20:01:15 <robcresswell> #startmeeting horizon
20:01:16 <openstack> Meeting started Wed Apr 12 20:01:15 2017 UTC and is due to finish in 60 minutes.  The chair is robcresswell. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:01:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:01:20 <openstack> The meeting name has been set to 'horizon'
20:01:38 <robcresswell> Anyone about for meeting?
20:01:50 <jeremy_moffitt> o/
20:01:52 <gary-smith> o/
20:01:56 <jgravel> o/
20:02:23 <choyj> \o
20:02:27 <betherly> o/
20:02:36 <gugl> o/
20:02:40 <rdopiera> o/
20:02:46 <ying_zuo> o/
20:03:10 <robcresswell> Hey everyone
20:03:14 <robcresswell> #link https://wiki.openstack.org/wiki/Meetings/Horizon
20:03:52 <robcresswell> #topic Notices
20:04:08 <robcresswell> First up, I'll be tagging Pike-1 tomorrow
20:04:34 <robcresswell> After a check on prominent bugs/relative stability etc
20:04:35 <david-lyle> o/
20:04:57 <robcresswell> I'll bump any bugs/bps that havent landed yet to Pike-2
20:05:36 <robcresswell> ... I said first up, but I think thats my only announcement :)
20:05:40 * robcresswell ponders
20:05:58 <robcresswell> Anyone else need to announce anything? Otherwise we'll move on to the agenda items
20:06:29 <robcresswell> #topic Drop nova-net support
20:06:48 <robcresswell> #link https://blueprints.launchpad.net/horizon/+spec/drop-nova-network
20:07:18 <robcresswell> So I'm raising this on behalf of amotoki because the meeting doesn't work for his tz
20:07:36 <robcresswell> There's been a discussion on the mailer, but wanted to raise it here too
20:08:10 <robcresswell> #link http://lists.openstack.org/pipermail/openstack-dev/2017-March/114674.html
20:08:12 <robcresswell> aaand
20:08:19 <robcresswell> #link http://lists.openstack.org/pipermail/openstack-dev/2017-April/115194.html
20:08:25 <robcresswell> These are also linked in the blueprint
20:08:43 <robcresswell> The general consensus has been that we're fine to drop nova-net support in Pike; I'm in support too
20:08:59 <robcresswell> We don't really have much of a choice either because we use the python clients
20:09:57 <robcresswell> Any yay/nay?
20:10:06 <gugl> it makes sense to drop it
20:10:41 <david-lyle> I would begrudgingly drop it
20:10:59 <robcresswell> \o/
20:11:02 <david-lyle> we typically support n-1 or n-2 releases
20:11:10 <robcresswell> Yeah, I'm aware
20:11:12 <david-lyle> for the other services
20:11:18 <robcresswell> and have generally tried to keep that up
20:11:29 <david-lyle> so this breaks that, and wasn't a consideration on the ML
20:11:52 <robcresswell> I'm not sure how we'd work around that
20:12:01 <david-lyle> but if you're on Ocata using nova-net, my sympathies are not too deep
20:12:15 <robcresswell> IIRC on the mailer it was deprecated in... Newton?
20:12:33 <robcresswell> and Ocata still supported it
20:13:13 <david-lyle> well for the handful of installations in Ocata using nova-net, they just won't be able to use the latest Horizon
20:13:19 <david-lyle> I think we're fine
20:13:48 <david-lyle> I hope that's a very small number
20:14:09 <robcresswell> Yeah
20:14:16 <david-lyle> drop away
20:14:33 <david-lyle> let the bugs fall where they may
20:14:49 <david-lyle> by drop do you mean gut code? or what exactly?
20:14:59 <david-lyle> how do we drop
20:14:59 <robcresswell> Gut code
20:15:10 <robcresswell> We make calls to the client that dont exist
20:15:16 <david-lyle> goto : let the bugs fall where they may
20:15:48 <david-lyle> is there a patch?
20:15:50 <robcresswell> Most of it is tied up in the security groups etc stuff that we have intermingled with Neutron
20:16:01 <robcresswell> No, amotoki said he's happy to own that though
20:16:17 <david-lyle> network wars
20:16:20 <david-lyle> ok
20:17:47 <robcresswell> #info Agreed to drop nova-net for Pike
20:18:04 <robcresswell> #topic Horizon feature deprecation policy
20:18:13 <robcresswell> This relates directly to what you said before david-lyle
20:19:08 <robcresswell> Generally I dont see any real reason not to keep up n-1/n-2 support, since its mostly just leaving code in the codebase and isnt much overhead to us
20:19:32 <robcresswell> This gets complicated when/if code is removed from the pythonclients
20:20:51 <david-lyle> I think the main issue arises from our ability to track all the deprecations in other services
20:21:29 <david-lyle> I would suggest we add a statement of some sort to our docs that other services deprecation notices also apply to horizon as we are just a mirror
20:21:45 <david-lyle> we can't support what they don't
20:22:14 <david-lyle> but we do tend to get caught by surprise, this buys us an out
20:22:47 <robcresswell> That goes against previous practice though right?
