22:01:46 <gabrielhurley> #startmeeting horizon
22:01:47 <openstack> Meeting started Tue Jul 16 22:01:46 2013 UTC.  The chair is gabrielhurley. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:01:48 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:01:51 <openstack> The meeting name has been set to 'horizon'
22:01:56 <gabrielhurley> Hello folks!
22:02:00 <jcoufal> hey
22:02:05 <gabrielhurley> #topic overview
22:02:23 <david-lyle> Hello
22:02:28 <absubram_> hi
22:02:32 <timductive> o/
22:02:34 <jpich> hey
22:02:40 <gabrielhurley> So, we're trying to cut H2 in the next 24 hours.
22:02:55 <gabrielhurley> I'm watching https://review.openstack.org/#/c/36351/ and https://review.openstack.org/#/c/35664/ and https://review.openstack.org/#/c/37335/
22:03:01 <vkmc> Hey o/
22:03:05 <gabrielhurley> those are the final three reviews that I'd really like to see in H2
22:03:19 <gabrielhurley> The keystone auth one is the only one that looks in danger :-/
22:03:33 <gabrielhurley> still trying to nail down client version problems apparently
22:03:45 <gabrielhurley> it looked promising for a bit, but just failed jenkins again
22:03:54 <gabrielhurley> lcheng is looking at it
22:04:09 <gabrielhurley> If that can't be fixed in the next few hours we'll have to defer it to H3
22:04:34 <gabrielhurley> Overall, good job on H2. https://launchpad.net/horizon/+milestone/havana-2
22:04:48 <gabrielhurley> 8 pretty sizable blueprints and 50+ bugs
22:05:10 <gabrielhurley> The big one that got bumped was the ceilometer stuff
22:05:32 <gabrielhurley> but with Heat generally in great shape now, I may be able to spend more time with the ceilometer stuff in H3
22:06:02 <gabrielhurley> I'd mostly like to spend this meeting looking forward at H3, and then at the end we can do other general questions and the like
22:06:09 <david-lyle> Looks like 35664 just merged
22:06:14 <gabrielhurley> Anybody have any concerns on H2 before we move forward?
22:06:40 <gabrielhurley> lol, of course that made https://review.openstack.org/#/c/36351/ fail due to a merge conflict
22:06:48 <gabrielhurley> timductive: https://review.openstack.org/#/c/36351/ needs a rebase
22:06:52 <gabrielhurley> then we'll push again
22:06:56 <timductive> ok
22:07:01 * gabrielhurley just loves last minute merge rushes
22:07:26 <gabrielhurley> okay then
22:07:31 <gabrielhurley> #topic blueprints
22:07:48 <gabrielhurley> let's look at H3: https://launchpad.net/horizon/+milestone/havana-3
22:07:56 <gabrielhurley> basically, we've got *way* too many blueprints
22:08:14 <gabrielhurley> some of them are smaller ones that'll only take a day or two to knock out, but even still
22:08:53 <gabrielhurley> Lemme ask this first: does anyone own a blueprint in H3 they don't think they'll finish
22:08:53 <gabrielhurley> ?
22:09:10 <gabrielhurley> no worries if you don't think you will, I just want to try and get a baseline
22:10:16 <gabrielhurley> I'll take that silence as a "no" from the people here
22:10:34 <gabrielhurley> let's tackle large topic areas
22:10:56 <gabrielhurley> I consider ceilometer support a must, but as far as I can tell there's a lack of consensus around how to integrate it currently...
22:11:10 <gabrielhurley> I've seen a lot of back-and-forth there
22:11:47 <jpich> I think we can get the separate panel into something useful during H
22:12:05 <jpich> I don't know if the ideal, put bits everywhere approach is possible to achieve during this cycle?
22:12:31 <gabrielhurley> Yeah, that seems to be the consensus. Has the UX group weighed in on the separate panel approach?
