12:00:07 <alex_xu> #startmeeting nova api
12:00:08 <openstack> Meeting started Tue Jan 19 12:00:07 2016 UTC and is due to finish in 60 minutes.  The chair is alex_xu. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:00:09 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
12:00:12 <openstack> The meeting name has been set to 'nova_api'
12:00:18 <alex_xu> who is here today?
12:00:21 <jichen> o/
12:00:24 <Kevin_Zheng> hi
12:00:53 <alex_xu> jichen: Kevin_Zheng hello
12:01:13 <alex_xu> let's wait one more minute for more people join
12:01:20 <Kevin_Zheng> sure
12:01:37 <gmann_> hi
12:01:52 <sdague> o/
12:02:27 <alex_xu> ok, cool, looks like have enough people now, let's start the meeting
12:02:34 <alex_xu> #topic content patches up for review
12:02:42 <alex_xu> #link https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/complete-todo-in-api-concept-doc,n,z
12:02:57 <alex_xu> looks good, just need more review
12:03:21 <alex_xu> I added one more about extension
12:03:35 <alex_xu> #link https://review.openstack.org/#/q/status:open+project:openstack/api-site+branch:master+topic:fix-compute-api-ref,n,z
12:03:51 <alex_xu> we have a lot of api ref fix merged last week, that thanks to jichen :)
12:04:09 <jichen> alex_xu: :)
12:04:24 <eliqiao> o/
12:04:27 <alex_xu> jichen: from your view, do we still have a lot of bugs in the doc?
12:04:30 <gmann_> nice, thanks jichen.
12:05:05 <jichen> alex_xu: I guess so, we might need more work on the wording and some consistency between code and doc
12:05:17 <alex_xu> jichen: ok, thanks
12:05:25 <jichen> np
12:05:26 <alex_xu> #link https://etherpad.openstack.org/p/nova-v2.1-api-doc
12:05:45 <alex_xu> from this etherpad looks like still have note didn't address yet for api ref
12:06:02 <alex_xu> #topic most needed next content patches
12:06:27 <alex_xu> for api concept, just left few concept in servers.
12:06:58 <alex_xu> hope we can at least have some initial doc for those
12:07:21 <alex_xu> if no more question, let's move on
12:07:40 <alex_xu> #topic remove project id
12:07:54 <alex_xu> sdague: looks like we have agreement on the solution last week
12:08:15 <alex_xu> #link https://review.openstack.org/233076
12:08:22 <sdague> yes, I responded to your review comments just now
12:08:34 <alex_xu> sdague: ^ this is ready to merge, right? I already begin to review it
12:08:54 <sdague> well, I have to rebase because another v2.16 merged
12:09:01 * gmann_ will review it tomorrow
12:09:14 <sdague> so it will be v2.17, it will take me a bit
12:09:35 <gmann_> sdague: should i hold all extension patches? not sure those will lead to conflict with this. need t check
12:10:01 <gmann_> #link https://review.openstack.org/#/q/topic:bp/api-sample-tests-with-all-extensions
12:10:08 <sdague> on https://review.openstack.org/#/c/233076/21 I added review comments
12:10:38 <sdague> gmann_: it would be nice if we could get this landed because we generate a bunch of additional comments
12:10:48 <sdague> sorry, conflict
12:10:52 <sdague> just waking up still :)
12:11:05 <gmann_> sdague: yea, sure. no issue.
12:11:25 <alex_xu> sdague: ok I got it
12:12:14 <alex_xu> I missed that point
12:12:45 <gmann_> sdague: alex_xu quick question (still need to look into patch)
12:12:59 <gmann_> we will not generate api samples for without project_id?
12:13:09 <sdague> gmann_: no, not yet
12:13:38 <sdague> auggy is working on some unit testing for the matcher in api_samples
12:13:51 <gmann_> sdague: ok. i see
12:14:04 <sdague> once she completes that, we can refactor that so we can have 1 template set, and 2 samples trees
12:14:14 <sdague> but there is so much magic in there, I didn't want to change it without tests
12:14:26 <gmann_> sdague: yea.
12:14:57 <alex_xu> sdague: do we still think about legacy v2 problem?
12:15:06 <gmann_> sdague: lot of magic logic will go away with all extension. and it will be more clear to refactor tests
12:15:16 <sdague> alex_xu: which problem is this?
