20:10:40 <robcresswell> #startmeeting horizon
20:10:41 <openstack> Meeting started Wed Jan 25 20:10:40 2017 UTC and is due to finish in 60 minutes.  The chair is robcresswell. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:10:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:10:45 <openstack> The meeting name has been set to 'horizon'
20:10:46 <robcresswell> :D
20:10:51 <robcresswell> Fire away rdopiera
20:11:57 <rdopiera> so the situation is, Nova went and changed that endpoint to always limit the number of entries it returns, but added pagination in microversion 2.40 so that we can get the rest of the entries
20:12:25 <rdopiera> that means we need to use microversion 2.40 of the api for that particular endpoint at least
20:12:37 <rdopiera> (if it's available, if it's not, there is nothing we can do)
20:13:26 <rdopiera> the initial patches I asked for an exception for just added that pagination, and a way for the user to set the nova api version to 2.40
20:13:42 <robcresswell> I haven't had time today to see if my microver patch works on new novaclient
20:13:48 <robcresswell> but your patch is fine
20:13:48 <rdopiera> but that is apparently risky, because we currently use version 2.1 for everything
20:13:59 <david-lyle> 2.40 for all API requests, or just that particular one?
20:14:09 <robcresswell> 2.40 for everything will be error prone
20:14:24 <david-lyle> that's why I'm asking for clarification
20:14:27 <ediardo> o/
20:14:30 <rdopiera> david-lyle: initially, my idea was to use it for all, since we didn't have microversion patch merged
20:14:50 <rdopiera> now I wrote a patch that only requests 2.40 for that one (actually two) endpoint
20:14:56 <david-lyle> and then if the client didn't support 2.40, no nova?
20:15:01 <rdopiera> using parts of robcresswell's microversion patch
20:15:03 <david-lyle> ok
20:15:08 <david-lyle> that would be better
20:15:20 <rdopiera> and of course if the nova doesn't support it, it uses whatever is the default
20:15:25 <rdopiera> and we get tuncated usages
20:15:32 <rdopiera> but there is nothing we can do about that
20:15:58 <rdopiera> the new patch is https://review.openstack.org/#/c/424585/
20:16:12 <rdopiera> I would love to hear any ideas or other feedback for this
20:16:32 <rdopiera> or suggestions about how we could approach this better
20:17:18 <rdopiera> one more note, since I can't catch the "wrong api version requested" error, because the novaclient catches it, I have to "look before jump" and check the available versions before doing the call
20:17:45 <robcresswell> So a double novaclient check?
20:17:47 <rdopiera> that creates a race condition if the nova version changes between the requests
20:18:02 <rdopiera> but I hope that's not a very common occurence
20:18:15 <rdopiera> and it would be fatal only if the version changes down
20:18:15 <robcresswell> Like, init novaclient, use novaclient to check version, re-init client
20:18:34 <rdopiera> yes, but that's memoized anyways
20:18:44 <robcresswell> I can't imagine many people are randomly downgrading clients during uptime
20:18:58 <robcresswell> Actually I take that back, people are crazy
20:19:00 <rdopiera> yeah, I just want to point out that it is there :)
20:20:13 <rdopiera> it could happen if you have nova behind a load balancer and only upgrade one
20:20:17 <david-lyle> patch seems reasonable
20:20:35 <david-lyle> the memoize only lasts the life of the request anyway
20:21:24 <rdopiera> but I think a lot of our logic would fail in that scenario
20:22:00 <robcresswell> I thought that the client memoize was meant to ignore some of the request parameters so it stayed cached
20:22:08 <robcresswell> Or something fancy
20:22:09 <rdopiera> yes
20:22:12 <rdopiera> it does
20:22:32 <rdopiera> but version is also one of the parameter for the memoization
20:22:42 <rdopiera> so the two versions get memorized separately
20:22:43 <robcresswell> Ah
20:22:49 <robcresswell> That was my next question :p
20:23:05 <robcresswell> I'm glad you are two steps ahead
20:23:59 <rdopiera> we will probably modify that code once the microversion patch lands
20:24:14 <rdopiera> this is supposed to be a stopgap
20:24:30 <robcresswell> Sure
20:24:34 <robcresswell> Fine with me
20:25:07 <rdopiera> oh, and I would be really grateful for a second +2 on that pagination patch, pretty please!
