12:00:41 <david-lyle> #startmeeting horizon
12:00:41 <openstack> Meeting started Wed Sep  2 12:00:41 2015 UTC and is due to finish in 60 minutes.  The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:00:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
12:00:44 <openstack> The meeting name has been set to 'horizon'
12:00:53 <r1chardj0n3s> o/
12:00:55 <fnordahl> o/
12:00:58 <tmazur> o/
12:01:03 <tsufiev> o|
12:01:11 <bradjones___> o/
12:02:12 <david-lyle> hi folks
12:02:25 <david-lyle> we are at L-3 tomorrow
12:02:47 <david-lyle> https://launchpad.net/horizon/+milestone/liberty-3 is what it's looking like so far
12:03:50 <david-lyle> that's a fair departure from https://etherpad.openstack.org/p/YVR-horizon-liberty-priorities
12:04:03 <david-lyle> #link https://launchpad.net/horizon/+milestone/liberty-3
12:04:12 <david-lyle> #link https://etherpad.openstack.org/p/YVR-horizon-liberty-priorities
12:04:37 <david-lyle> but significant progress was made on made items in the priority list
12:05:05 <david-lyle> I expect a few FFE
12:05:15 <david-lyle> but I don't know which yet
12:05:43 <david-lyle> One thing that is almost finished is Django 1.8 support
12:06:00 <david-lyle> we have support in both d-o-a and horizon and g-r now allows 1.8
12:06:27 <david-lyle> I'm going to release d-o-a one more time to pull in the requirements.txt changes
12:06:48 <david-lyle> and hopefully get a little WebSSO refresh
12:07:07 <david-lyle> that should happen in the next couple of days
12:07:31 <david-lyle> I would also like to advertise trove and sahara reviews
12:07:57 <david-lyle> #link https://etherpad.openstack.org/p/sahara-reviews-in-horizon
12:08:07 <david-lyle> I've spent a fair amount of time merging several of those
12:08:11 <david-lyle> and
12:08:31 <david-lyle> https://etherpad.openstack.org/p/trove-reviews-in-horizon
12:08:39 <david-lyle> #link https://etherpad.openstack.org/p/trove-reviews-in-horizon
12:09:17 <david-lyle> I start on the first trove patch but it's 1700 lines
12:09:31 <david-lyle> but has been around for over a year now and still not merged
12:09:36 <r1chardj0n3s> 1700 lines is never going to be reviewed :/
12:09:45 <tsufiev> close eyes and merge ))
12:09:49 <r1chardj0n3s> lol
12:09:51 <masco> :)
12:09:55 <david-lyle> r1chardj0n3s: I'm still working through it
12:10:22 <david-lyle> honestly I'd rather see 1700 lines and a complete feature than 200 lines of a stub
12:10:27 <david-lyle> but maybe that's  just me
12:10:44 <david-lyle> but 1700 lines is difficult to review for sure
12:11:02 <r1chardj0n3s> scientists (well, this one guy) have proven that patches over 400 lines don't get reviewed
12:11:09 <r1chardj0n3s> (I saw it on twitter, it must be true)
12:11:13 <bradjones___> speaking of patches that have been around for a while it would be great to get the curvature work in liberty
12:11:18 <bradjones___> #link https://review.openstack.org/#/c/141078/
12:11:25 <david-lyle> and doc patches get over reviewed
12:11:47 <tsufiev> david-lyle, should Horizon cores just ensure that the contrib/ stuff doesn't break Horizon and that's all? (with the rest being checked by service team)
12:13:16 <fnordahl> the fact that doc patches get over reviewed probably has to do with the bikeshed-theory (http://bikeshed.com) :-)
12:13:20 <david-lyle> tsufiev: I like to do a little more diligence than that, but some things I can't test really well
12:14:00 <david-lyle> bradjones___: I've been watching that, had a couple -1s for a while
12:14:17 <david-lyle> I think that may be worthy of a FFE, should it come to it
12:14:31 <bradjones___> david-lyle: have been moving house past week but addressed those issues this morning so I hope it is good to go now
12:14:49 <david-lyle> bradjones___: no worries
12:14:56 <david-lyle> life happens
12:15:01 <david-lyle> :)
12:15:05 <bradjones___> david-lyle: :)
12:15:32 <david-lyle> any thing else people have on their minds that would warrant a FFE?
