12:01:08 <david-lyle> #startmeeting horizondrivers 12:01:09 <openstack> Meeting started Wed Oct 21 12:01:08 2015 UTC and is due to finish in 60 minutes. The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:01:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 12:01:12 <openstack> The meeting name has been set to 'horizondrivers' 12:01:44 <david-lyle> anyone driving? 12:01:52 <tsufiev> o/ 12:01:56 <doug-fish> \o 12:02:19 <tsufiev> also robcresswell said that he will soon after beginning of the meeting 12:02:53 <masco> o/ 12:03:51 <david-lyle> Only general item 12:04:15 <david-lyle> #link http://mitakadesignsummit.sched.org/overview/type/horizon 12:04:29 <david-lyle> #link https://wiki.openstack.org/wiki/Design_Summit/Mitaka/Etherpads#Horizon 12:04:44 <david-lyle> for summit stuff 12:06:45 <david-lyle> any pressing bps people want to discuss, or should I just pull from the back of the deck? 12:07:09 <doug-fish> I don't have any priority blueprints today 12:07:22 * doug-fish still looking at the schedule 12:07:29 <tsufiev> neither do I 12:07:56 <david-lyle> ok, the zaqar-ui bp should wait until the summit 12:08:29 <robcresswell> Sorry for lateness 12:08:38 <david-lyle> #topic https://blueprints.launchpad.net/horizon/+spec/volume-type-access 12:08:43 <tsufiev> david-lyle, is tests discussion going to happen on the contributors meetup? 12:08:43 <david-lyle> #link https://blueprints.launchpad.net/horizon/+spec/volume-type-access 12:09:02 <david-lyle> tsufiev: it sure can 12:09:28 <tsufiev> ok, that's good ) 12:09:36 <david-lyle> add things you'd like to discuss on the etherpad 12:09:48 <david-lyle> for the meetup 12:10:12 <tsufiev> ack 12:11:34 <david-lyle> for the bp seems like one piece was overlap otherwise just rounding out API coverage 12:11:48 <tsufiev> seems that this blueprint is a duplicate of https://blueprints.launchpad.net/horizon/+spec/volume-type-description is it? 12:12:06 <david-lyle> just the first of 3 patches 12:12:15 <david-lyle> the next two alter access 12:12:36 <david-lyle> IIUC more like images 12:12:39 <tsufiev> ah, ok 12:13:05 <robcresswell> Perhaps just leave a comment asking the author to fill it out more 12:13:34 <david-lyle> robcresswell: the code is all up already 12:13:59 <david-lyle> the commit messages are a bit more informative, a bit 12:14:11 <robcresswell> :( 12:14:40 <david-lyle> second one is better 12:14:56 <robcresswell> I dislike trawling code to understand what features have been added, but we shouldnt be blocking on it. "red tape" etc. 12:15:17 <david-lyle> I'm going to approve, seems like new basic functionality we should support 12:15:22 <robcresswell> Looks like its useful functionality 12:15:26 <robcresswell> agreed ^^ 12:15:37 <david-lyle> robcresswell: feel free in the reviews to ask for a more complete commit message 12:16:03 <david-lyle> #info https://blueprints.launchpad.net/horizon/+spec/volume-type-access approved 12:16:10 <robcresswell> Yes, I think so for the first patch. 12:16:32 <david-lyle> the first of 3 is abandoned 12:16:42 <david-lyle> the 2nd could use a better commit message 12:16:55 <david-lyle> I'm going to skip ceilo bp for now too 12:17:08 <robcresswell> Ya, I meant of the two active patches. 12:17:23 <david-lyle> there will be a ceilo fishbowl where we'll learn more 12:17:24 <doug-fish> I added a couple of comments to the whiteboard on https://blueprints.launchpad.net/horizon/+spec/volume-type-access 12:18:28 <david-lyle> #topic https://blueprints.launchpad.net/horizon/+spec/uuid-integration-tests 12:18:38 <david-lyle> #link https://blueprints.launchpad.net/horizon/+spec/uuid-integration-tests 12:18:53 <david-lyle> seems logical and I don't see a reason to block 12:19:23 <robcresswell> Yeah, it doesn't really need any context, the linked bp explains enough. 12:19:28 <tsufiev> yes, I think it won't hurt 12:19:41 <david-lyle> we could make the integration tests pluggable to tempest at some point 12:19:41 <robcresswell> +1 for approval 12:19:47 <tsufiev> I like the assumption that we'll have as many tests as in Tempest :) 12:20:24 <doug-fish> david-lyle: FWIW I don't think our Horizon tests will _ever_ go into Tempest 12:20:47 <david-lyle> doug-fish: not in tempest, a plugin to tempest 12:20:57 <david-lyle> in fact our current test will be doing the same 12:21:28 <doug-fish> I wasn't aware of that - version-less-ness of tempest isn't a concern? 12:21:42 <tsufiev> david-lyle, so, are going to move tests out of horizon repo? 12:21:45 <doug-fish> is there a blueprint for this effort? 