12:00:29 <david-lyle> #startmeeting Horizon
12:00:29 <openstack> Meeting started Wed Nov 11 12:00:29 2015 UTC and is due to finish in 60 minutes.  The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:00:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
12:00:33 <openstack> The meeting name has been set to 'horizon'
12:00:45 <r1chardj0n3s> o/
12:00:58 <pkarikh> hi!
12:01:04 <robcresswell> o/
12:01:09 <amotoki> o/
12:01:27 <mrunge> hey o/
12:01:52 <doug-fis_> \o morning all
12:02:27 <david-lyle> Let's get rolling
12:03:16 <david-lyle> Several of us attended the summit in Tokyo two weeks ago
12:04:12 <david-lyle> as a reminder, the gist of the conversations were recorded
12:04:15 <david-lyle> #link https://wiki.openstack.org/wiki/Design_Summit/Mitaka/Etherpads#Horizon
12:04:28 <kzaitsev_mb> o/
12:04:54 <david-lyle> amotoki: there was little contention and people were ready to do other things :)
12:05:05 <amotoki> :)
12:05:16 <neillc_away> o/
12:05:24 <david-lyle> 2 people were there at 2 pm, so we adjourned and enjoyed Tokyo
12:05:38 <doug-fish> A much different Friday than the previous summit.
12:05:44 <r1chardj0n3s> we were very efficient, yes
12:05:57 <r1chardj0n3s> I think the rapid agreement on priorities helped :-)
12:06:12 <neillc> So much agreement...
12:06:15 <david-lyle> I also want to thank amotoki for arranging the meetup on Monday night, excellent location and a good time. Thank you!
12:06:37 <robcresswell> +1, although I was late :)
12:06:42 <r1chardj0n3s> I was sad I missed it, but I flew in too late
12:06:58 <doug-fish> yes - thanks amotoki!
12:07:05 <amotoki> very happy to hear you all enjoyed Tokyo (and more Japan) !
12:08:08 <david-lyle> the resulting priority list for Mitaka can be found
12:08:11 <david-lyle> #link https://etherpad.openstack.org/p/mitaka-horizon-priorities
12:09:28 <david-lyle> Highest priority is really documenting plugins and theming and testing, all of which are in progress
12:10:18 <david-lyle> basically give users and operators enough information to successfully use all the nice features we've been adding
12:11:09 <david-lyle> r1chardj0n3s: I saw that our pbr issue is unlikely to become unblocked
12:11:10 <david-lyle> ?
12:11:25 <david-lyle> which was another critical priority
12:11:31 <r1chardj0n3s> that's a problem because it's blocked on lifeless
12:11:56 <r1chardj0n3s> I've engaged with him post-summit but I think he's got other priorities
12:12:05 <r1chardj0n3s> and he's the only pbr core
12:12:26 <david-lyle> we may need to consider other solutions :/
12:12:29 <mrunge> so, that leaves us hanging somewhere?
12:12:54 <r1chardj0n3s> I'm not sure what other solutions there are, but yeah, we're blocked on being able to release thru openstack infra :/
12:13:37 <mrunge> if releasing is blocked, isn't that a super-urgent-high-critical issue?
12:13:41 <david-lyle> adding another group with publishing privileges other than openstack-infra?
12:14:13 <r1chardj0n3s> so the core problem is that we need some credentials to publish to pypi that we can all share
12:14:26 <r1chardj0n3s> openstack infra have solved that problem
12:14:38 <r1chardj0n3s> without using their solution, we'll have to implement our own solution
12:14:55 <david-lyle> are there groups on pypi?
12:15:02 <r1chardj0n3s> no, sadly
12:15:09 <robcresswell> :/
12:15:13 <mrunge> I *know* I released at least one django_openstack_auth just by tagging it
12:15:15 <r1chardj0n3s> you can have multiple maintainers for a package
12:15:25 <mrunge> but that has been a year since then...?
12:15:32 <david-lyle> so openstack-infra is just an account that multiple people have access to?
12:15:33 <amotoki> IIRC openstack infra use openstack-ci account to access PyPI.
