19:01:19 <dtroyer_zz> #startmeeting OpenStackClient
19:01:20 <openstack> Meeting started Thu Jul 16 19:01:19 2015 UTC and is due to finish in 60 minutes.  The chair is dtroyer_zz. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:21 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:01:23 <openstack> The meeting name has been set to 'openstackclient'
19:01:26 <dtroyer_zz> Who is here today?
19:01:32 <stevemar> dtroyer_zz: half here
19:01:35 <terrylhowe> o/
19:01:37 * shaleh raises hand
19:01:53 <dtroyer_zz> stevemar: I hope it is the pretty half
19:01:59 <stevemar> dtroyer_zz: bknudson is around too, and pushing patches
19:02:16 <dhellmann> o/
19:02:17 <bknudson> hi
19:02:28 <dtroyer_zz> ping: lhcheng, sigmavirus24
19:02:33 <bknudson> keystone meetup!
19:02:44 <dtroyer_zz> Oh!  Say hi to the gang!
19:02:44 <bknudson> conversation here is uninteresting
19:03:01 <dstanek> bknudson: nice
19:03:09 <stevemar> bknudson: yeah, little bit
19:03:09 <dtroyer_zz> #topic reviews
19:03:43 <dtroyer_zz> I suppose the starting point might be WTF is going on?  I haven't had a chance yet to dig into the test failures, has anyone else?
19:04:00 <terrylhowe> no
19:04:07 <bknudson> I tried to use devstack to install and it didn't work.
19:04:10 * sigmavirus24 didn't know there was a failure =P
19:04:20 <bknudson> I think due to https://bugs.launchpad.net/python-openstackclient/+bug/1475127
19:04:20 <openstack> Launchpad bug 1475127 in python-openstackclient "Cannot manipulate group member by ID" [High,In progress] - Assigned to Steve Martinelli (stevemar)
19:05:33 <stevemar> yeah, the recent devstack patches to use keystone v3 only kinda murked things up
19:06:01 <stevemar> dtroyer_zz: i brought this up yesterday and you thought i was crazy!
19:06:06 <dtroyer_zz> that's what I was wondering.  did it to this across the board?
19:06:21 <dtroyer_zz> is this what you were asking?  I couldn't duplicate it here
19:06:42 <stevemar> you'll only hit it if its osc master + devstack master
19:06:47 <bknudson> it's passing domain_id=None to groups.get(), but that function only takes a group parameter
19:06:49 <stevemar> so CI is fine
19:07:00 <stevemar> (well, not _our_ CI)
19:07:02 <dtroyer_zz> oh, right
19:07:23 <bknudson> http://git.openstack.org/cgit/openstack/python-keystoneclient/tree/keystoneclient/v3/groups.py#n81
19:07:23 <dtroyer_zz> then I suppose that's the priority for this afternoon…
19:07:28 <bknudson> that's the groups.get() call
19:08:04 <bknudson> http://git.openstack.org/cgit/openstack/python-openstackclient/tree/openstackclient/common/utils.py#n29
19:08:14 <dtroyer_zz> so the dueling reviews, face-off in 10 minutes?  ;)
19:08:20 <bknudson> there's find_resource, which would call manager.get(name_or_id, **kwargs)
19:08:57 <stevemar> dtroyer_zz: basically, the find_resource call (or the refactoring we did a while ago, find_group_resource) changed the way .get() and .find() are called with UUIDs
19:09:02 <bknudson> so if kwargs has domain_id=None, it calls get('uuid-jdsfkdsa', domain_id=None)
19:09:05 <stevemar> it ended up doing this: GET /v3/users?name=f84f7276f2ae4e068869d9e5afd433ed
19:09:09 <stevemar> which is silly pants
19:09:32 <stevemar> dtroyer_zz: both fix it, just at different layers
19:09:42 <stevemar> dtroyer_zz: you can decide which fix you like better :P
19:10:03 <bknudson> find_resource is considering any exception from manager.get(name_or_id, **kwargs)  to mean "resource not found"
19:10:04 <dtroyer_zz> ok, at a glance I don't know yet…will dig in after the meeting
19:10:07 <bknudson> which hides the actual error
19:10:32 <dtroyer_zz> that sounds bad, bknudson is in the lead now ;)
19:10:32 <bknudson> the actual error was that .get() was called with an invalid parameter
19:10:39 <stevemar> yep
19:10:47 <bknudson> I didn't try to fix that
19:10:55 <stevemar> bknudson: slacker
19:11:07 <dtroyer_zz> so then I still have to actually read both reviews?
19:11:35 <dtroyer_zz> what other reviews need attention here?
19:11:44 <stevemar> dtroyer_zz: it'll be cool, i already approved mine, to get the gate unstuck (and i did this early in the morning)
19:11:49 <bknudson> it's interesting that the manager.get(int(name_or_id), **kwargs) case has special exception handler but manager.get(name_or_id, **kwargs) just skips
19:12:05 <stevemar> bknudson: you're more than welcome to fix it :D
19:12:12 <bknudson> maybe the exception handlers should be the same. I'll propose a patch!
