12:00:40 <david-lyle> #startmeeting horizon 12:00:41 <openstack> Meeting started Wed Oct 14 12:00:40 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' 12:02:06 <robcresswell> o/ 12:02:12 <amotoki> hi 12:02:12 <r1chardj0n3s> o/ 12:02:23 <tsufiev> o/ 12:02:26 <masco> o/ 12:03:00 <kzaitsev_mb> o/ 12:03:20 <david-lyle> looks like we can get rolling 12:03:44 <david-lyle> First release topics 12:03:45 <mrunge> o/ 12:04:21 <david-lyle> RC-2 is finalized and we haven't found a bug/fix serious enough to roll an RC-3 12:04:48 <david-lyle> so tomorrow we'll release RC-2 as the liberty version of Horizon 12:05:05 <david-lyle> Great job everyone 12:05:14 <r1chardj0n3s> yay us :) 12:05:20 <robcresswell> \o/ 12:05:35 <robcresswell> I had a question, possibly release-note related 12:06:12 <amotoki> https://bugs.launchpad.net/horizon/+bug/1415712 12:06:12 <openstack> Launchpad bug 1415712 in OpenStack Dashboard (Horizon) ""Volumes" tab doesn't show up if cinder v1 isn't registered" [High,In progress] - Assigned to IWAMOTO Toshihiro (iwamoto) 12:06:14 <amotoki> robcresswell: this one? 12:06:20 <robcresswell> Yessir 12:06:22 <robcresswell> https://review.openstack.org/#/c/151081/ 12:07:26 <robcresswell> I think Cinder v1 was deprecated in Kilo? It seems deployments with both v1 and v2 behave fine, but v2-only will have issues. This will likely be more prevalent in Liberty deployments. 12:07:53 <david-lyle> so cinder v1 was deprecated, but that clock has been reset, I believe 12:08:06 <mrunge> it hits kilo deployments as well 12:08:21 <david-lyle> python-cinderclient didn't fully support v2 until mid liberty 12:08:27 <amotoki> it is a good candidate of backport, but I am not sure it is release blocking. I think we can backport it for 2015.2.1. 12:08:46 <mrunge> 8.0.0 you mean? 12:08:49 <robcresswell> Thats 8.1.0 now :p 12:09:10 <amotoki> mrunge: yes. 2015.2.1 is a version of collection. 12:09:17 <david-lyle> I think all of those should have checked for either not just v2 12:09:26 <david-lyle> there's no reason to bar v1 support 12:09:32 <mrunge> yupp 12:09:34 <amotoki> mrunge: 8.0.1 I think. 12:09:45 <mrunge> amotoki, yes 12:09:48 <david-lyle> it's not quite dead yet 12:09:59 <robcresswell> Yeah, that was my question in the comments 12:10:10 <robcresswell> Code-wise its fine, but I'm unsure of the deployment status 12:10:23 <david-lyle> additionally if deployers chose to use volume as the endpoint type for volumev2, it still works 12:10:41 <david-lyle> I'd like to see that fixed before we consider backporting 12:11:08 <mrunge> at least we should have a note in release notes, right? 12:11:50 <david-lyle> that we require v1 12:11:52 <david-lyle> ? 12:11:54 <robcresswell> If 'volume' can be either v1 or v2 12:11:59 <robcresswell> how do we differentiate 12:12:25 <mrunge> I guess we should describe, what we require 12:12:38 <mrunge> and that it can be different from cli 12:12:41 <robcresswell> Or do we only support v2, but it can be named 'volume' or 'volumev2' 12:12:54 <david-lyle> this is one of the reasons for the service catalog cross-project session 12:13:08 <david-lyle> creating a new endpoint type is just wrong 12:13:13 <amotoki> agree 12:13:30 <david-lyle> but we have to try to handle it as is 12:13:48 * david-lyle wonders if there is a version API 12:13:51 <david-lyle> for cinder 12:14:47 <david-lyle> this should be listed in a "Known Issues" section 12:14:57 <david-lyle> that we require cinder v1 12:15:41 <robcresswell> Yes, I'll update the notes 12:15:58 <david-lyle> thanks robcresswell 12:16:11 <robcresswell> Also here 12:16:16 <robcresswell> #link https://bugs.launchpad.net/horizon/+bug/1506036 12:16:16 <openstack> Launchpad bug 1506036 in OpenStack Dashboard (Horizon) "Horizon only supports a 'volumev2' endpoint" [Undecided,New] 12:16:23 <robcresswell> For further discussion etc. 12:16:52 <david-lyle> that patch hasn't merged yet BTW 12:17:04 <robcresswell> wait I'm muddling my service names in that bug too 12:17:10 <robcresswell> I think 12:17:51 <robcresswell> ignore me. 12:17:53 <david-lyle> I can pull out of gate 12:18:24 <david-lyle> objections amotoki robcresswell? 