19:01:58 <dtroyer_zz> #startmeeting openstackclient
19:01:59 <openstack> Meeting started Thu Sep 15 19:01:58 2016 UTC and is due to finish in 60 minutes.  The chair is dtroyer_zz. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:02:00 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:02:03 <openstack> The meeting name has been set to 'openstackclient'
19:02:15 <rtheis> hi
19:02:41 <stevemar> o/
19:02:44 <stevemar> howdy
19:03:41 <dtroyer_zz> #topic release
19:04:00 <dtroyer_zz> so last week we proposed a 3.3.0 release, which is being held until g-r is branched
19:04:09 <dtroyer_zz> and a lot has merged since then (yay!)
19:04:32 <dtroyer_zz> so I was thinking of updating the release hash this afternoon to include most (all?) of the new bits
19:04:34 <dtroyer_zz> thoughts?
19:04:41 <stevemar> dtroyer_zz: the g-r branch should be soon, maybe late sept at the latest
19:04:45 <stevemar> dtroyer_zz: definitely
19:04:56 <rtheis> sounds good to me
19:05:20 <dtroyer_zz> stevemar: I was under the (possibly mistaken) impression that would happen once all of the RCs are cut, like this weekend-ish
19:05:21 <stevemar> a lot of good feature work landed
19:05:30 <stevemar> dtroyer_zz: oh, entirely possible
19:05:43 <stevemar> i heard later
19:06:00 <stevemar> ¯\_(ツ)_/¯
19:06:06 <dtroyer_zz> if it is much later, we can update it again
19:06:58 <stevemar> yarp
19:07:08 <dtroyer_zz> ok, so beyond that, anything else release-specific?
19:07:38 <stevemar> nope
19:07:49 <rtheis> nothing else
19:07:52 <dtroyer_zz> ok…
19:07:57 <dtroyer_zz> #topic reviews
19:08:10 <dtroyer_zz> I finally have had some time to start working through my backlog
19:08:19 <dtroyer_zz> had a question or two
19:08:50 <dtroyer_zz> #link https://review.openstack.org/365910
19:09:04 <dtroyer_zz> any idea what the rationale for that one is?
19:09:28 <stevemar> nope, i've been avoiding it hehe
19:10:15 <dtroyer_zz> ok, I've left it low on my list, was just curious if there was something I had not seen
19:10:28 <rtheis> I haven't looked at it
19:10:55 <stevemar> maybe he thinks "image project" is a resource
19:11:18 <stevemar> and "add" should always be last?
19:11:22 <stevemar> since it's the verb
19:11:34 <stevemar> but for add we always (?) use it as a transition
19:11:53 <dtroyer_zz> my problem is I started heading in this direction with the mapping stuff in one of the network port reviews
19:12:07 <stevemar> object2 is added to object1  <object1> add <object2>
19:12:15 <dtroyer_zz> and now I need to figure out why these are different things
19:13:04 <dtroyer_zz> stevemar: right.   I thought project was an attribute attached to an image, not a separate resource that had a one-to-many relationship with an image
19:13:54 <stevemar> dtroyer_zz: i couldn't tell you confidently
19:14:56 <dtroyer_zz> ok.  this can wait a bit then
19:15:40 <dtroyer_zz> the other one I'm pushing back on a bit is https://review.openstack.org/#/c/353931/
19:15:47 <stevemar> dtroyer_zz: yeah, i'm in no hurry to merge it, but we should let the author know if hes going in the wrong direction
19:16:10 <dtroyer_zz> stevemar: will do
19:16:31 <stevemar> ohh that one, huanxuan and i had a good discussion on this earlier, see comments in PS3
19:17:17 <dtroyer_zz> I did, and it still seems a bit clunky
19:17:20 <dtroyer_zz> but maybe that is just me
19:17:56 <stevemar> dtroyer_zz: i like your idea of using --force to overwrite
19:18:32 <stevemar> it's the inverse of what i was suggesting with --new-volume, and it's cleaner
19:18:33 <dtroyer_zz> that was one of the tings I want to talk about in general, quickly here, and probably in our FB session
19:19:02 <dtroyer_zz> do we do that anywhere else?  I think I've seen it once, and that was due to the API requiring it, otherwise I don' tthink we have any danger flags
19:19:41 <stevemar> theres the usual "force" option with volume deletes
19:19:52 <stevemar> but no, i think this is new
19:20:52 <stevemar> dtroyer_zz: thought...
19:21:02 <stevemar> dtroyer_zz: can't multiple volumes have the same name?
19:21:28 <stevemar> so even if --force is used, then there's a chance that we'll overwrite the wrong one?
19:21:45 <dtroyer_zz> so, uh, yeah, that's an issue I didn't catch
19:21:56 <stevemar> or the cinder API will return something that is not a conflict
19:22:00 <dtroyer_zz> what does create do about that now?