20:22:49 <david-lyle> s/ability/inability/
20:23:07 <robcresswell> Like with other things previously, we've kept the code for older clouds to be run with newer Horizons
20:23:48 <david-lyle> yes we have, but typically things haven't been removed
20:24:30 <david-lyle> if our deprecation bleeds to 3-4 cycles, so be it
20:25:17 <david-lyle> but this would give fair warning that we may follow the other services more directly, as the nova-net case is dictating
20:25:59 <robcresswell> Sure
20:26:01 <robcresswell> Makes sense to me
20:26:09 <robcresswell> Anyone else want to weigh in? :)
20:26:36 <david-lyle> ideally we could have better communication between projects and not have surprises, but we have what we have
20:26:48 * david-lyle stops typing
20:27:15 <robcresswell> haha
20:29:14 <robcresswell> So, document that we will follow service deprecations, but don't force removal of code ASAP
20:29:52 <gugl> just curious, how we watch the deprecation list of the services horizon uses during each release?
20:29:59 <robcresswell> If it remains in a deprecated state we can still hold the code, but there are no guarantees past the service's timeline
20:30:16 <robcresswell> gugl: Generally speaking we just keep things until someone points out its deprecated, or it stops working
20:30:32 <robcresswell> Things rarely stop working though, because OpenStack tends to add code, not remove it
20:30:54 <gugl> robcresswell, oh
20:31:13 <robcresswell> I mean, I think this nova-net change is the first time I can really recall it happening? Apart from neutron stadium stuff, but even a lot of that is just in a weird unmaintained state I think.
20:32:11 <david-lyle> oh yeah, I was supposed to remove some aaSs
20:35:04 <robcresswell> I think that's everything as far as I'm concerned :)
20:35:08 <robcresswell> #topic Open Discussion
20:35:20 <robcresswell> If anyone has any patches, bugs, blueprints, feel free to raise them
20:35:25 <betherly> Qos basic table panel passing tests now - would love reviews on it :) https://review.openstack.org/#/c/418828/
20:36:41 <robcresswell> \o/
20:36:43 <robcresswell> nice
20:36:48 <robcresswell> Sure, I'll take a look
20:36:58 <betherly> robcresswell: thanks so much for your help getting it there :)
20:37:34 <gary-smith> I've got a bug fix that is causing theming to break: https://review.openstack.org/#/c/453919/. Would love a core review on that
20:38:39 <robcresswell> Oh yeah, I remember that being mentioned
20:38:48 <robcresswell> Also starred
20:39:58 <robcresswell> Anything else?
20:40:04 <jeremy_moffitt> not tied to a specific bug, but is the "npm" target in tox functional? we can follow-up in openstack-horizon if the answer is yes... getting chrome errors trying to launch tests that way, but using the karma CLI it launches successfully
20:40:30 <gary-smith> On the recent open list (within last week) there are several (mostly theming) changes authored Diana Whitten and/or Matt Borland that would be great to have reviewed.
20:41:01 <robcresswell> jeremy_moffitt: Yeah, it should work fine, I've been using it without issue
20:41:06 <gary-smith> Many were abandoned due to stagnation and releases and have been freshened up again
20:41:11 <jeremy_moffitt> great thanks, will follow-up in the normal room
20:41:28 <robcresswell> gary-smith: Yeah, a lot of that fell to the bottom of the pile
20:41:53 <robcresswell> Once the theming problems were resolved, most of it seemed more of a nice-to-have, than critical issues
20:42:04 <gary-smith> robcresswell: undertstood
20:42:27 <robcresswell> I'll try and spend some review time on it, if its being maintained again
20:42:33 <robcresswell> I'm pretty familiar with it all
20:45:02 <robcresswell> If there's nothing else, I think we can end here :)
20:45:22 <ying_zuo> I have a question for tox
20:45:30 <robcresswell> ying_zuo: Fire away
20:45:52 <ying_zuo> I used tox -e npm locally and it passed but failed in gerrit.
20:46:08 <ying_zuo> maybe I should be using a different command?
20:46:25 <david-lyle> link to failed job?
20:46:29 <robcresswell> ying_zuo: Which patch?
20:47:00 <ying_zuo> it was a obvious failure but I wish I could find it out when I run it locally
20:47:49 <david-lyle> so the failure wasn't installing npm packages related? but an actual issue?
20:48:01 <ying_zuo> the test failed
20:48:07 <robcresswell> You should be able to run it identically locally
20:48:17 <ying_zuo> hmm
20:48:26 <ying_zuo> maybe I just missed that error then
20:48:44 <ying_zuo> cool. thanks for confirming
20:49:18 <robcresswell> ying_zuo: It wasn't linting or something?
20:49:31 <robcresswell> tox -e npm just runs the tests iirc...
20:49:32 <ying_zuo> no
20:49:55 <robcresswell> hmm, I'm not sure then I'm afraid. tox in that instance is literally just a wrapper for npm
20:52:22 <robcresswell> Anything else?
20:54:48 <robcresswell> Okay, thanks for your time everyone :)
20:54:52 <robcresswell> #endmeeting