12:18:44 <amotoki> in my understanding, the patch itself is good. I don't think we need to pull it out. 12:18:45 <david-lyle> or do we feel this is an incremental improvement? 12:18:56 <robcresswell> Personally, I feel its a step in the right direction. 12:19:05 <robcresswell> We just need to argue abit more over 'volume' 12:19:10 <david-lyle> I think it's jumping the gun 12:19:15 <amotoki> it makes horizon works only with cinder v2 API as long as volumev2 is used. 12:19:41 <robcresswell> the problem atm is that you need volume and volumev2 I thought 12:19:48 <david-lyle> yo ducttape_ you support cinder v2? and what is the endpoint type? 12:19:50 <robcresswell> this patch makes it volumev2 only 12:20:08 <mrunge> yupp, we're using it that way 12:20:11 <ducttape_> I'd have to look if we have it, lemme check 12:20:25 <david-lyle> ducttape_: thanks 12:20:46 <amotoki> ducttape_: https://review.openstack.org/#/c/151081 is the review we discussed. 12:21:09 <robcresswell> Do we officially support v1? I thought we dropped that 12:21:20 <david-lyle> I think there were several problems with cinder v2 support across other projects and SDKs 12:21:31 <mrunge> so, puppet-modules currently deploy 2 endpoints for cinder, volume and volumev2 12:21:32 <david-lyle> which is why I believe the clock was reset on deprecation 12:22:16 <amotoki> there was a mailing thread http://lists.openstack.org/pipermail/openstack-dev/2015-October/thread.html#75934 12:22:45 <amotoki> it is marked as deprecated (for future removal) but cinder team seems have no plan to drop it soon. 12:23:04 <ducttape_> we have vol 1 and 2 deployed 12:23:34 <david-lyle> ducttape_: ok, good, thanks, we won't be breaking you today then, at least that way 12:23:44 <david-lyle> the day is young 12:23:47 <ducttape_> no, it will be in another way 12:23:52 <ducttape_> indeed 12:24:44 <ducttape_> why the rush to dump the old vol1 stuff? 12:24:57 * ducttape_ reads scroll back logs 12:25:02 <david-lyle> ducttape_: misuse of the term deprecated 12:25:46 <amotoki> it loosk similar of kyestone v2 deprecation :-( 12:25:56 <amotoki> *simiar to* 12:25:59 <david-lyle> yes 12:26:06 <mrunge> but keystone just uses a single endpoint 12:26:14 <david-lyle> which is better 12:26:15 <mrunge> and some magic inside client 12:26:28 <david-lyle> and provides a version API 12:26:41 <david-lyle> so we have a fighting chance 12:27:24 <tsufiev> but we still define keystone version to be used by Horizon in 2 places :) 12:27:36 <tsufiev> there is a room for improvement as well 12:27:48 * ducttape_ thinks if this is similar to keystone v2, then I'd expect everyone to still use v1 for the next 10 years 12:28:24 <tsufiev> I mean OPENSTACK_KEYSTONE_URL and OPENSTACK_API_VERSIONS 12:28:31 <david-lyle> tsufiev: preferred version, but we work with either 12:28:47 <david-lyle> and the OPENSTACK_KEYSTONE_URL doesn't need to have the version 12:29:14 <ducttape_> it just does, b/c thats the convention w keystone, right? 12:29:24 <tsufiev> david-lyle, I suspect that will fail w/o version prefix, but need to recheck that 12:29:31 <tsufiev> s/prefix/suffix/ 12:29:37 * david-lyle wonders where we are now and if we can see the main thread from here 12:29:48 <mrunge> tsufiev, there is some automagic in place 12:30:07 <tsufiev> david-lyle, sorry for leading the discussion away 12:30:09 <mrunge> iirc, it works in all ways, /me tried a few weeks back 12:30:21 <david-lyle> mrunge: yes 12:30:31 <david-lyle> we just haven't changed our default 12:30:38 <robcresswell> FYI, I updated the release notes in the wiki, and the in-tree patch. 