20:00:19 <robcresswell> #startmeeting horizondrivers
20:00:21 <openstack> Meeting started Wed Sep 16 20:00:19 2015 UTC and is due to finish in 60 minutes.  The chair is robcresswell. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:25 <openstack> The meeting name has been set to 'horizondrivers'
20:00:45 <robcresswell> Anyone about for the meeting?
20:00:56 <mrunge> o/
20:01:05 <bpokorny> o/
20:01:08 <jwy> hi
20:01:19 <ducttape_> o/
20:01:25 <lhcheng> o/
20:01:29 <robcresswell> David is away atm, so have stolen chair. bpokorny, just a few announcements to make then will pass on to you.
20:01:40 <bpokorny> Sounds good
20:01:55 <robcresswell> #link https://wiki.openstack.org/wiki/Meetings/HorizonDrivers
20:02:05 <robcresswell> Agenda, not much on it
20:02:07 <tsufiev> O/
20:02:30 <robcresswell> First thing is: The etherpad is up for Mitaka Summit Session Topics
20:02:42 <robcresswell> #link https://etherpad.openstack.org/p/horizon-mitaka-summit
20:03:10 <robcresswell> Note the bits at the top too, so we don't end up with a huge mess.
20:03:44 <robcresswell> Second notice is: rel mgmt wants us to aim for RC1 at the end of this week I believe.
20:03:54 <robcresswell> #link https://launchpad.net/horizon/+milestone/liberty-rc1
20:04:45 <robcresswell> Cores really need to be picking up bugs. We have 3/4 bps and <20 bugs, so get on with it :p
20:05:42 <mrunge> ah yes, david mentioned earlier, bugs <= medium prio might not make it into rc1
20:05:53 <robcresswell> Yes, I've been removing a few already.
20:06:14 <robcresswell> The translations appear to be fixed again now.
20:06:23 <mrunge> \o/
20:06:42 <mrunge> but selenium is broken again?
20:06:46 <robcresswell> But we still have several High bugs that really need reviews. A lot are not too complex.
20:06:55 <robcresswell> Yes, it broke late last week again I think
20:07:00 <mrunge> darn
20:07:11 <robcresswell> Integration is still up though I *think*. tsufiev may have more to say on the matter.
20:07:45 <tsufiev> robcresswell: no news = good news
20:07:52 <robcresswell> Excellent.
20:08:23 <robcresswell> That's all from me. Expect pings about bugs/bps, sorry if I get annoying.
20:08:31 <robcresswell> bpokorny, your turn :)
20:08:39 <bpokorny> Thanks!
20:08:45 <mrunge> robcresswell, you're doing fine :)
20:08:50 <mrunge> keep pinging!
20:09:02 <bpokorny> Just wanted to publicize this one for the M release:  https://blueprints.launchpad.net/horizon/+spec/nova-server-groups
20:09:18 <bpokorny> Ability to manage server groups via Horizon.
20:09:28 <robcresswell> #link https://blueprints.launchpad.net/horizon/+spec/nova-server-groups
20:09:35 <lhcheng> bpokorny: is that different as service groups?
20:09:45 <tsufiev> mrunge: iirc selenium fix is blocked by DOA issue
20:09:54 <bpokorny> I put a review out already for setting the server groups.
20:10:16 <bpokorny> lhcheng:  I think so...  I'll have to look into what service groups are.
20:10:18 <mrunge> tsufiev, by which doa issue?
20:10:46 <bpokorny> But server groups are managed on the cli with the "nova server-group-XXXX"  commands.
20:11:34 <tsufiev> mrunge: the one Richard was trying to fix with lhcheng, but then it appeared that we need to rename DOA ). Sorry, I'm from tablet, difficult to paste links
20:11:59 <robcresswell> BP looks solid IMO.
20:12:08 <mrunge> tsufiev, ok, thanks. I'll look that up or ask you tomorrow
20:12:14 <lhcheng> bpokorny: we're using the term service group in-house, sounds like the same actually :) it allows a user to set a scheduler hint
20:12:20 <robcresswell> Has code, and is supported elsewhere.
20:12:27 <lhcheng> bpokorny: that's going to be useful for us, thanks
20:12:31 <robcresswell> SO we aren't waiting on service updates :D
20:12:45 <bpokorny> lhcheng: Yep, that's what's needed.
20:12:46 <ducttape_> nova service list  (etc) is usually the service status.   more confusion
20:12:59 <bpokorny> Scheduler hints
20:13:49 <bpokorny> So as a next step, I'll ping David later about approval for the bp for the M release.
20:14:05 <bpokorny> And that's all I had on that one :)
20:14:10 <robcresswell> We can just vote now
20:14:18 <robcresswell> There should be several cores here
20:14:18 <lhcheng> bpokorny: cool
20:14:24 <lhcheng> + for me
20:14:27 <bpokorny> Ah, sounds good.
20:14:36 <tqtran> + for me as well
20:14:51 <robcresswell> Yeah I think it looks good. Approved. We dont have an M tag yet, so will mark for 'next'
20:14:55 <TravT> bpokorny: is this just adding hints?
