12:00:09 <alex_xu> #startmeeting nova api
12:00:10 <openstack> Meeting started Tue Feb 16 12:00:09 2016 UTC and is due to finish in 60 minutes.  The chair is alex_xu. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:00:12 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
12:00:14 <openstack> The meeting name has been set to 'nova_api'
12:00:28 <alex_xu> who is here today?
12:00:30 <jichen> o/
12:00:41 <gmann_> hi
12:01:10 <alex_xu> super quite recently
12:01:25 <alex_xu> I guess due to ops meetup?
12:01:55 <cdent> o/
12:02:21 <cdent> February is a dreary month in the norther hemisphere...
12:02:40 <alex_xu> For the concept doc, all the existed patch already merged
12:03:04 <alex_xu> cdent: ok...
12:03:18 <cdent> make people tired and quite :)
12:03:26 <alex_xu> cdent: :)
12:03:29 <gmann_> :)
12:04:09 <jichen> :)
12:04:14 <alex_xu> I remember johnthetubaguy already closed the concept doc bp, but I think people still can free feel to add more doc
12:04:29 <gmann_> alex_xu: +1, nice work
12:04:39 <alex_xu> #topic API futures - patches for approved specs
12:04:49 <alex_xu> any patches we want to discussion?
12:05:28 <alex_xu> For the allow server_id param as DB index
12:05:59 <alex_xu> agree with gmann_ and sdague should check ec2 first
12:06:27 <gmann_> alex_xu: yea at first glance i grep into ec2 repo, so there their use uuid
12:06:35 <alex_xu> if ec2 didn't depend on that, I think we can remove that totally and without microversion, as there isn't a way to get that 'id' from nova api
12:06:50 <sdague> o/
12:06:55 <alex_xu> gmann_: cool, thanks
12:06:55 <gmann_> but we need to completely check their usage from external API too
12:07:12 <sdague> it might be an some extended response
12:07:19 <sdague> I though that's what 2.4 was all about
12:07:40 <alex_xu> oops, I may only check the 2.1
12:07:59 <sdague> that being said, it's not as easy to remove as you think, the get_id assumes either numeric or uuid at a very early stage in the code
12:08:00 <gmann_> sdague: you mean 2.3
12:08:05 <sdague> gmann_: yeh, 2.3
12:08:20 <sdague> it was one of the first microversions
12:08:25 <gmann_> yea
12:10:23 <alex_xu> #link https://github.com/openstack/nova/blob/master/doc/api_samples/os-extended-server-attributes/v2.3/server-get-resp.json
12:10:35 <alex_xu> I didn't find out one field for that id
12:11:37 <alex_xu> sdague: what is 'get_id' point to?
12:12:19 <gmann_> i also remember we do not put that in response
12:12:23 <gmann_> #link https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/views/servers.py#L107
12:13:11 <alex_xu> but it still worth for checking to find out why ec2 needn't that for now
12:13:59 <gmann_> yea
12:14:07 <sdague> https://github.com/openstack/nova/blob/dbdf27021cc23bf8cec8d3f51fa03c6dac5da53a/nova/compute/api.py#L1939
12:14:28 <sdague> that's where pretty deep in the compute api we go a lot of different directions
12:15:30 <alex_xu> sdague: you mean there may have code still use compute_api.get with id in nova code?
12:15:46 <sdague> yes
12:16:13 <sdague> because this is pretty deep in the code, the fallout could be a pretty big deal to unwind
12:16:18 <alex_xu> ok, let's double check it
12:16:32 <sdague> I consider this a really low priority issue to address
12:16:39 <jichen> yeah, I did a quick check on this, network related functions used it
12:16:41 <alex_xu> yes
12:16:45 <jichen> +1
12:16:51 <alex_xu> it didn't hurt anything
12:16:54 <gmann_> sdague: yea, we can do later.+1
12:16:57 <jichen> https://bugs.launchpad.net/nova/+bug/1522536
12:16:58 <openstack> Launchpad bug 1522536 in OpenStack Compute (nova) "500 error returned from server GET by numeric id" [Undecided,Confirmed] - Assigned to Sriteja (sriteja-alapati)
12:17:03 <jichen> so this bug should be invalid ?Just saw it
12:17:11 <sdague> jichen: that's a little different
12:17:17 <sdague> because we're getting a 500
12:17:45 <jichen> sdague: yeah, actually I also got 500 when I did test
12:18:35 <sdague> I think that's just because we're not checking if the instance is really found
12:19:07 <gmann_> sdague: yea from cache
12:19:24 <jichen> ok
12:19:53 <sdague> anyway, that's a real bug to be fixed, and I think it's because something is not entirely safe in the db fetch
12:20:17 <alex_xu> I guess it due to we save the instance into the case with uuid, not the id?