12:30:45 <david-lyle> robcresswell: thanks 12:31:04 <ducttape_> I'd think the important part right now is to find out if nova runs ok with no v1 endpoint type. I would expect nova to have some strange bug if no v1 was present 12:31:06 * david-lyle still contemplating blocking the v1 removal patch 12:31:36 <david-lyle> I know enabling v2 as the default in devstack created all sorts of problems 12:32:05 <robcresswell> If we're unsure still, block it. It's not incredibly urgent is it? 12:32:19 <david-lyle> done 12:32:28 <tsufiev> mrunge, I recall I got 404 w/o version suffix, but that may have been in some older version, like Juno... 12:32:49 <tsufiev> anyways, will look again at master branch behavior 12:33:23 <david-lyle> so anyway RC-2 is finalized 12:33:34 <david-lyle> now all mitaka all the time 12:34:02 <robcresswell> awesome 12:34:36 <david-lyle> #topic finalizing summit sessions 12:34:41 <david-lyle> #link https://etherpad.openstack.org/p/horizon-mitaka-summit 12:35:20 <david-lyle> the topic list is at the bottom unless we have last minute changes, I will be posting those today or tomorrow 12:36:20 <david-lyle> additionally Ceilometer has invited us to a fishbowl session Wed 12:36:26 * david-lyle checking on time again 12:36:33 <r1chardj0n3s> "beer and pytest" is still there. I'm happy to extoll the virtues of pytest, of course, but I think it's not very sensible to consider replacing our current test suite ;) 12:37:04 <r1chardj0n3s> mind you, I can see keystone running over ;) 12:37:20 <doug-fish> I'm a bit uncertain of our plan for the Wed 5:30 work session 12:37:23 <david-lyle> wed 12:05 around the poor state of ceilometer view in horizon and what we can do about it 12:37:30 <tsufiev> r1chardj0n3s, will pytest work w/o beer :)? 12:37:45 <ducttape_> I'd like to see who is working on what panels. I see a lot of re-work on panels that are not very important / used much 12:37:58 <r1chardj0n3s> tsufiev: actually, it does, and unittest drives me to beer, so it's a win-win! 12:38:05 <david-lyle> hmm hadn't notice the pytest bit 12:38:21 <robcresswell> I'm interested in hearing more about which panels are used the most by various deployers :p 12:38:36 <tsufiev> r1chardj0n3s, ok, got it 0 12:38:45 <david-lyle> ducttape_: +1 12:38:56 <ducttape_> anything admin is less used, vs the project / _memeber_ panels 12:39:44 <david-lyle> and the improvements from a rewrite are at best negligible 12:39:48 <ducttape_> nova / networking / swift are used more.... then stuff like glance cinder etc are used less (within the project panels) 12:40:40 <robcresswell> Interesting 12:40:47 <david-lyle> it will be a good follow on to wed 5:30 and can segue into mitaka priorities? 12:40:58 <ducttape_> keystone / id is kind of sometimes used - but not day to day / all the time 12:42:02 <david-lyle> any pytest discussion would be better for Fri 12:42:40 <david-lyle> any other concerns? 12:43:37 <david-lyle> ok, doug-fish is on the hook for social arrangements 12:43:46 <david-lyle> IIRC 12:43:59 <doug-fish> my suggestion is that we meet here: http://www.princehotels.com/en/shinagawa/top-of-shinagawa 12:44:23 <doug-fish> at about 8pm, after the women of openstack thing 12:44:50 <amotoki> on monday? 12:44:56 <doug-fish> oh yes, on monday 12:44:58 <r1chardj0n3s> I might even make it, if you kick off at 8pm. my hotel's a 30 minute walk from there IIRC 12:45:26 <mrunge> I might be a bit late then 12:45:47 <amotoki> If it is full, perhaps i can help finding a place. 12:46:01 <doug-fish> amotoki: that would be outstanding! 12:46:04 <david-lyle> amotoki: any suggestions are welcome 12:46:07 <r1chardj0n3s> \o/ amotoki 12:46:17 <r1chardj0n3s> my hotel is more like 40 minutes from there :/ 12:46:21 <david-lyle> we're just choosing blindly at this point 12:46:35 <david-lyle> 40 minutes? staying at the airport? 12:46:42 <r1chardj0n3s> david-lyle: don't ask 12:46:47 <doug-fish> amotoki: do you think planning a different location is wise? 12:46:48 <amotoki> :) 12:46:52 * david-lyle really wants to ask 12:46:52 <robcresswell> :( 12:47:06 <r1chardj0n3s> david-lyle: Rackspace "organisation" strikes again 12:47:11 <david-lyle> ah 12:47:17 <amotoki> I think more of us will stay near Shinagawa, so it is a good choice. 12:47:26 <david-lyle> yes 12:47:27 <r1chardj0n3s> actually, I can metro to take ~25 minutes off that 12:47:32 <r1chardj0n3s> so that's good 12:47:48 <r1chardj0n3s> yes, definintely center on Shinagawa 12:48:02 <amotoki> Another option is to go Japanese bar (or more casual than usual bar) but it might be difficult to reach.. 12:49:54 <doug-fish> amotoki: more casual does seem suitable for our group 12:50:00 <doug-fish> do you mean that it's too far to walk? 12:50:01 <mrunge> amotoki, I think you'll have all our ears and eyes, meaning we'll probably trust you blindly ;-) 12:50:02 <amotoki> about how many folks are interested in the meetup? 