12:15:41 <david-lyle> mrunge: d-o-a is not an issue
12:15:50 <r1chardj0n3s> mrunge: that package doesn't run into the problem that the xstatic packages do: 4 digit versions are rejected by infra because pbr
12:15:55 <david-lyle> only xstatic with x.x.x.x version strings
12:16:18 <amotoki> ahh... pbr now supports only x.y.z
12:16:28 <david-lyle> yes
12:16:46 <david-lyle> and not looking highly likely to change
12:17:36 <r1chardj0n3s> the spec to change pbr is https://review.openstack.org/#/c/205623/ fyi
12:17:48 <robcresswell> Are pbr changes just at the mercy of a single developer?
12:17:54 <r1chardj0n3s> robcresswell: yes
12:18:15 <robcresswell> *facepalm*
12:18:16 <r1chardj0n3s> personally I think that's an amazing risk that OpenStack is taking
12:19:07 <robcresswell> What are the alternatives? We produce our own horizon-infra procedure?
12:19:32 <david-lyle> another alternative is we deviate from the upstream version numbers
12:19:39 <david-lyle> which will be confusing
12:19:44 <r1chardj0n3s> yeah, had considered that, but that would be tres confusing
12:20:23 <david-lyle> just increment each release and document in a version file or something the real version
12:20:36 <r1chardj0n3s> I think we considered using .postN at one point, but I believe pbr rejects those too
12:20:50 <r1chardj0n3s> but I'm not 100% sure on that. and testing it is difficult ;-)
12:20:55 <david-lyle> but since we already have releases that would have us starting in weird places
12:21:18 <david-lyle> and the requirements changes would be challenging
12:21:28 <r1chardj0n3s> I know we can't use +internal extensions to versions because *pypi* rejects those
12:21:38 <david-lyle> r1chardj0n3s: I thought we tried the postN bit at the midcycle
12:21:46 <r1chardj0n3s> david-lyle: yeah, that was my recollection too
12:21:53 <david-lyle> to no success
12:22:09 <r1chardj0n3s> however, I recall some muttering about .postN in another place since then
12:22:27 <r1chardj0n3s> I'm gonna write myself a note to look into that, *and* to poke Robert again to see if I can get some movement
12:22:49 <r1chardj0n3s> even if that movement is a final "no"
12:23:01 <david-lyle> we can try again I suppose, maybe it wasn't postN exactly that we tried, but another of the key words in the 4th value
12:23:03 <r1chardj0n3s> which would suck, but at least we know where we stand (not a real project)
12:23:14 <david-lyle> haha
12:25:17 <david-lyle> The other critical item from the summit, not document related was moving forward with some angular content on master
12:26:00 <david-lyle> any general feedback on the summit this time?
12:26:51 <robcresswell> Felt like it was good overall. Cross-project sessions seemed pretty unproductive
12:27:16 <robcresswell> I think next time we should work with the project teams beforehand to discover issues, then discuss how to solve them at the summit
12:28:02 <david-lyle> robcresswell: re: x-project or horizon related issues?
12:28:14 <robcresswell> x-project
12:28:23 <robcresswell> Ceilometer was the one I was specifically thinking of
12:28:34 <robcresswell> discussion seemed a little vague
12:28:47 <robcresswell> That may just be my lack of experience with it though.
12:29:04 <r1chardj0n3s> I learned a lot about ceilometer in that session actually ;-)
12:29:19 <r1chardj0n3s> I was starting from a pretty ignorant position though, admittedly
12:29:40 <david-lyle> robcresswell: I thought you were referring to the grander x-project sessions
12:30:21 <robcresswell> Oh, no sorry
12:30:34 <david-lyle> I agree those two sessions could have been more productive, but sometimes the summit is just a good way to start the conversation with the interested parties in the room
12:30:58 <david-lyle> get everyone up to speed on the unknown together :)
12:31:27 <r1chardj0n3s> yep :)
12:31:39 <amotoki> it usually happens in neutron sessions..
12:31:44 <robcresswell> Fair enough, t'was just a suggestion
12:31:56 <david-lyle> robcresswell: not a bad one
12:32:34 <david-lyle> any other general feedback?