12:16:00 <alex_xu> sdague: the project_id will optional for legacy v2 compatible mode also
12:16:19 <sdague> alex_xu: yes, that's true
12:16:33 <sdague> there are only so many things we can do with route changes like this
12:16:51 <sdague> the microversion here is really a signal for new allowed behavior
12:17:34 <alex_xu> anyway with project_id still work with legacy v2 mode
12:17:39 <sdague> right
12:17:51 <alex_xu> ok, it isn't a big problem
12:18:28 <alex_xu> ok, so we are good at here, any more question on this?
12:18:43 <alex_xu> so we just need help on another around review
12:19:09 <gmann_> +1
12:19:17 <alex_xu> cool, let's move on
12:19:24 <alex_xu> #topic API futures - patches for approved specs
12:19:34 <alex_xu> #link https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/api-sample-tests-with-all-extensions
12:19:42 <alex_xu> thanks gmann_ work on this
12:20:01 <gmann_> alex_xu: yea it was on hold for long time.
12:20:11 <alex_xu> I think we should metion this work on the meeting, otherwise I afraid we always forget to help on review those patches
12:20:21 <gmann_> yea
12:20:43 <alex_xu> so hope people help on review!
12:20:54 <gmann_> i will try to push more on this tomorrow but will check project_id patch should not conflict with this
12:21:02 <jichen> +1
12:21:07 <alex_xu> gmann_: thanks
12:21:28 <gmann_> looking at first glance it should not but will confirm and update accordingly
12:21:31 <gmann_> alex_xu: np
12:21:42 <alex_xu> I'm just thinking we should show all the fields in the sample file. that will be more useful in the api-ref
12:22:07 <alex_xu> but that may can be done separately
12:22:27 <gmann_> alex_xu: +1, once we get rid of extensions specific tests, i will go through all with schema files and add if anything mising
12:22:36 <alex_xu> gmann_: thanks a lot
12:22:54 <alex_xu> ok, so next one
12:22:58 <alex_xu> #link https://review.openstack.org/#/c/254950/
12:23:05 <alex_xu> I have question for this patch. I found rebuild action allow to change the instance name, if the name changed, the meaning of description may changed also. So I think user should have requirement update description when rebuild instance also.
12:23:06 <gmann_> alex_xu: i found some tests missing there for 2.3
12:23:24 <gmann_> only doc sample files are there but no template and tests. ll add that too
12:23:25 <alex_xu> gmann_: emm...interesting
12:23:35 <alex_xu> gmann_: thanks again :)
12:23:38 <gmann_> alex_xu: np
12:23:48 <sdague> alex_xu: is name change mandatory?
12:24:14 * alex_xu is checking
12:24:43 <alex_xu> sdague: no, it's not
12:24:47 <alex_xu> optional
12:25:07 <gmann_> sdague: i think no, that is optional same as update API
12:25:09 <gmann_> yea
12:25:17 <sdague> so description being optional is probably fine
12:25:17 <gmann_> only image is mandatory
12:25:41 <alex_xu> ok, probably need update spec also
12:25:53 <alex_xu> but that will be quickly, hope it still can catch up the freeze
12:26:03 <gmann_> sdague: alex_xu related to this, i feel update should show only updated field in response
12:26:33 <gmann_> update response is kinda replica of show server and keep adding more and more attribute with show API changes
12:26:33 <alex_xu> gmann_: currently it isn't only show updated field?
12:26:43 <gmann_> alex_xu: no it shows all the filed
12:27:05 <alex_xu> I'm not, but I think that should be another propose
12:27:22 <alex_xu> s/I'm not/I'm not sure/
12:27:23 <gmann_> alex_xu: yea another not in this
12:27:44 <sdague> gmann_: why do you only want to show updated fields?
12:28:03 <sdague> the point of UPDATE is that you update some number of things, and you get a whole resource back
12:28:11 <gmann_> alex_xu: with rebuild addition spec for that patch should add about adding the description field in update Response also
12:28:28 <alex_xu> gmann_: ok, got it
12:28:34 <gmann_> sdague: but with that it will be more clear what all being updated.
12:29:11 <gmann_> sdague: show resource should be showing all the info
12:29:37 <gmann_> i feel that but may be m missing any such use case of update
12:30:08 <alex_xu> is there any chance one field updated may lead to other field changed also? but for now, I guess we don't have this case.
