20:38:57 <esp_> mrunge and ducttape_: we know there is still on going work to combine the domain token with the project token and are prepared to refactor when it’s ready.
20:39:11 <doug-fish> mrunge: I'm not sure I fully understand your concern about the default session backend. You think it's unreasonable to say some options require other, non default options?
20:39:45 <mrunge> doug-fish, I was just suggesting to change the default anyways
20:39:56 <doug-fish> got it
20:40:03 <mrunge> since that's, what deployers will need to do
20:40:12 <mrunge> and this is *one* reason.
20:40:16 <david-lyle> esp_: I'd really like to see proper v3 support merge
20:40:26 <mrunge> yes!
20:40:39 <ducttape_> so thats the two patches above, right?
20:40:41 * david-lyle will resurrect his devstack patch for memcache
20:40:54 <david-lyle> ducttape_: that's the start
20:41:12 <david-lyle> there's a mess in horizon beyond that
20:41:38 <david-lyle> the reuse of the admin role by keystone in the sample file is pure lunacy, but we need to handle it
20:41:46 <david-lyle> *sample policy file
20:42:06 <ducttape_> so proper v3 support really means better policy file usage, right?
20:42:07 <esp_> david-lyle: agreed, it’s a first step but I think there is enough there to be useful.  bpokorny and others are already using it.
20:42:12 <david-lyle> and different tokens have to be used for different resources
20:42:33 <ducttape_> whoah, you need different tokens for different policy checks ?
20:42:45 <ducttape_> that's not optimal
20:42:52 <david-lyle> ducttape_: yes and proper use of the domain scoped token when warranted over the project scoped token
20:42:59 <david-lyle> it's idiotic, but it's v3
20:43:30 <ducttape_> so can we just place both v2 and v3 tokens to the policy engine, and have the policy file pick and choose which to use?
20:44:17 <esp_> ducttape_: I don’t think we want to run v2 and v3 concurrently
20:44:19 <david-lyle> well we'd have a project scoped token and a domain scoped token instead of v2 v3 but yeah
20:44:28 <david-lyle> that's what the second patch to d-o-a does
20:44:36 <esp_> aye
20:44:36 <ducttape_> if you hard code to pass one token vs the other for a particular check, that is going to come back and bite us
20:44:46 <ducttape_> my bad.... I mean project scoped vs domain scoped
20:45:00 <david-lyle> well only keystone wants a domain scoped token
20:45:31 <david-lyle> so the toggle is simpler, except you have to know whether you're asking at the domain or project level
20:45:34 <esp_> right, it’s only used for identity mgmt (talking to the keystone api) https://review.openstack.org/#/c/196328/
20:45:45 <esp_> ducttape_: ^
20:45:57 <david-lyle> I have days worth of rants that esp_ has heard about the brilliance here
20:46:11 <esp_> def :)
20:46:13 <david-lyle> but I'll refrain for now
20:46:39 <david-lyle> there was no earthly reason for a second token type
20:47:00 <ducttape_> that is going to be a problem for deployment
20:47:08 <david-lyle> and that's actually in discussion to correct
20:47:17 <david-lyle> 3 years later
20:47:47 <ducttape_> you are going to want to write policy files that would have both credentials and domain_credentials support at the same time
20:48:07 <david-lyle> ducttape_: maybe
20:48:21 <david-lyle> depends how much you're wanting to expose to a domain admin
20:48:29 <david-lyle> outside of keystone
20:48:36 <david-lyle> but yeah, it's a mess IMO
20:48:42 <ducttape_> lets dig back out of the rathole we are in.   esp_ and brad have a couple of patches, lets get those merged I guess
20:49:20 <david-lyle> the other options aren't pretty either and this code has been thoroughly tested
20:49:26 * ducttape_ hopes we are done with the airing of grievances
20:49:43 * david-lyle puts away the pole
20:49:46 <esp_> thx, I owe some more docs upstream as well
20:50:02 <mrunge> and release notes
20:50:14 <david-lyle> hmm
20:50:25 * david-lyle wonders about d-o-a release notes
20:50:29 <robcresswell> Does d_o_a do its release notes in Horizon?