22:12:51 <gabrielhurley> If we're gonna go that route I want to make sure we're making something that's actually useful there
22:12:58 <jpich> Not yet, and agreed
22:12:59 <david-lyle> And I still worry about the performance implications of it everywhere unless the calls are asynchronous
22:13:10 <gabrielhurley> and I'm nt sold on the utility of just dumping a bunch of data into the dashboard and saying "figure it out yourself"
22:13:33 <jpich> david-lyle: Exactly, that's why the real time stuff should be in first (from my memories of the summit discussion)
22:13:37 <gabrielhurley> david-lyle: yeah, any general integration would need to be async
22:13:48 <jpich> Yes, at the moment it's difficult to read, especially with ids instead of names
22:13:51 <gabrielhurley> we had discussed doing it via AJAX for the time being
22:13:52 <jcoufal> jpich: Was there discussion about separate ceilometer panel in UX group? Havn't noticed that
22:13:56 <gabrielhurley> but I could be convinced to wait
22:14:11 <gabrielhurley> jcoufal: I don't believe there was
22:14:19 <gabrielhurley> there should be if that's the proposal on the table
22:14:35 <jpich> jcoufal: No, not yet, this needs to be brought up. Soon! :)
22:14:41 <jcoufal> gabrielhurley: me neither, can someone relevant post it there?
22:14:46 <gabrielhurley> jcoufal and jpich: can you two coordinate on getting the UX team to bring some thought to bear on the issue?
22:15:08 <david-lyle> If that's in place then I think it's useful asa separate panel, but also integrated into other panels
22:15:10 <jcoufal> deffinitely, if someone enlighten to the problem can help to formulate the question
22:15:12 <jpich> Note the current tables's content is ajaxified, though that may not be enough yet
22:15:35 <jpich> gabrielhurley, jcoufal: Will do
22:15:47 <david-lyle> Just status though and it doesn't handle transient objects at all
22:15:47 <gabrielhurley> okay. I'll leave it in y'alls hands to get a UX consensus and then get the code in line
22:15:51 <jpich> Can't wait to have a new place that is not the G+ community to have those discussions...
22:16:15 <jcoufal> jpich: you are aware of what is going on about ceilometer panel?
22:16:20 * gabrielhurley is staying out of this one and is just happy there *can* be discussions
22:16:47 <jpich> jcoufal: I have been reviewing and testing it so I can post screenshots and my understanding of the background at least
22:16:50 <jcoufal> jpcih: me too, G+ is being really bad, but at least as gabrielhurley said, we have somwhere to discuss
22:17:00 <jcoufal> jpich: perfect, sounds great
22:17:06 <gabrielhurley> cool. let's jump to the next set of BPs
22:17:22 <gabrielhurley> There's a slew of Quantum VPNaaS and FWaaS which I've seen nothing on so far...
22:17:33 <gabrielhurley> s/Quantum/Neutron
22:17:40 <gabrielhurley> I haven't followed what's happening with the development of those in Neutron
22:17:54 <gabrielhurley> but I'm inclined to postpone those, especially given what happened with LBaaS last time
22:18:09 <gabrielhurley> landing big stuff like that should go in the first or second milestone
22:18:29 <gabrielhurley> so I'm gonna bump those unless someone jumps forward with amazing code at some point during the cycle
22:18:47 <absubram_> :)
22:19:37 <david-lyle> If I remember correctly vpnaad is on gerrit as a work in progress at least
22:19:57 <david-lyle> vpnaas
22:20:12 <absubram_> yes.. it's in progress and there have been plenty of reviews.. its ongoing from what I understand
22:20:32 <gabrielhurley> If it gets proposed all the way in the earlier portion of H3 I'll consider it, but I really don't want to push half-finished features this time around
22:20:48 <gabrielhurley> we can always re-target blueprints
22:20:53 <david-lyle> Agreed
22:21:49 <gabrielhurley> so with that said, there's a couple of keystone-related BPs which I understand david-lyle and lcheng are still feeling good about, so I won't question those (for now ;-) )
22:22:29 <gabrielhurley> the Nova-related BPs are most of the small ones. I may actually grab some of those 'cuz I know a lot about 'em and they're pretty easy
22:23:18 <gabrielhurley> Honestly, the ones I'm questioning the most are the ones that are about Horizon itself
22:23:38 <gabrielhurley> things like the "building on OpenStack Dashboard" blueprint https://blueprints.launchpad.net/horizon/+spec/building-on-openstack-dashboard
22:23:54 <gabrielhurley> and inline editing for data tables https://blueprints.launchpad.net/horizon/+spec/inline-table-editing
22:24:26 <jcoufal> Inline editation is in process
22:24:40 <jcoufal> currently I got feedback on designes
22:24:43 <jpich> I will likely try and poke people's brains about the messages-on-login-page bp at some point. The approaches I tried so far didn't really work out
22:24:59 <jcoufal> and will share result with Ladislav who is asignee and has already some part done
22:25:02 <jpich> There's a WIP code review for inline editing I think
22:25:14 <gabrielhurley> jcoufal: I realize that, but the technical problems there are pretty significant. I haven't reviewed the code closely, but I know why *i've* been putting it off so long and I'm not sure we'll be able to get it ready in time.