12:15:47 <david-lyle> (Feature Freeze Exception)
12:16:04 <david-lyle> a bp merging after L-3
12:17:19 <david-lyle> Other release item: translations
12:17:37 <david-lyle> unlike other services, our string freeze will be RC-1
12:17:56 <david-lyle> and then merge the final strings in RC-2
12:18:12 <david-lyle> as we do very little without string changes
12:18:54 <david-lyle> So as a reminder, once we tag L-3, master will be closed to features not granted a FFE
12:19:27 <david-lyle> I will most likely go -2 a ton of patches as usual because the tools are inadequate to control this
12:19:37 <david-lyle> bug fixes are still ok
12:19:51 <david-lyle> for RC-2 the criteria for merging a bug fix will be higher
12:20:21 <tsufiev> david-lyle, this https://review.openstack.org/#/c/215196/ is a bugfix, but most likely it will include a minor UI change
12:20:36 <tsufiev> piecharts for disabled quotas
12:20:54 <david-lyle> once RC-1 is tagged, a release branch will be made and master will be open for Mitaka work again.
12:21:09 <tsufiev> would love to land it in Liberty because as of now it's one of our blockers for Launch Instance NG
12:22:12 <david-lyle> tsufiev: I've added it to my list, will add to RC-1 as well
12:22:29 <tsufiev> david-lyle, thanks
12:22:42 <david-lyle> ok, I think that covers release items
12:22:51 <david-lyle> any questions/concerns?
12:24:33 <mrunge> what about enabling new launch instance as default?
12:24:38 <mrunge> if it's not already?
12:24:49 <mrunge> I haven't checked it lately
12:24:54 * ducttape_ can of worms
12:25:05 <mrunge> :D
12:25:13 <tsufiev> mrunge, btw, the above bug is the exact reason we didn't so in our distro :)
12:25:25 <mrunge> tsufiev, we don't either
12:25:52 <ducttape_> I think we'd want more time to flush out bugs and ux issues before it is the default
12:26:26 <tsufiev> ducttape_, is there some kind of actual bugs and ux issues for Launch Instance NG?
12:26:27 <mrunge> the issue is, we're sending quite unclear signals to projects basing on horizon
12:26:34 <tsufiev> *some kind of list
12:27:10 <mrunge> tsufiev, I'd say, that depends on who you're asking ;-)
12:27:14 <ducttape_> tsufiev - coworkers that have complained to me, I asked for bugs.  I *think* there are bugs for items
12:27:42 <mrunge> I don't think launchpad supports something like tracker bugs
12:27:57 <ducttape_> mrunge - unclear signals just b/c of launch instance?
12:28:15 <mrunge> ducttape_, uhm, unclear about the direction
12:28:15 <tsufiev> mrunge, well, we could use tags for maintaining this list
12:28:39 <mrunge> I have been asked again and again, what happens with plugins based on python code
12:28:58 <ducttape_> mrunge - I think that is unfortunate, but probably natural.  we are still trying to find the best way to add ng based features
12:29:07 <mrunge> it's somehow clear we're deprecating that
12:29:27 <ducttape_> python is a known thing, and I think it will be around for M and maybe N releases even
12:29:44 <ducttape_> (my $.02)
12:30:05 <mrunge> ducttape_, yes; that's my understanding, and I agree
12:30:07 <david-lyle> ok my connection failed silently I believe
12:30:11 <ducttape_> we are in the middle right now, and have a mix.  it's weird
12:30:41 <tsufiev> ducttape_, like the whole thing with python2 and python3?