20:50:46 <ducttape_> yep, please include docs with patches.  docs are not difficult.  yes robcresswell, most of the time
20:50:52 <david-lyle> robcresswell: I usual put important items in the release proposal
20:51:01 <robcresswell> ducttape_: I realise we have done previously, I meant with reno
20:51:05 <mrunge> didn't we want to merge d-o-a in horizon?
20:51:12 <david-lyle> d-o-a is easier to track
20:51:24 * esp_ will need to spend some time looking at reno
20:51:44 <david-lyle> I think the clients may use reno
20:51:49 <david-lyle> might need to add to d-oa-
20:52:03 <david-lyle> rearrange chars as necessary
20:52:12 <david-lyle> ok, doc it
20:52:15 <david-lyle> merge it
20:52:22 <david-lyle> #topic Open Discussion
20:52:34 <robcresswell> Yeah, we'll need to set up reno in doa then
20:52:37 * krotscheck points at eslint patches.
20:52:44 * krotscheck goes back to lurking
20:52:49 <david-lyle> hey look eslint patches
20:52:57 <krotscheck> Actually, I lied.
20:53:15 <krotscheck> There's a new rule under discussion here: https://review.openstack.org/#/c/244257/
20:53:19 <r1chardj0n3s> just back to the midcycle - snow is lovely and all but let's please not hold it in a location where travel could be affected by adverse weather
20:53:27 <robcresswell> https://review.openstack.org/#/c/244910/ << thats the top of the chain, I think, for eslint.
20:54:07 <ducttape_> r1chardj0n3s +1.   south pacific island would be lovely
20:54:09 <krotscheck> robcresswell is correct.
20:54:13 <david-lyle> krotscheck: so there was a patch to add space after ";" but then you end up with "blah = {var b = 'super'; }" which felt odd
20:54:17 * krotscheck volunteers Honolulu
20:54:44 <krotscheck> david-lyle: Yeah, I saw that too and was trying to figure out if they fix it in a future version of the rule.
20:54:53 * krotscheck promptly got distracted.
20:55:13 <krotscheck> david-lyle: Feel free to -1 that one for more investigation. I may be able to update eslint upstream to get it fixed.
20:55:21 <david-lyle> krotscheck: no worries, I was just hoping there was another rule coming along to rationalize that
20:55:29 <robcresswell> I updated this, btw: https://wiki.openstack.org/wiki/Horizon/WeeklyBugReport If anyone wants some bugs to look at :)
20:55:30 * r1chardj0n3s notes that accountants don't like approving midcycles in pacific islands ;-)
20:55:53 <r1chardj0n3s> thanks again for that robcresswell!
20:56:04 <robcresswell> david-lyle: We also need to discuss bug day
20:56:06 <david-lyle> robcresswell: ah, was missing on the agenda
20:56:14 <robcresswell> david-lyle: Yes, my fault :(
20:56:17 <david-lyle> bug report that is
20:56:35 <david-lyle> yes, so my latest look at the bug list show ~1000 bugs
20:56:44 <david-lyle> which is just wow
20:56:55 <david-lyle> some are 2+ years old
20:57:17 <david-lyle> while age is not an indication of validity, some may not exist
20:57:31 <david-lyle> and actually about 600 of those are not even prioritized
20:57:57 <david-lyle> what I would like to do is let robcresswell organize a bug day to try to clean up that list
20:58:10 <doug-fish> hooray robcresswell!
20:58:41 <robcresswell> ...trapped again
20:58:45 <david-lyle> the concept would be to take a day to have people focus on triaging/classifying bugs and cross-talk when there are questions
20:59:09 <david-lyle> just take different parts of the list and have at it
20:59:46 <robcresswell> Originally thinking next week, but apparently thanksgiving
20:59:58 <david-lyle> I've been told we've discussed this previously and not followed through
21:00:09 <david-lyle> hence robcresswell is taking the helm
21:00:22 <robcresswell> I *will* find those logs
21:00:29 <david-lyle> and I can rue about v3
21:00:50 <robcresswell> I'll send out an email on the dev mailer tomorrow
21:01:01 <david-lyle> thanks!
21:01:03 <r1chardj0n3s> yay bug day
21:01:20 <david-lyle> ok time's up. Thanks everyone, have a lovely week.
21:01:23 <david-lyle> #endmeeting