20:25:26 <rdopiera> https://review.openstack.org/#/c/410337
20:26:50 <david-lyle> honestly I think those two should be squashed
20:27:02 <david-lyle> but as long as both merge, I'm ok with it
20:27:09 <david-lyle> as is
20:27:27 <rdopiera> they depend on each other
20:27:35 <rdopiera> I mean, the second one on the first one
20:29:28 <robcresswell> Anything else to discuss?
20:29:55 <robcresswell> Oh PTL elections are this week if anyone else wants to run
20:30:12 <betherly> this cycle has gone super fast
20:30:22 <robcresswell> It was really short
20:30:36 <robcresswell> I think Richard picked this one on purpose
20:30:42 <betherly> :D
20:30:43 <david-lyle> reminder for those headed to the PTG: https://etherpad.openstack.org/p/horizon-ptg-pike
20:30:55 <robcresswell> Ah, yeah, thanks david-lyle
20:31:03 <robcresswell> Its just us two right?
20:31:13 <david-lyle> up to 7 now
20:31:16 <david-lyle> err 8
20:31:21 <betherly> thanks david-lyle!
20:31:24 <david-lyle> including richard
20:31:28 <betherly> and meeeee!
20:31:34 <betherly> woot
20:31:36 <david-lyle> I believe he confirmed
20:31:49 <david-lyle> I was including you betherly :)
20:32:09 <rdopiera> betherly: I just realized we never met
20:32:12 <robcresswell> Huh someone has added the lbaas dashboard to our discussion :p
20:32:15 <betherly> yay! just being hyper ignore me
20:32:37 <david-lyle> I added my favorite topic of removing code :)
20:32:49 <rdopiera> david-lyle: thank you for the +2s \o/
20:32:57 <david-lyle> rdopiera: thanks for working on it
20:32:58 <betherly> rdopiera: so weird right! the joys of international working :)
20:33:08 <rdopiera> robcresswell: should we discuss dropping nova support? ;)
20:33:19 <david-lyle> where better to meet than Atlanta, USA
20:33:22 <betherly> ha!!!
20:33:32 <rdopiera> david-lyle: I'm not coming, sadly
20:33:35 <david-lyle> doh
20:33:37 <betherly> david-lyle: sadly I don't think rdopiera is going
20:33:50 * rdopiera plays a really small violin
20:33:55 <betherly> hahaha
20:34:07 <david-lyle> :'(
20:34:13 <robcresswell> I'm glad we're meeting in the USA, it makes a change to all the other OpenStack events
20:34:14 <betherly> y'all make my wednesday evenings :'D
20:34:48 <betherly> robcresswell: sarcasm detector alerted me
20:34:54 <robcresswell> :)
20:34:59 <david-lyle> we recognize there are other parts of the world on occasion and then promptly forget
20:35:35 <betherly> :D
20:35:56 <robcresswell> :(
20:36:07 <david-lyle> robcresswell:  other than your panel split work
20:36:15 <david-lyle> are there any other FFE's in the wind
20:36:24 <robcresswell> Glorious work that it is
20:36:27 <robcresswell> K2K
20:36:35 <david-lyle> I consider that merged
20:36:39 <robcresswell> That vagrant setup doesn't like me
20:36:45 <david-lyle> git may disagree
20:36:52 <robcresswell> Lol
20:38:29 <david-lyle> I just merged the doa version bump
20:38:35 <david-lyle> so that's in place
20:38:47 <david-lyle> it's gating that is
20:38:58 <robcresswell> I'll try again in the morning
20:39:13 <robcresswell> My eyes are stinging from too much computers
20:39:18 <robcresswell> :D
20:39:34 <rdopiera> computers are the worst
20:41:11 <betherly> robcresswell: go rest your eyes and play a computer game or watch tv or something :p
20:41:40 <david-lyle> first end the meeting
20:41:49 <robcresswell> Yeah good idea
20:41:53 <robcresswell> Thanks everyone
20:42:00 <rdopiera> thanks, bye!
20:42:02 <betherly> Have a good one all
20:42:02 <robcresswell> #endmeeting