12:30:44 <mrunge> david-lyle, when did your connection drop?
12:30:53 <ducttape_> exactly tsufiev ;)
12:31:02 <david-lyle> I think about 12:24 UTC
12:31:18 <mrunge> david-lyle, we were unclear about new launch instance becoming default
12:31:19 <ducttape_> talking about python vs ng - what do plugins do
12:31:40 <mrunge> yes, exactly that
12:31:51 <david-lyle> I don't think the new launch instance is ready
12:32:12 <ducttape_> right, I think most people in the room agree then
12:32:13 * tsufiev feels guilty for not paying enough attention to reviewing angular stuff in horizon
12:32:25 <mrunge> david-lyle, I believe, we agree here
12:32:39 <mrunge> I would expect a different picture at 20:00 UTC
12:32:44 <ducttape_> lol
12:32:46 <david-lyle> there was a fix to static_url that made some progress, but otherwise it was largely neglected after the late Kilo push
12:32:51 <tsufiev> david-lyle, it would be great to define some criteria of 'readiness' for it
12:33:20 <david-lyle> I probably haven't recorded that tsufiev
12:33:44 <mrunge> and I think we need to talk about deprecation of plugins in python
12:33:47 <david-lyle> There are is a couple of areas where we lack parity with the existing interface
12:34:05 <david-lyle> still
12:34:13 <david-lyle> the first page needs to be split
12:34:36 <david-lyle> the transfer tables could use a redesign (wish list)
12:34:49 <david-lyle> and support for add network should be added
12:35:23 <ducttape_> I think we want to look at tables again, to see if what we have is ideal.  before we go follow that pattern 20 times
12:35:28 <david-lyle> while add network is above existing functionality that's one of the main reasons we started the rewrite road
12:35:48 <david-lyle> alleviating dead ends in the wizard
12:35:55 <david-lyle> we still have all the same dead ends
12:36:27 <david-lyle> I think part of what's missing is, we aren't rewriting to prove it can be done, of that I have no doubt
12:36:37 <david-lyle> we are rewriting things to improve on them
12:37:18 <david-lyle> as far as plugins, I think angular only as a plugin path is going to be a tough sell
12:37:38 <david-lyle> unless we have a great story around how easy it is to build them
12:37:46 <david-lyle> we are not there yet
12:38:12 <david-lyle> I probably echoed other people because I dropped
12:40:39 <david-lyle> ducttape_: I think the tables are getting closer, the transfer tables are still a work in progress
12:40:45 <mrunge> would that be a topic for next summit?
12:40:47 <david-lyle> and overused I think
12:40:59 <david-lyle> mrunge: which?
12:41:10 <ducttape_> how to move a table to ng
12:41:17 <ducttape_> that would be like 80% of the work
12:41:20 <mrunge> david-lyle, readiness of new launch instance? etc?