12:21:56 <david-lyle> hold up 12:22:03 <david-lyle> we have one test in tempest 12:22:18 <david-lyle> we would move that scenario into the horizon repo 12:22:28 <david-lyle> to work as a plugin to tempest 12:22:56 <david-lyle> the when we run the gate jobs, our plugin is loaded into tempest 12:23:23 <david-lyle> there is currently no point to have the horizon test scenario run for nova and neutron patches 12:23:39 <david-lyle> but random failures can block their patches 12:23:51 <david-lyle> more big tent stuff 12:24:01 <david-lyle> #info https://blueprints.launchpad.net/horizon/+spec/uuid-integration-tests approved 12:24:39 <tsufiev> hm... still there is a possibility that changes in nova/neutron/cinder/keystone will break horizon 12:24:56 <doug-fish> I'm reading here: http://docs.openstack.org/developer/tempest/plugin.html (Didn't know this existed) 12:25:02 <tsufiev> something we might to consider in future 12:25:04 <david-lyle> tsufiev: only keystone 12:25:12 <david-lyle> we only test login 12:25:27 <david-lyle> and none of the clients off master, only released versions 12:25:34 <tsufiev> well, there was a recent change in cinder quotas that broke some of Horizon buttons 12:25:52 <david-lyle> tsufiev: that wasn't caught by our login scenario 12:26:00 <tsufiev> yes, that's true 12:26:25 <tsufiev> I meant integration tests breaking from changes in other openstack components 12:26:38 <david-lyle> for which projects our plugin gets loaded is a matter for later debate 12:26:50 <david-lyle> tsufiev: but they only run on horizon jobs 12:26:52 <doug-fish> do we have a plugin yet? 12:26:56 <david-lyle> they are not in tempest 12:27:03 <david-lyle> doug-fish: no, it's on my todo list 12:27:06 <doug-fish> :-0 12:27:09 <doug-fish> :-) 12:27:17 <doug-fish> got it 12:27:19 <david-lyle> maybe we can do it at the meetup on Friday 12:27:22 <tsufiev> david-lyle, should we make horizon integration tests be a tempest plugin then? 12:27:29 <david-lyle> then more than one person understands it 12:27:30 <david-lyle> :) 12:27:55 <david-lyle> tsufiev: perhaps 12:27:57 <doug-fish> tsufiev: maybe it's a collection of plugins, so we can run the Horizon Cinder tests with cinder, etc 12:28:02 <david-lyle> they need to be highly stable 12:28:15 <tsufiev> I'm working on that :) 12:28:26 <david-lyle> tsufiev: I'm well aware and grateful 12:28:30 <david-lyle> :D 12:28:33 <tsufiev> :D 12:28:41 <david-lyle> #topic https://blueprints.launchpad.net/horizon/+spec/update-jasmine 12:28:41 <doug-fish> david-lyle: it sounds like a good thing to do on Friday 12:28:54 <david-lyle> #link https://blueprints.launchpad.net/horizon/+spec/update-jasmine 12:28:56 <david-lyle> is this done? 12:29:34 <doug-fish> it doesn't look done to me 12:29:38 <david-lyle> requirements indicate we're using it 12:29:48 <doug-fish> the upgrade has happened, but no changes to the tests 12:30:14 <david-lyle> but if we're using it, what changes are required? 12:30:51 <doug-fish> I'm looking at the patches references in the blueprint https://review.openstack.org/#/c/135619/ and https://review.openstack.org/#/c/133175/ 12:31:00 <doug-fish> I don't have a full understanding here 12:31:04 <david-lyle> oh, I auto-abandoned those patches 12:31:18 <david-lyle> just saw they were abandoned 12:31:22 <doug-fish> sure - not sure that was wrong 12:31:25 <robcresswell> https://github.com/openstack/horizon/blob/master/package.json#L12 12:31:29 <robcresswell> Jasmine is on 2.2.0 12:31:38 <robcresswell> So the title at least, has been achieved 12:31:42 <doug-fish> :-) 12:31:49 <david-lyle> so perhaps there are optimizations this allows 12:32:06 <doug-fish> yeah, that's what I get from the 2nd sentence 12:32:56 <doug-fish> Maybe it should be marked done anyway? It's half done and I think no further work is planned 12:33:18 <david-lyle> another bp can take on the updating if still desired 12:33:26 <tsufiev> +1 12:34:19 <david-lyle> #info https://blueprints.launchpad.net/horizon/+spec/update-jasmine implemented 12:35:17 <david-lyle> #topic https://blueprints.launchpad.net/horizon/+spec/task-event-history 12:35:19 <david-lyle> #link https://blueprints.launchpad.net/horizon/+spec/task-event-history 12:36:14 <david-lyle> I don't believe there is API support for this 12:36:23 <doug-fish> agreed 12:36:27 <david-lyle> we have something for nova 12:36:37 <doug-fish> usage? 12:36:44 <david-lyle> no actions tab 12:36:51 <doug-fish> oh, yes, I see 12:37:10 <david-lyle> this bp is too broad and ill defined to be useful IMO 12:37:17 <doug-fish> agreed 12:38:33 <robcresswell> agreed 12:39:15 <david-lyle> #info https://blueprints.launchpad.net/horizon/+spec/task-event-history obsolete 12:39:20 <david-lyle> left comment 12:39:35 <doug-fish> oops, me too 12:39:51 <david-lyle> #topic https://blueprints.launchpad.net/horizon/+spec/session-token-improvement 12:41:11 <david-lyle> I'm all for this 12:42:06 <david-lyle> smaller session size is a high priority 12:42:45 <robcresswell> Thats a nice bp. 12:42:50 <doug-fish> yeah, this is what Lin shared at the last summit I think 12:43:05 <david-lyle> yeah, the formal write up part 12:44:03 <david-lyle> any negative votes? 