12:32:50 <david-lyle> or specific, I suppose
12:34:07 <r1chardj0n3s> personally, I'd like to be more prepared for having time like we did on Friday - with the assumption that the next summit will go just as swimmingly :-)
12:35:17 <david-lyle> r1chardj0n3s: after the last summit I just assumed we'd need all that time
12:35:21 <robcresswell> Difficult to judge though. Travis was saying its the first summit he hasn't left early friday evening, and its the first one he could've left early and not missed anything.
12:35:34 <robcresswell> david-lyle: Agreed
12:35:41 <r1chardj0n3s> yep
12:35:56 <amotoki> agree.
12:36:10 <doug-fish> maybe it's not a good assumption we won't meet Friday, but maybe we could reach some agreement on Thursday if we have any other work to do.
12:36:10 <david-lyle> the other option is not schedule something in the afternoon and have informal time if needed
12:36:41 <doug-fish> I'd rather have the time scheduled - it has been important and useful in past summits
12:36:50 <r1chardj0n3s> I think if I had've been more prepared I could've sat with a couple of people and been productive on Friday arvo
12:36:57 <r1chardj0n3s> er afternoon
12:37:09 <david-lyle> r1chardj0n3s: sure
12:37:14 <neillc> We were all pretty tired by Friday afternoon. That's likely constant.
12:37:29 <r1chardj0n3s> I just wasn't expecting to have to, and then Friday came and I was too exhausted to think so hard ;-)
12:37:46 <amotoki> Friday afternoon is a spare time. if we have topics, let's discuss. otherwise enjoy sightseeing or others!
12:38:15 <r1chardj0n3s> the one thing that not hanging around on Friday meant was no Horizon post-summit drinks, which I also missed
12:38:23 <r1chardj0n3s> because we disintegrated early
12:38:37 <david-lyle> r1chardj0n3s: true
12:38:54 <doug-fish> maybe we need to schedule more mid-week drinking?
12:38:58 <neillc> post summit drinks would have been good :)
12:38:59 <david-lyle> we didn't properly close things out
12:39:13 <robcresswell> I think we should do the meetup later in the week
12:39:22 <amotoki> :)
12:39:39 <r1chardj0n3s> well, I don't want to be that guy, but the *swift* team all got a harbor cruise
12:40:06 <r1chardj0n3s> of course, swift is a Real Project so someone paid for it  :-/
12:40:27 <robcresswell> Hey Horizon is real, HP made t-shirts this time
12:40:55 <robcresswell> I loved in Vancouver when they told Travis "we don't really have anyone working on Horizon"
12:41:01 <amotoki> i think swift is special in this summit. no sponsor in neutron meetup  too
12:41:13 <r1chardj0n3s> anyway, I did want to make the point that I think things went so smoothly partly because we met only three months prior for the midcycle
12:41:28 <david-lyle> r1chardj0n3s: I would agree with that
12:41:30 <r1chardj0n3s> so I think there's value in doing that thing again
12:41:32 <amotoki> robcresswell: really? we use horizon for production
12:41:40 <r1chardj0n3s> amotoki: so do we
12:41:50 <robcresswell> amotoki: Yeah, just some naive marketing person I think.
12:42:01 <r1chardj0n3s> (though not in the public cloud, of course)
12:42:42 <robcresswell> So, another mid-cycle?