12:41:34 <david-lyle> mrunge: definitely
12:41:54 * ducttape_ has to drop, thanks all
12:41:56 <david-lyle> ducttape_: I think we'll have a story by then
12:42:02 <david-lyle> thanks ducttape_
12:42:02 <ducttape_> good :D
12:42:30 <mrunge> when putting this up, I've been asked several times, how to deal with plugins based on python
12:42:44 <mrunge> we need to have something like a deprecation timeline
12:42:58 <tsufiev> mrunge, +1
12:43:01 <mrunge> or how we'll act on that, even if we don't know exact dates
12:43:03 <zhenguo> agree, many plugins use python now
12:43:17 <tsufiev> other PTL need to plan in advance
12:43:33 <david-lyle> mrunge: do we need to deprecate python plugins? I think we're a long way off from that being a need
12:43:42 <tsufiev> so they need to allocate some time for rewriting horizon-based plugins
12:44:13 <david-lyle> and if I tell Sahara to go rewrite all their stuff in angular, I think they may tell me to go stuff myself
12:44:22 <mrunge> david-lyle, it's more the message, we need to deliver to dependent projects
12:44:52 <mrunge> and to give them some timeline
12:44:53 <tsufiev> david-lyle, I suspect that if python in horizon is going to be deprecated in some _indefinite_ time, Horizon users will postpone learning NG for the same _indefinite_ time :)
12:45:21 <tsufiev> or, sorry, not Python, I meant Django views
12:46:18 <david-lyle> my problem is, I can't in good faith tell someone to start an angular based plugin now
12:46:31 <david-lyle> there are a couple who have started
12:47:28 <mrunge> I have the feeling, if we had a message, krotscheck wouldn't started something completely new
12:47:48 <david-lyle> mrunge: no the use case is different
12:47:56 <david-lyle> ironic wants standalone
12:48:13 <david-lyle> you can run ironic without openstack
12:48:19 <david-lyle> like swift
12:48:38 <zhenguo> yeah, I propose an ironic-dashboard, but finally droped :(
12:48:56 <david-lyle> so a pure horizon plugin doesn't meet their needs
12:49:16 <mrunge> true
12:50:22 <zhenguo> for ironic, should we support that in horizon, like disable some unsupported ironic hypervisor actions?
12:50:39 <mrunge> zhenguo, not sure.
12:50:49 <zhenguo> by default, ironic devstack not enable horizon now
12:51:03 <mrunge> this was discussed more than once, zhenguo
12:51:20 <mrunge> it's likely the same as with tuskar-ui
12:51:44 <mrunge> you'd be configuring underlying hardware
12:51:56 <mrunge> this is not horizons focus
12:52:08 <david-lyle> I think the aim was for the ironic webclient work to have reusable bits that might be leveraged in Horizon, but I'd like to know the use case where the ironic driver in nova doesn't meet the needs of users
12:52:18 <zhenguo> mrunge: ok got it
12:52:39 <mrunge> that being said, I still believe we have some experience or knowledge around that
12:53:02 <mrunge> and it will probably fit better in horizon than ... zaquar?
12:53:05 <zhenguo> david-lyle: please have a look at https://review.openstack.org/#/c/215419/
12:53:25 <mrunge> not trying to fingerpoint to a specific project
12:53:41 <david-lyle> zhenguo: that makes sense to me
12:54:09 <david-lyle> but  again that's an interaction with nova
12:54:37 <zhenguo> david-lyle: yeah
12:54:47 <david-lyle> added to my list
12:55:14 <zhenguo> does it make sense to use another policy file to disable some unsupported ironic actions?
12:55:32 <david-lyle> zhenguo: I don't think policy is the right mechanism
12:55:49 <david-lyle> as you need an API call result to determine support
12:56:03 <david-lyle> I think logic in the code is the best answer
12:56:15 <zhenguo> ok, agree
12:56:54 <david-lyle> 4 minutes left, any other topics
12:58:36 <tsufiev> integration tests again broken, now because of some change in cinder (most likely)
12:58:54 * tsufiev starting to think about testing horizon against less fluid openstack
12:59:04 <r1chardj0n3s> ugh, yeah
12:59:28 <david-lyle> what changed in cinder? the client?
12:59:53 <tsufiev> david-lyle, some issue related to getting the quotas, https://review.openstack.org/#/c/219620/2 fixes that
13:00:02 <fnordahl> well, actually. I beleive the fault is our error checking
13:00:06 <fnordahl> I am working on it right now.
13:00:18 <fnordahl> coming with a new patchset with a more specific except line
13:00:29 <tsufiev> fnordahl, but it was working before, which implies that something changed in cinder as well
13:01:27 <david-lyle> ok, time's up, but the fragile nature of the integration tests is something we need to discuss at the summit
13:01:51 <david-lyle> I know tsufiev has already been working on it, but we may need something more drastic
13:01:58 <mrunge> can we get horizon integration tests running in other projects?
13:02:08 <david-lyle> Thanks everyone!
13:02:12 <david-lyle> #endmeeting