12:44:29 <doug-fish> nope. +1 12:44:37 <david-lyle> #info https://blueprints.launchpad.net/horizon/+spec/session-token-improvement approved 12:44:38 <tsufiev> I suspect it may increase response time a bit (cache retrieval), but only a bit 12:45:15 <tsufiev> +1 12:46:51 <david-lyle> #topic https://blueprints.launchpad.net/horizon/+spec/select-all-checkbox-object-containers 12:46:55 <david-lyle> #link https://blueprints.launchpad.net/horizon/+spec/select-all-checkbox-object-containers 12:47:16 <david-lyle> I have no concerns with this other than the bp is 7 months old and no code 12:47:47 <david-lyle> it could almost be treated as a bug rather than a bp 12:48:32 <david-lyle> merely adding a select-all 12:48:40 * robcresswell mumbles something about small bugs have well defined bps, and very complex bps having no content. 12:48:42 <tsufiev> if pagination will be implemented for Object Storage (we have such plans for Mitaka), this feature won't have much usefulness 12:48:52 <robcresswell> This looks more like a bug to me 12:48:57 <tsufiev> because Select All works only within a page 12:49:01 <tsufiev> AFAIK 12:49:28 <david-lyle> right, but all the other paginated tables have select-all 12:49:44 <david-lyle> what I would really prefer is a rewrite of the entire view 12:49:52 <david-lyle> that would be a good bp 12:49:58 <tsufiev> yes, it'd be good in terms of UX consistency, I agree 12:50:06 * tsufiev makes a mental note 12:50:25 <david-lyle> I have a feeling that select-all is more complicated to implement than it's worth 12:51:25 <doug-fish> just to be clear - select-all means select-page, right? 12:51:35 <david-lyle> doug-fish: yes 12:51:43 <tsufiev> yes 12:51:53 <doug-fish> ok great 12:52:31 <david-lyle> now I'm lost 12:54:26 <david-lyle> #info https://blueprints.launchpad.net/horizon/+spec/select-all-checkbox-object-containers approved 12:54:54 <david-lyle> may need to culled later pending progress 12:56:35 <david-lyle> #topic https://blueprints.launchpad.net/horizon/+spec/openstack-dashboard-logs 12:56:39 <david-lyle> #link https://blueprints.launchpad.net/horizon/+spec/openstack-dashboard-logs 12:57:10 <david-lyle> there is no openstack path to get to this. I think culling this makes sense 12:57:24 <doug-fish> +1 12:57:28 <robcresswell> Agreed 12:57:39 <robcresswell> Also a year old, no code. 12:58:25 <tsufiev> +1. I would rather mention here some inconsistencies between different openstack clients inside apache/dashboard log, but that's a topic for Logging group 12:59:14 <david-lyle> #info https://blueprints.launchpad.net/horizon/+spec/openstack-dashboard-logs obsolete 12:59:19 <tsufiev> (when Horizon is slow and someone switches on total debug in logs, it's not very convenient to parse the results) 12:59:19 <david-lyle> I added a note 12:59:49 <david-lyle> for admin it would be possible to show the horizon server log 13:00:24 <david-lyle> but that would be a different bp 13:00:41 <doug-fish> plus clustering? 13:00:45 <doug-fish> how to do that? 13:01:05 <david-lyle> plus clustering? 13:01:15 <doug-fish> sorry, it's early to use all of the words 13:01:33 <tsufiev> I think Doug meant what if we have several Apache servers? 13:01:34 <doug-fish> logging in a more difficult problem if there are multiple clustered instances of Horizon running 13:01:45 <doug-fish> *is 13:01:48 <doug-fish> *sigh* 13:01:50 <doug-fish> nevermind. 13:02:01 <doug-fish> it's a different blueprint. 13:02:22 <david-lyle> so in production, I've found the hardest part was figuring out which server you were connected to to view the log 13:02:40 * tsufiev imagines looking through a centralized log being logged itself, thus resulting in a positive feedback :) 13:02:40 <david-lyle> in this case, you wouldn't have to make that determination 13:02:59 <david-lyle> as long as you're able to reproduce the error 13:03:10 <doug-fish> and as long as all of your requests go to the same server 13:03:26 <tsufiev> david-lyle, I usually tailf as many logs as I have Apache instances and see which changes :) 13:04:00 <david-lyle> if you are getting loadbalanced that often I think there's something else wrong 13:04:11 <david-lyle> we're at time 13:04:19 <david-lyle> Thanks everyone for your time. 13:04:23 <david-lyle> #endmeeting