22:25:44 <gabrielhurley> again, it's one that I'd be more favorable to seeing work continue on but target it to land in I1 or some such
22:25:44 <jcoufal> gabrielhurley: just to be sure, when si H3?
22:25:50 <gabrielhurley> 6 weeks
22:26:06 <gabrielhurley> I don't have the date off the top of my head
22:26:12 <gabrielhurley> oh
22:26:20 <gabrielhurley> Sept 9th
22:26:29 <jcoufal> ok, based on talks with Ladislav, I am not very optimistic about getting it in time...
22:26:31 <gabrielhurley> it's on the launchpad page, duh
22:26:38 <jcoufal> but will ask him to make sure
22:26:43 <gabrielhurley> I definitely want to see work continue on it
22:26:55 <gabrielhurley> but I'd rather make it awesome and land it early in I then rush it for H
22:27:37 <jcoufal> yup, I will encourage Ladislav in that work, I knwo he was blocked for some time now with other issues
22:28:04 <gabrielhurley> I think trying to get the "realtime" work done is a massive project for the Horizon framework as a whole, and will probably eat up a lot of developer attention
22:28:19 <gabrielhurley> and I think there's more group excitement around that one than any of the others
22:28:53 <gabrielhurley> so let me defer a few of these things we've talked about and see where the list is... one moment
22:29:28 <david-lyle> I would like to free up the centralized color pallette blueprint.  I have another issue I'd rather tackle after rbac.  https://blueprints.launchpad.net/horizon/+spec/domain-context-services
22:30:10 <david-lyle> I don't consider the pallete essential and it would fall in the horizon work
22:30:51 <absubram_> So umm.. gabriel, can I add a blueprint to the H3 list? If I'm able to get a code review out say end of July or within the first few days of August? :)
22:31:37 <gabrielhurley> absubram_: yep
22:31:57 <gabrielhurley> david-lyle: fine with me, though I'm sure some of the people who want to skin the dashboard will be sad ;-)
22:32:09 <gabrielhurley> those people should put their own devs on it.
22:32:44 <gabrielhurley> absubram_: I've added it to the list for H3
22:33:17 <absubram_> garielhurley: super! thanks! the neutron plugin for my horizon blueprint should go upstream within the next 10 days or so.. I should be able to push my stuff out soon after.. I need to test it first in Havana as most of my testing has been in grizzly
22:34:08 <absubram_> I did have a quick question about the api-detection blueprint though.. that's no longer in H3? so there's no longer a way to detect which netron plugin is being used?
22:35:39 <gabrielhurley> absubram_: for now just use a setting (you'll see other examples of how this works in the local_conf.py.example file)
22:36:35 <gabrielhurley> Okay. We're down to 20 BPs now...
22:36:42 <gabrielhurley> I guess that's *less* overkill
22:37:00 <gabrielhurley> I think I'll follow up with some of the folks who own blueprints that aren't at this meeting and see what their plans are
22:37:03 <david-lyle> what could go wrong?
22:37:34 <gabrielhurley> Also worth knowing is that now the "Low" priority on blueprints won't be tracked by ttx, so those indicate "nice to have" items that we can slip without issue.
22:37:44 <gabrielhurley> So you'll see me lower the priority on a number of things in this list
22:37:56 <absubram_> cool.. I'll check that out
22:38:01 <gabrielhurley> great!