19:22:15 <stevemar> i think it just creates a new one with a different ID
19:23:54 <dtroyer_zz> to a user, I am thinking this is either a 'replace exiting' or a 'create new', and the two need to be clear
19:24:05 <dtroyer_zz> so yeah, —new-volume helps in that regard
19:24:38 <stevemar> the issue is with 'replace existing' -- they can supply a name and many can be returned
19:24:59 <stevemar> or the first find, i forget how the API acts
19:26:02 <dtroyer_zz> this is one area we should start exposing the same behaviour to users…
19:26:14 <dtroyer_zz> and ambiguity is not our friend ;)
19:27:38 <dtroyer_zz> that was my short list, any others?
19:28:09 <stevemar> umm
19:28:25 <stevemar> oh the manage commands
19:28:30 <stevemar> for volume stuff
19:28:36 <stevemar> those are... weird as heck
19:29:01 <dtroyer_zz> yeah, I don't want manage/unmanage, that's almost as bad as 'action'
19:29:58 <dtroyer_zz> as much of that needs to be mapped into standard verbs as possible…
19:30:19 <dtroyer_zz> in https://review.openstack.org/#/c/363574/ I'm pushing back on import/unimport a bit, not sure that is right, but a good example of what we need to do
19:30:47 <stevemar> dtroyer_zz: the trouble with pulling that all into 'create' is that we already use --source too :)
19:30:58 <stevemar> to copy from an existing volume
19:31:28 <dtroyer_zz> that's a detail, then —backup-source or something
19:32:24 <stevemar> i think i had one more
19:32:45 <stevemar> meh
19:32:47 <stevemar> no matter
19:33:29 <stevemar> open discussion?
19:35:02 <dtroyer_zz> sorry, yes
19:35:06 <dtroyer_zz> #topic open discussion
19:35:38 <stevemar> i was gonna tackle this one if i have the time: https://blueprints.launchpad.net/python-openstackclient/+spec/compute-action-list
19:36:00 <stevemar> but it'll depend on https://review.openstack.org/#/c/280726/14
19:37:41 <dtroyer_zz> interesting… the term 'action' scared me for a bit as it is also ised badly in Nova
19:38:49 <stevemar> `server list actions``
19:38:52 <stevemar> i dunno
19:40:23 <dtroyer_zz> so I haven't looked at purge yet
19:40:37 <stevemar> eh, take your time
19:40:38 <dtroyer_zz> but I do think I have an outline of how we'll make that, and other commands, extensible to plugins
19:41:01 <stevemar> spread the knowledge :P
19:42:20 <dtroyer_zz> simple: plugins will register callbacks that get called when necessary
19:42:41 <dtroyer_zz> and as usual, the majority of the setup only happens once the command is being processed in take_action()
19:44:01 <stevemar> makes sense
19:44:26 <dtroyer_zz> I've been wanting to refine the plugin interface anyway, so this will be v3 and we'll roll it out slowly
19:44:40 <dtroyer_zz> and be a nice carrot for some projects that really want hooks, like in quotas and purge
19:44:56 <stevemar> i didn't realize how elaborate the purge command got
19:46:42 <stevemar> they offer dry runs, keep the project, delete the project
19:46:48 <stevemar> works with ceilometer, heat
19:46:50 <stevemar> snazzy stuff
19:46:54 <stevemar> a boatload of network stuff
19:51:13 <stevemar> dtroyer_zz: there's also https://review.openstack.org/#/c/370574/2
19:51:27 <stevemar> but i pulled it down, and couldn't get it to work
19:51:35 <dtroyer_zz> I'm sorry, keep getting interrupted
19:51:51 <stevemar> dtroyer_zz: oh its fine
19:51:53 <dtroyer_zz> uh, extension?
19:53:43 <stevemar> extension?
19:53:52 <dtroyer_zz> ya, why is there an optional thing?
19:54:06 <dtroyer_zz> how do we detect if it is there?
19:55:50 <stevemar> dtroyer_zz: oh damn, i sent you the wrong link
19:56:07 <stevemar> dtroyer_zz: dammit
19:56:12 <stevemar> dtroyer_zz: here's the right one: https://review.openstack.org/#/c/368768/
19:56:34 <stevemar> i guess the extension thing is still a question
19:56:52 <dtroyer_zz> oooohhhh, I hadn't seen that yet
19:57:56 <dtroyer_zz> I need to study that a bit
19:58:00 <stevemar> thats the one i couldn't get to work
19:58:09 <stevemar> but i assumed i was doing something wrong
19:58:46 <stevemar> oh this one just came in hot: https://review.openstack.org/#/c/371081/1
19:58:58 <dtroyer_zz> best part, though, is that bit all goes away soon
19:59:58 <dtroyer_zz> yay, more work!
20:00:24 <stevemar> we should scan for places that called auth_ref -_-
20:00:37 <stevemar> anywho, good meeting
20:00:43 <stevemar> lots of reviewing to do
20:00:45 <dtroyer_zz> ya, thanks guys
20:00:49 <dtroyer_zz> #endmeeting