12:30:13 <sdague> gmann_: the idea is we are working with a server representation
12:30:31 <sdague> that could be passed around
12:31:18 <sdague> I guess I don't understand what use you would have with partial resource returns
12:31:52 <gmann_> sdague: but in current case also update API does show all info as Show does
12:32:05 <gmann_> some attribute is in Show only
12:32:33 <alex_xu> before I hope we have some return for show and update
12:33:13 <alex_xu> I guess we should fix that in the future?
12:33:17 <gmann_> like https://github.com/openstack/tempest-lib/blob/master/tempest_lib/api_schema/response/compute/v2_1/servers.py#L159-L178
12:33:37 <gmann_> i feel either we should show all or only updated one
12:33:55 <sdague> right, I think we should be returning everything, or nothing honestly.
12:34:28 <gmann_> sdague: yea
12:34:53 <alex_xu> +1
12:35:16 <alex_xu> let find some point to fix that in the future
12:35:27 <gmann_> alex_xu: yea +1
12:35:40 <sdague> gmann_: it would probably be good to write up the issue, maybe as a spec, and where it shows up
12:35:54 <alex_xu> sdague: +1
12:35:59 <gmann_> sdague: sure, will add spec
12:36:27 <alex_xu> so let's move to next one
12:36:31 <alex_xu> #link https://review.openstack.org/#/c/128940/106
12:36:44 <alex_xu> For this patch, I just think we should give some help. As it already have 106 patchsets....
12:37:01 <alex_xu> API layer patch looks like pretty closes
12:38:05 <alex_xu> anyway people free to help review it
12:38:29 <Kevin_Zheng> OK
12:38:30 <gmann_> alex_xu: is author active or need help on implementation ?
12:38:46 <gmann_> alex_xu: i can help if needed
12:38:55 <alex_xu> gmann_: author is active, implementation looks good
12:39:12 <gmann_> alex_xu: ok, will review then. Thanks
12:39:23 <alex_xu> in my memory, this patch already been here 2 or 3 release. I guess due to wait too long time, people begin to forget it....
12:39:29 <alex_xu> gmann_: thanks
12:40:04 <alex_xu> good to see we have a helpful team
12:40:17 <alex_xu> #topic open
12:40:18 <alex_xu> any open today?
12:40:23 <Kevin_Zheng> yes
12:40:34 <alex_xu> Kevin_Zheng: please go ahead
12:40:47 <Kevin_Zheng> I have a question like this
12:40:56 <Kevin_Zheng> I have a microversion
12:40:59 <Kevin_Zheng> say 2.18
12:41:14 <Kevin_Zheng> it added vm_state and task_state
12:42:07 <Kevin_Zheng> John and Sean suggested that we shouldn't let list and show API in lower microversion to show those states
12:42:39 <Kevin_Zheng> but to doing this, I may have to add microversion for list and show API also?
12:42:54 <alex_xu> yes, I guess so
12:43:13 <Kevin_Zheng> also in 2.18?
12:43:24 <sdague> Kevin_Zheng: yes, this should all be in the same bump
12:43:28 <gmann_> i think we have those in show API right
12:43:40 <Kevin_Zheng> hm...
12:43:50 <sdague> basically when < 2.18 those values should show up to something that could exist earlier
12:44:01 <Kevin_Zheng> OK then
12:44:11 <Kevin_Zheng> I will work on that
12:44:41 <alex_xu> cool~
12:44:42 <sdague> alex_xu: https://review.openstack.org/233076 is rebased now, if you could take a look before you stop your day it would be appreciated.
12:44:56 <alex_xu> sdague: yea, I will check after the meeting
12:45:01 <jichen> hi , thanks for talking https://review.openstack.org/#/c/257261/ , last week ;we had a discussion, as follow up, I can't find use case for mac address, but I think the 'ip' type should be something helpful to user?
12:46:37 <alex_xu> is there any other way to get the type? for neutron, I guess probably user need check the network type
12:47:44 <alex_xu> jichen: maybe we can update the usecase in the spec, then let people feedback on the spec?
12:48:11 <jichen> um.. I can do that, let me update the spec and invite folks to take a look
12:48:25 <alex_xu> jichen: cool, thanks
12:48:34 <jichen> thanks
12:48:50 <alex_xu> if no more open, I will close the meeting early
12:49:13 <alex_xu> 3...
12:49:17 <alex_xu> 2..
12:49:19 <alex_xu> 1.
12:49:24 <alex_xu> thanks all!
12:49:31 <gmann_> Thanks all
12:49:32 <alex_xu> #endmeeting