13:00:01 <alex_xu> #startmeeting nova api
13:00:01 <openstack> Meeting started Wed Aug 30 13:00:01 2017 UTC and is due to finish in 60 minutes.  The chair is alex_xu. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:04 <openstack> The meeting name has been set to 'nova_api'
13:00:08 <alex_xu> who is here today?
13:00:19 * stephenfin is lurking
13:00:26 <gmann_> o/
13:00:38 <tetsuro> here
13:01:00 <alex_xu> ok, we can start the meeting
13:01:37 <alex_xu> but I didn't have too much update, since I spend most of time for the traits related spec and some internal training
13:01:49 <alex_xu> gmann_: do you have any update?
13:02:09 <gmann_> alex_xu: not much
13:02:32 <gmann_> wan to bring api change to know we want to bump version or not
13:02:43 <alex_xu> gmann_: yea, I saw that one
13:02:55 <gmann_> #link https://review.openstack.org/#/c/490722/
13:02:58 <gmann_> ah you replied
13:03:03 <gmann_> matt is not here to discuss
13:03:09 <alex_xu> gmann_: yea, I agree with you
13:03:30 <alex_xu> it needs a microversion if we change the schema of aggregate API
13:03:55 <alex_xu> but I think we fix the availability_zone attribute in the server create API
13:03:57 <gmann_> alex_xu: just read your comment, i was thinking the same that somehow we can interpret the : things and return 400 from there
13:04:13 <gmann_> and document ':' properly for aggregate that it is risky
13:04:35 <gmann_> alex_xu: yea, it is restricting the API and can be interop issue
13:05:00 <alex_xu> gmann_: annther way is that we can make the ':' works
13:05:26 <alex_xu> we can split the value by ':' from the right side
13:05:39 <gmann_> you means it:isazname:name:example
13:05:53 <gmann_> and start from example and 3rd : is all az name
13:06:31 <alex_xu> yes
13:06:42 <gmann_> yea that also ok
13:06:48 <gmann_> stephenfin: what you say?
13:07:54 <alex_xu> for example, 'ZONE:A:HOST1:NODE1'.rsplit(':', 2), you will get ['ZONE:A', 'HOST1', 'NODE1']
13:08:39 <gmann_> i see
13:08:45 <alex_xu> and fix that without microversion
13:08:50 <gmann_> and with that sch hint will be able to use AZ right
13:09:29 <alex_xu> gmann_: you mean the availability_zone attributes in the server create API? that isn't a sch hint
13:09:59 <gmann_> alex_xu: i mean for finding the right host
13:10:15 <alex_xu> yea, it will works
13:10:49 <alex_xu> s/works/work/
13:10:58 <gmann_> i do not know who used ':' in AZ. i even did not know this is allowed.
13:11:29 <gmann_> anyways m ok with your solution
13:11:29 <alex_xu> yea, I guess it is very corner case
13:11:55 <tetsuro> i can ammend the patch with the way you said.
13:12:19 <tetsuro> my collegue hit the bug somehow
13:12:22 <alex_xu> with the rsplit, it won't change the server create API I think (I didn't check all the corner case)
13:12:33 <alex_xu> tetsuro: thanks
13:12:41 <gmann_> because  as mriedem mentioned in patch, if do microversion it will be difficult to fix or correct the problem completely
13:12:54 <gmann_> tetsuro: i see
13:13:25 <gmann_> alex_xu: change the server API you mean functionality wise ?
13:13:45 <alex_xu> gmann_: mriedem's concern is in the old microversion?
13:13:58 <gmann_> alex_xu: yea
13:14:01 <alex_xu> gmann_: I mean the API behaviour
13:14:11 <gmann_> humm
13:14:46 <gmann_> alex_xu: if someone badly mentioned 4 param there az:host:node:mistake
13:15:05 <gmann_> in that case we will be having 'az:host' as az name
13:15:41 <alex_xu> gmann_: yea, but that case I checked :)
13:15:44 <alex_xu> #link https://github.com/openstack/nova/blob/master/nova/compute/api.py#L1611
13:15:57 <alex_xu> gmann_: we have a check for the valid az
13:16:09 <alex_xu> so the user get a 400
13:16:17 <alex_xu> as before
13:16:33 <alex_xu> s/as/same as/
13:16:40 <stephenfin> Sorry gmann, I'm in three meetings 🙈
13:16:47 <gmann_> alex_xu: nice.
13:16:57 <gmann_> stephenfin: this is 4th one or 3rd :)
13:17:06 <stephenfin> 3rd :D
13:17:08 <alex_xu> stephenfin: three meetings, that will be a top record
13:17:16 <stephenfin> thankfully
13:17:20 <gmann_> :)
13:17:26 <stephenfin> alex_xu: Brings me back to my Intel days :D
13:17:34 <alex_xu> lol
13:18:16 <alex_xu> gmann_: it is also good to check if there is any other corner case
13:18:39 <alex_xu> probably I can check tomorrow
13:18:49 <gmann_> alex_xu: i am ok with that solution. let's get stephenfin and mriedem feedback also on review.
13:18:58 <gmann_> alex_xu: thanks. that will be nice
13:19:05 <alex_xu> gmann_: cool, thanks
13:19:11 <tetsuro> that check on the link looks good to me, but how about the case that a host name or node name contains : ??
13:19:24 <gmann_> we can just check API is able to get az with ':' in name without any issue
13:19:42 <alex_xu> tetsuro: ':' isn't a valid char in the hostname
13:20:05 <tetsuro> okay, got it, thanks
13:20:15 <alex_xu> I guess also for nodename?
13:21:10 <alex_xu> tetsuro: thanks, remember to verify all the corner case also
13:21:32 <alex_xu> gmann_: back to the topic about queens release
13:21:45 <tetsuro> i'll check it out.
13:22:04 <gmann_> tetsuro: thanks
13:22:19 <alex_xu> gmann_: I guess we probably just work on the no-more-extension, hide-address, API deprecation
13:22:40 <alex_xu> for the concept doc, I can try that in the idle time
13:22:54 <gmann_> alex_xu: yea and our api-ref and api guide which are default for every cycle
13:23:04 <gmann_> alex_xu: yea, me too ll try.
13:23:31 <alex_xu> gmann_: for the microversion tests, we should continue with the idea, and some poc
13:24:02 <gmann_> alex_xu: the schema one or the on i have BP?
13:24:09 <gmann_> s/on/one
13:24:32 <alex_xu> gmann_: I mean the test on the lastest microversion, you already have a BP?
13:24:51 <gmann_> alex_xu: ok.
13:25:07 <gmann_> alex_xu: i had other one which i did not progress actually. top bottom one
13:25:19 <gmann_> alex_xu: let's see if i can push some patches before PTG
13:25:34 <alex_xu> gmann_: thanks
13:26:17 <alex_xu> anything else we can discuss, otherwise, let us end the meeting, and go to bed...
13:26:47 <gmann_> alex_xu: i am fine. we can close may be
13:27:09 <alex_xu> ok, thanks all!
13:27:11 <alex_xu> #endmeeting