12:50:12 * mrunge interested 12:50:13 <tsufiev> yeah, +1 for more casual 12:50:16 <doug-fish> \o 12:50:23 <david-lyle> o/ 12:50:27 <tsufiev> o/ 12:50:38 <robcresswell> o/ 12:50:42 <fnordahl> o/ 12:50:44 <r1chardj0n3s> I definitely arrive too late to make it to Shinagawa for drinks, so I'll catch you folks on the Tuesday 12:50:54 <amotoki> 10~15 people? 12:51:04 <fnordahl> (i know it's a bit rude to come into the meeting when discussing drinks, but would like to join if I may :-) ) 12:51:25 <david-lyle> fnordahl: all are welcome 12:51:25 <robcresswell> amotoki: Probably more like 15/20, this tz meeting tends to miss a lot of people 12:51:25 <doug-fish> fnordahl: of course! 12:51:39 <r1chardj0n3s> fnordahl: I dunno about you, but it's 10 to midnight here and I'm about to finish my whiskey ;) 12:51:40 <david-lyle> informal face to face before the summit 12:51:50 <fnordahl> r1chardj0n3s: lol 12:51:51 <amotoki> perhaps i can try to look for a place early next week or this weekend. 12:52:04 <david-lyle> let's add to the etherpad the tentative plan and amotoki please update with improved suggestion 12:52:12 <robcresswell> amotoki: That would be excellent! 12:52:18 <doug-fish> amotoki: thank you 12:52:31 <r1chardj0n3s> amotoki: places to converge on to have lunch might be appreciated, to save us wandering around aimlessly? 12:52:49 <r1chardj0n3s> I remember a lot of aimless wandering in Vancouver, which wasted valuable lunch time :/ 12:53:27 <david-lyle> #topic Open Discussion 12:53:40 <david-lyle> oh wait 12:53:43 <david-lyle> #topic 12:53:48 <david-lyle> argh 12:53:57 <amotoki> you can use #undo 12:54:10 <david-lyle> #undo 12:54:11 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0x9416250> 12:54:20 <david-lyle> #topic Bug Report 12:54:26 <david-lyle> #link https://wiki.openstack.org/wiki/Horizon/WeeklyBugReport 12:54:52 <robcresswell> Woop. So I dropped the searchlight panel from the bps, as all patches are either -1 Jenkins or merge conflicts -.- 12:55:12 <robcresswell> THe others bps have active patches (especially Images table, has about 8 in a chain, all passing) 12:55:29 <david-lyle> robcresswell: searchlight content is external for now 12:56:03 <robcresswell> As for the bugs, they are all carry overs from rc-potential. 12:56:16 <david-lyle> makes sense 12:56:20 <david-lyle> thanks robcresswell 12:56:27 <robcresswell> The two high bugs specifically, have patches that could be looked at. 12:56:37 <david-lyle> potential suggestion 12:56:42 * david-lyle pushes his luck 12:56:46 <robcresswell> ha, go ahead 12:57:02 <david-lyle> number of new bugs/new blueprints/closed bugs/closed bps ? 12:57:26 <robcresswell> As in totals week on week? 12:57:28 <david-lyle> the first two more than the latter two 12:57:40 <david-lyle> yeah, not imperative though 12:57:47 <david-lyle> just a thought 12:57:49 <robcresswell> Sure, I'll have a look at collating it 12:58:03 <david-lyle> not sure how easy launchpad would make such a thing 12:58:19 <david-lyle> guess not very 12:58:23 <robcresswell> As always, if you have chance to do a handful of reviews this week, please look at the bug report so we can focus on moving a few things forward. Thanks! 12:58:32 <david-lyle> thanks robcresswell 12:58:37 <mrunge> has launchpad an api? 12:58:42 <david-lyle> mrunge: yes 12:58:47 <robcresswell> mrunge: Yes, but not a useful one. 12:58:59 <mrunge> yay and boo! 12:59:41 <david-lyle> I think that brings us to time 13:00:03 <amotoki> in neutron, armando retrieves a bug list and convert it to Google spreadsheet for better filtiring. 13:00:12 <amotoki> it is not a best way thouth. 13:00:25 <david-lyle> we'll have the Horizon Driver's meeting later today to continue to sort through the blueprint backlog 13:00:31 <david-lyle> 2000UTC 13:00:38 <david-lyle> same channel for all interested 13:00:50 <david-lyle> amotoki: anything has to be better than launchpad 13:00:55 <david-lyle> :) 13:01:09 <robcresswell> I've put up a potential list for drivers meeting btw. 13:01:14 <david-lyle> thanks everyone 13:01:17 <david-lyle> #endmeeting