12:20:19 <sdague> jichen: I'd start with a new test case in tree for that
12:20:48 <jichen> sdague: ok, will take a look ,thx
12:20:51 <sdague> no, I think it's an assumption in the code that get_by_id isn't going to fail on a context check
12:21:51 <alex_xu> ok
12:22:59 <alex_xu> ok, I think we are done this problem, anymore want to bring up?
12:23:37 <alex_xu> if no, let's jump to open
12:24:00 <alex_xu> #topic open
12:24:27 <alex_xu> let's talk about newton ideas?
12:24:50 <gmann_> alex_xu: can we discuss tempest microversion before that
12:25:22 <alex_xu> unfortunately johnthetubaguy and ken'ichi isn't here today
12:25:24 <alex_xu> gmann_: sure
12:25:34 <gmann_> sdague: as discussed, i submitted patch for using fixture for microversion setting
12:25:40 <gmann_> #link https://review.openstack.org/#/c/279430/
12:25:48 <gmann_> if you can have look
12:25:59 <sdague> gmann_: ok, cool. I'll look today
12:26:13 <sdague> mtreinish is proposing that tempest-lib spins back into tempest
12:26:13 <gmann_> sdague: Thanks.
12:26:22 <gmann_> sdague: yea
12:26:22 <sdague> that will make some of this easier
12:26:38 <gmann_> sdague: yea that will be really helpful for miceroversion testing
12:26:55 <gmann_> sdague: m also almost +2 on that patch
12:27:01 <alex_xu> gmann_: sdague, a question, which microversion we will test in the tempest?
12:27:31 <gmann_> alex_xu: All but with specific tests for that microversion
12:27:40 <sdague> right, what gmann_ said
12:27:46 <alex_xu> gmann_: ok, got it, thanks
12:28:03 <sdague> so for projects that need specific microversions tested we'll have dedicated tests for those (especially ones with behavior changes)
12:28:05 <gmann_> alex_xu: example v2,2 - https://github.com/openstack/tempest/blob/master/tempest/api/compute/keypairs/test_keypairs_v22.py
12:28:12 <alex_xu> we probably late on lastest version a lot
12:28:19 <gmann_> sdague: yea
12:29:11 <alex_xu> gmann_: cool
12:29:26 <gmann_> alex_xu: yea, at end we would like to run a experimental job running with latest to check any version tests missing etc.
12:31:07 <gmann_> alex_xu: that's all on this, we can move to next topic
12:31:59 <alex_xu> sorry for unstable network...
12:32:33 <alex_xu> sdague: do you want to talk about newtown idea today?
12:33:02 <alex_xu> or just wait johnthetubaguy and ken'ichi at here someday
12:33:14 <sdague> let's postpone until next week when they are back
12:33:20 <sdague> we're not really in a rush on it
12:33:26 <alex_xu> ok
12:33:38 <gmann_> yea, Ken'ichi shifted to USA last week
12:33:50 <alex_xu> gmann_: ah, that's cool
12:34:09 <alex_xu> I can't catch him in my working time anymore :(
12:34:20 <gmann_> :), me too
12:34:28 <alex_xu> so...no more from today, anyone have more bring up today?
12:34:44 <alex_xu> s/no more from/no more from me/...
12:35:00 <gmann_> nothing from my side
12:35:06 <alex_xu> 3...
12:35:06 <sdague> no, I'm good for now
12:35:13 <alex_xu> 0..
12:35:31 <alex_xu> ok, let's close the meeting early today, thanks all!
12:35:34 <jichen> thanks
12:35:36 <gmann_> Thanks all
12:35:40 <alex_xu> #endmeeting