22:38:07 <gabrielhurley> #topic open discussion
22:38:19 <gabrielhurley> now for everything else! what do people feel like talking about?
22:38:40 <absubram_> forum for discussion? :p
22:38:51 <absubram_> if not g+
22:39:24 <jcoufal> haha
22:39:26 <gabrielhurley> the discussion continues on the ML... are we getting closer to an answer or not?
22:39:38 <jcoufal> Not very sure
22:39:57 <gabrielhurley> As the OpenStack Dashboard PTL I'm sort of de facto involved, but it's ultimately not my decision, so I'm trying to let the community reach consensus
22:40:32 <jcoufal> there was just one acceptable recommendation from infra team for discourse, but not much people look very happy about that
22:40:34 <gabrielhurley> Normally I'd say "let's start trying things and if they don't work try another" but I worry that that will frustrate/fragment the community
22:40:45 <jcoufal> yes, I agree
22:40:57 <jcoufal> changing tools all the time is one thing what I don't want to do as well
22:41:32 <jcoufal> I will leave that for about 1 max 2 weeks and if there is no other proposal, I will go with the suggestion with most plus points
22:41:41 <gabrielhurley> sounds good
22:41:46 <absubram_> yeah
22:41:52 <gabrielhurley> maybe a summary email and ask for people to give a final opinon/vote
22:42:00 <gabrielhurley> then we just go with it
22:42:00 <jcoufal> sounds great
22:42:10 <jpich> Yes
22:42:27 <gabrielhurley> cool
22:42:28 <david-lyle> There's been the beginnings of discussion regarding potential navigation changes in horizon, I haven't seen a lot of input.  Maybe others aren't feeling the pain?
22:42:44 <gabrielhurley> I think the pain comes when we try to add more
22:42:49 <jcoufal> yeah, it also started in ML, just one answer on that
22:43:05 <gabrielhurley> that's one that I think we're better off solving via UX group/internally
22:43:15 <gabrielhurley> the ML at large doesn't care/isn't gonna be helpful
22:43:33 <gabrielhurley> but I would like to continue thinking on/addressing that issue
22:43:46 <jcoufal> david-lyle: I'd leave that 1 week maximum and if nobody replies, just summarize the issues and go for designs
22:43:50 <gabrielhurley> creating a better solution would help enable people to add their own dashboard content
22:44:54 <gabrielhurley> other items?
22:45:06 <jcoufal> gabrielhurley: I think david-lyle wanted to point out that there is ongoin discussion and not much people are sharing issues, I think navigation re-design is needed
22:45:38 <david-lyle> I hope rbac will enable collapsing some of existing duplication, but the dashboard needs better extensibility regardless
22:45:51 <gabrielhurley> agreed
22:47:28 <david-lyle> I have one other concern
22:48:25 <david-lyle> A lot of nova functionality is coming which is great, but i think we need a better approach than a new panel per feature
22:48:58 <gabrielhurley> not gonna argue with that. did you have a thought on that?
22:49:05 <david-lyle> Problem is that takes some ux work or discussion
22:49:51 <jcoufal> david-lyle: why is that a problem? time pressure?
22:50:11 <gabrielhurley> I think most of what's in H3 for Nova fits into existing sections, whether as added data in instance tables, extra tabs on instance details, etc.
22:50:30 <gabrielhurley> in the longer term I agree that our organizational groupings may warrant reexamination
22:51:09 <david-lyle> I guess the org is big issue.  But more than just panels
22:51:24 <absubram_> I've gtg guys.. thanks.. see you tomorrow
22:51:39 <gabrielhurley> yep. Let's revisit that at the next meeting
22:52:01 <gabrielhurley> I think we should call this one for today, and maybe start discussing design issues again next time.
22:52:06 <gabrielhurley> sound reasonable?
22:52:06 <david-lyle> Requiring ux thought is not a problem, just pointing out a work item that's open :)
22:52:26 <jcoufal> gabrielhurley: sounds good
22:52:29 <david-lyle> Sure
22:52:42 <gabrielhurley> cool. thanks everyone! this has been a very productive meeting. have a great week!
22:52:45 <gabrielhurley> #endmeeting