20:15:15 <bpokorny> The first part is being able to boot a VM with hints.
20:15:42 <bpokorny> But as a next step, we'll also probably want the ability to create/edit/delete server groups.
20:16:28 <ducttape_> bpokorny - there are some scheduler hints available to admins - like place the vm on a specific host
20:16:54 <ducttape_> we'd need to check the auth reqmts on some of those options, perhaps ?
20:16:58 <lhcheng> ducttape_: aggregates + flavor metadata (I think)
20:17:29 <TravT> this type of information you can get from the metadata catalog in glance.
20:18:11 <TravT> you can create metadata items and then set their visibility
20:18:26 <lhcheng> the hints available to user should be configurable, depending how the deployer wants their scheduler to work
20:18:28 <TravT> and then only show to users based on their ability to see them
20:18:41 <TravT> that is what glance metadata catalog gives you
20:18:49 <ducttape_> bpokorny - your life is getting complex ^  ;)
20:19:09 <TravT> we already use it in horizon.
20:19:11 <bpokorny> ducttape_:  Heh - always how it goes :)
20:19:18 <ducttape_> but it seems like this is interesting work / supported by most all
20:19:40 <TravT> it isn't too hard...
20:19:50 <TravT> there is an api for it.
20:20:11 <bpokorny> As far as I've seen, there's not a way to specify server groups currently via Horizon.
20:20:14 <ducttape_> boot from vol TravT, just to throw a monkey wrench at you
20:20:33 <mrunge> bpokorny, yes, that's right
20:21:14 <ducttape_> right now you'd have to use live migration to simulate something close to that
20:24:53 <robcresswell> Has the dust settled?
20:25:35 <bpokorny> I think so?  :)
20:25:40 <robcresswell> If there are still people around, I'd like to discuss the remaining RC1 BPs and see if people have time to review them.
20:26:14 <robcresswell> #info Nova Server Groups blueprint approved
20:26:56 <robcresswell> lhcheng: Were you still intending to look over https://blueprints.launchpad.net/horizon/+spec/database-clustering-support ?
20:27:31 <robcresswell> I know I have it on my list. It's a big bp, so would be good to get 2 core reviews if possible.
20:27:56 <lhcheng> ugh, I just did a first pass on the code. have to look at it again
20:28:15 <lhcheng> robcresswell: yeah, its a big chunk of  code
20:28:24 <lhcheng> will try to look at it tomorrow
20:28:47 <robcresswell> lhcheng: Yup. Thanks. I will do the same.
20:29:13 <robcresswell> There is also a sahara bp
20:29:18 <robcresswell> #link https://blueprints.launchpad.net/horizon/+spec/unified-job-interface-map-ui
20:29:51 <robcresswell> It's had a number of reviews from Horizon folk too in the last patch set.
20:30:12 <robcresswell> I don't suppose anyone will volunteer to review?
20:30:36 <robcresswell> tqtran, clu_ *hint hint*
20:31:44 <tsufiev> robcresswell: I'm still against embedding html into js, but an author didn't respond
20:32:01 <mrunge> then it's easy: -1
20:32:10 <mrunge> I can do that tsufiev :P
20:32:25 <robcresswell> tsufiev: I had a discussion with them on IRC. IMO with Sahara, since it is likely to move out soon, it isn't really a blocker.
20:32:45 <robcresswell> As long as it's not insanity, I'm inclinced to be lenient with it.
20:32:57 <robcresswell> But I'm not sure what davids stance is on that
20:32:57 <tsufiev> Heh. Then I don't see much sense in reviewing it at all
20:33:12 <tsufiev> If it doesn't beak horizon, it's fine
20:33:24 <mrunge> We have to keep that code alive for quite a time
20:33:27 <robcresswell> Well, difference between "where does a one-time-usage-template live" and "this code is totally insane"
20:33:29 <mrunge> cough
20:33:53 <robcresswell> As I understood, they are planning on moving it out in M, and contrib/ was the first stage to that.
20:33:59 <robcresswell> But not sure on absolute plans.
20:34:34 <tsufiev> robcresswell: insanity creeps in with small steps ;)
20:34:48 <mrunge> I mean, embedding js in html is not the worst thing in sahara dashboard
20:35:04 <robcresswell> tsufiev: Haha. If you feel strongly, -1 the patch set.
20:35:18 <robcresswell> I'll look at it after the databases code.
20:35:34 <tsufiev> mrunge: js inside html is okay, reverse is not IMO
20:35:56 <mrunge> ouch :x
20:35:57 <tsufiev> Okay, i'll say my word and -1 it
20:37:04 <robcresswell> Sure thing
20:37:37 <robcresswell> Right I think that is all on the agenda. Does anyone else have any other blueprints or points they'd like to raise (or argue about?)
20:37:43 <TravT> yeah
20:38:00 <TravT> re: angularizing the tables.
20:38:17 <TravT> i think there are a couple bug fixes we should still do on master with what is there.