12:42:55 <neillc> +1
12:43:02 <doug-fish> +1
12:43:05 <david-lyle> I took the action to try and track potential venues for a midcycle
12:43:28 <david-lyle> but since I've been back at it for 1 day now since the summit, I have not made much progress
12:43:40 <david-lyle> west coast US would be easy to arrange
12:43:54 <r1chardj0n3s> david-lyle: San Diego has been bandied about by several folks
12:43:59 <david-lyle> SJC or PDX would be easy
12:44:10 <r1chardj0n3s> not sure precisely who the host org would be there
12:44:30 <david-lyle> I'm not sure who that cares about horizon has an office in SD
12:44:43 <david-lyle> HP used to, but not since post split, I belive
12:45:15 <david-lyle> Seattle is also probably a possibility
12:46:24 <david-lyle> I'll work on some concrete options
12:46:58 <r1chardj0n3s> coolo
12:47:23 <david-lyle> I know I can offer to host in Santa Clara or Portland, but I'll build the list
12:48:20 <david-lyle> #topic bug report
12:48:24 <david-lyle> #link https://wiki.openstack.org/wiki/Horizon/WeeklyBugReport
12:49:37 <robcresswell> \o/
12:50:30 <robcresswell> So I updated it for this week, and updated the criteria too. It's a wiki, so feel free to add to it. I'll keep adding new things and removing merged ones each week.
12:51:01 <robcresswell> Been speaking with matt borland too, so we're synced on priorities with some of the angular patches
12:51:16 <david-lyle> thanks
12:51:21 <david-lyle> looks good
12:51:53 <robcresswell> Ideally, we can merge the majority of these each week, and keep on top of the priority list, but it will require active participation so bookmark it folks :)
12:53:24 <david-lyle> isn't the magic search one merged, that bug report is a mess, but there is a merge message at the bottomw
12:53:41 <david-lyle> launchpad disagrees
12:53:57 <david-lyle> but launchpad
12:54:48 <robcresswell> I think there are more for it
12:55:05 <robcresswell> At least, there seem to be active patches still attached to it
12:55:13 <david-lyle> ok
12:55:19 <david-lyle> just scanning
12:55:37 <david-lyle> #topic reno
12:55:41 <david-lyle> #link http://lists.openstack.org/pipermail/openstack-dev/2015-November/078301.html
12:55:48 <david-lyle> robcresswell: o/
12:55:51 <robcresswell> Yup
12:55:58 <robcresswell> So, lhcheng has a patch to hadd this
12:55:59 * david-lyle admits still hasn't reread
12:56:16 <robcresswell> #link https://review.openstack.org/#/c/243897/
12:56:28 <david-lyle> so our release notes are to be yaml rather than rst?
12:56:38 <robcresswell> This is how we'll manage our release notes now. It's not as draconian as a I first thought.
12:56:52 <amotoki> I haven't figured out the full picture yet
12:56:59 <david-lyle> well yaml that -> rst
12:57:35 <amotoki> another one, the last call for juno stable update has been sent, but horizon juno gate is broken. can we handle it, or no action?
12:57:50 <robcresswell> Yeah, yaml then rst
12:58:10 <robcresswell> Makes it easier to do it as we go, and we control what needs to be reno'd
12:58:47 <robcresswell> So it just means a little bit more "stuff" with our larges features, so bps and larger bugfixes I assume.
12:58:59 <david-lyle> the positive for me is that we can require release notes as part of a change
12:59:09 <robcresswell> Yeah
12:59:09 <david-lyle> amotoki: good topic
12:59:16 <david-lyle> jumping
12:59:19 <david-lyle> #topic juno
12:59:20 <amotoki> we have less than 1 min.
12:59:29 <david-lyle> the gate is broken
12:59:30 <amotoki> just link http://lists.openstack.org/pipermail/openstack-dev/2015-November/079065.html
12:59:42 <robcresswell> Is the gate issue a known bug that we can fix? Or new?
12:59:42 <david-lyle> there is a bit of a hacky patch to fix it
12:59:54 <david-lyle> robcresswell: known issue
13:00:00 <david-lyle> dependency hell
13:00:00 <amotoki> it is not our fault, but dependency is broken.
13:00:29 <amotoki> dark side of oslo dependency
13:00:50 <david-lyle> we should fix it, but I'm not sure I saw much meat for juno patches outstanding
13:01:07 <david-lyle> but this will be the last formal juno release
13:01:45 <david-lyle> fixing the gate would be a nice start
13:01:52 <r1chardj0n3s> a co-worker has a phrase be uses to describe the dependency hell around juno "tangled web of onions"
13:02:04 <david-lyle> ok, time's up
13:02:08 <david-lyle> thanks everyone
13:02:13 <david-lyle> #endmeeting