20:38:19 <TravT> but otherwise:
20:38:20 <TravT> https://openstack.nimeyo.com/58582/openstack-dev-horizon-patterns-for-angular-panels
20:39:04 <robcresswell> Yep, I've yet to reply to this
20:39:52 <robcresswell> TravT: Which bugs? Tag them for rc1 please
20:40:15 <robcresswell> TravT: They may not make it, but still be good to get visibility.
20:40:23 <TravT> here's one.
20:40:29 <mrunge> yes, robcresswell +1
20:40:33 <TravT> https://review.openstack.org/#/c/217295/
20:40:39 <TravT> it is under the larger bp
20:40:56 <tsufiev> robcresswell: btw, if an rc1 bug has 2 not tagged dependency bugs/commits, should I tag them as well?
20:41:42 <mrunge> TravT, is that a bug or a feature?
20:41:51 <robcresswell> tsufiev: Wow, what a question. I would say yes, tag them, but don't expect them to get in or not.
20:42:12 <mrunge> since we're that late, I'd try to fix bugs and not to introduce new features
20:42:46 <TravT> probably could change to a bug and fix the usage of '-' in numerous other spots in angular horizon.
20:42:48 <tsufiev> robcresswell: it's about https://review.openstack.org/#/c/215196/
20:42:49 <mrunge> sounds more like it's fixing a bug?
20:43:10 <TravT> it is a bug
20:43:32 <robcresswell> TravT: Probably better to tag as a bug, rather than feature work
20:43:40 <robcresswell> tsufiev: looking...
20:44:05 <robcresswell> tsufiev: Yep, tag all as RC1, we'll see where we get to
20:44:15 <tsufiev> robcresswell: okay
20:44:32 <TravT> yeah, i could do that... could then track down other places we need to replace '-' with using the filter.
20:45:05 <robcresswell> tsufiev: As I said, I'm being pretty sharp about the bug list, so it may be removed again shortly, but best to draw attention to it first so its an educated decision.
20:47:00 <mrunge> uhm, bugs still can be fixed after rc1 release
20:47:17 <mrunge> but they will need backports from mitaka to liberty branch then
20:47:24 <robcresswell> Ha, yeah ofc
20:47:28 <mrunge> just as a reminder
20:47:39 <mrunge> that's not that easy for features
20:47:48 <robcresswell> We may also have an RC2 depending. I just meant they may get removed from rc1 etc.
20:47:54 <TravT> i'm working on a searchlight panel which has uncovered some bugs in magic search.  I might submit them up too.  kfox1111 also experienced one of them on his app catalog plugin
20:48:23 <mrunge> there are quite a few bugs currently
20:48:35 <robcresswell> The remaining BPs should get in IMO, if people will review. I've +2'd Curvature, merged shelving, looking into dbs/ sahara
20:49:15 <robcresswell> mrunge: Indeed, but it is shrinking. I've got it down to about half of Monday, obsoleted a few, reviewed/ merged a lot :p
20:49:39 <mrunge> robcresswell, pain points are still in
20:49:57 <robcresswell> mrunge: pain points?
20:50:11 <mrunge> uhm xss
20:50:27 <mrunge> the re-login bug
20:50:47 <mrunge> etc. just walk through the list of high prio bugs in horizon
20:50:53 <mrunge> not to mention doa bugs
20:51:25 <mrunge> which currently fall out of the view, because it's a different project
20:51:26 <robcresswell> Yep. I believe the re-login bug had a fix out
20:51:34 <mrunge> nope
20:51:41 <mrunge> there's a proposal
20:51:47 <mrunge> but it's not merged
20:51:47 <robcresswell> Ah, my mistake
20:52:07 <mrunge> and tsufiev said, it breaks project switching
20:52:21 <robcresswell> :(
20:52:43 <tsufiev> mrunge: did I?
20:52:52 <mrunge> tsufiev, I believe so
20:54:22 <robcresswell> Well, we've got a busy couple of days ahead then :)
20:54:30 <mrunge> yes!
20:54:37 <tsufiev> Okay, let's first fix DOA job that's failing, and then re-login bug and see what's next...
20:55:11 <robcresswell> Yes. Please focus reviews on those bugs targeted for RC1, and mention any others that may be blockers.
20:56:03 <robcresswell> Curvature needs another core review, me and lhcheng will target dbs, and then sahara I'll look at afterwards, see if I can drag someone else in too.
20:56:28 <robcresswell> That covers the FFEs I think, unless angular images is looking ready to go live.
20:58:01 <TravT> robcresswell: there is good progress on them, but I don't think they'll be ready.
20:58:27 <robcresswell> TravT: h'okay. I'm glad development was kept open though.
20:59:01 <TravT> yeah, we've really made some improvements in the last few weeks and found some additional ways to reduce code redundancy
20:59:37 <robcresswell> Good stuff. I'm looking forward to settling on patterns so I can update my networks panel and push it.
21:00:11 <robcresswell> Anyway, thats the hour
21:00:18 <robcresswell> Thanks everyone!
21:00:27 <mrunge> thanks robcresswell
21:00:32 <robcresswell> #endmeeting