19:12:56 <stevemar> dtroyer_zz: so, a cool review that was recently merged
19:13:01 <topol> bknudson, sounds like an easy one to review!  add me as a reviewer!!!
19:13:05 <stevemar> thanks to shaleh: https://review.openstack.org/#/c/201258/
19:13:24 <stevemar> now --help | -h works
19:13:35 <shaleh> I like good UX :-)
19:13:52 <shaleh> I have some more ideas to propose once I get 'image create' working properly
19:13:54 <dtroyer_zz> yes!  thanks shaleh.  and it was a cliff fix
19:14:32 <stevemar> dtroyer_zz: logistically, i kept the bug attached to OSC too, so we remember to release a new cliff before we release OSC
19:14:44 <stevemar> dtroyer_zz: another recent cliff change: https://review.openstack.org/#/c/202053/
19:14:50 <stevemar> moar UX!
19:15:11 <shaleh> I was actually thinking of writing that patch when I saw the review land
19:15:24 <stevemar> if you type $ openstack user foo
19:15:35 <topol> stevemar Need moar UX!!! keep going!!!
19:15:41 <stevemar> it'll spit out other `openstack user` commands instead of non-sense
19:15:41 <dtroyer_zz> ah, that's where that came from!  I saw it when testing the help bit
19:16:07 <stevemar> dtroyer_zz: terrylhowe and i were waiting for you and/or dhellmann to be cool with it
19:16:45 <dtroyer_zz> I didn't realize it was sitting there… conceptually I am, will look after the other thing
19:16:56 <stevemar> dtroyer_zz: it landed yesterday?
19:17:00 <terrylhowe> both nice fixes
19:17:03 <stevemar> err.. arrived
19:17:04 <dtroyer_zz> my queue polling has been dropping packets lately
19:17:18 <stevemar> dtroyer_zz: nah, it showed yo 24 hrs ago :P
19:17:48 <dhellmann> stevemar: I'll be looking at it later today
19:18:04 <stevemar> dhellmann: ty sir! you're a gentleman and a scholar
19:18:40 <dtroyer_zz> so in https://review.openstack.org/#/c/198506/ Roxana makes a point about starting with interface and changing to endpoint-type (I may have prompted that).  evidently only Identity uses 'interface'.  Do we still want to make that change?  I'm waffling now…
19:19:05 <stevemar> i thought i approved that
19:19:19 <dtroyer_zz> no +A yet
19:19:29 <stevemar> i like interface since it makes jamie happy :P
19:19:43 <stevemar> and staying consistent with KSC is good
19:19:53 <shaleh> Roxana just wants her patch to land and be done :-)
19:19:54 <dtroyer_zz> we have been giving Jamie plenty of grief  ;)
19:20:09 <dtroyer_zz> terrylhowe: does the SDK have a single usage internally for this?
19:20:16 <terrylhowe> I have a change in occ for that
19:20:22 <stevemar> shaleh: i already apologized to roxana in person at keystone mid cycle!
19:20:29 <shaleh> stevemar: :-)
19:20:41 <terrylhowe> the sdk should probably go to interface.  It is using endpoint_type now.  i think I made a bug for it already.
19:21:05 <dtroyer_zz> it's me too, I've basically had a mental cache flush (and will again in a few weeks)
19:21:34 <dtroyer_zz> terrylhowe: ok, if you are confident that is going to happen, we need to be consistent there for sure
19:21:35 <stevemar> empty caches are good sometimes
19:21:55 <terrylhowe> yeh endpoint_type->interface will happen in sdk
19:22:11 <dtroyer_zz> ok, then I'm back on-board with it here too
19:22:14 <terrylhowe> in occ Monty just asked for a minor change in the backward compat
19:22:45 <stevemar> dtroyer_zz: aside from those, theres a whole whack of functional tests being added by some dude
19:22:55 <stevemar> i tried to keep up, but he's cranking them out
19:23:05 <terrylhowe> same here stevemar
19:23:17 <dtroyer_zz> how do they look?  consistent with the way we've been doing them?
19:23:25 <dtroyer_zz> I havne't looked at them at all yet
19:23:43 <stevemar> dtroyer_zz: yep, i weighed in heavily on the early ones
19:23:50 <dtroyer_zz> ok, good
19:23:51 <stevemar> but all subsequent ones are looking good
19:24:40 <stevemar> he's also going around adding --project-domain | --user-domain | --group-domain to the few spots we missed
19:25:03 <dtroyer_zz> cool
19:25:11 <dtroyer_zz> any other reviews for attention?
19:25:40 <stevemar> a few image related ones are up, haven't reviewed them yet
19:26:51 <dtroyer_zz> not counting network, and after volume v2, image v2 needed the most love…
19:27:24 <dtroyer_zz> #topic bugs
19:27:55 <stevemar> dtroyer_zz: volume v2 is pretty awesome now btw
19:28:01 <stevemar> abhide did great work there
19:28:12 <dtroyer_zz> https://bugs.launchpad.net/python-openstackclient/+bug/1472909 looks interesting… is this a case of OSC not doing the name -> ID translation properly?
19:28:12 <openstack> Launchpad bug 1472909 in python-openstackclient ""role assignment list" fails if two users in different domains have the same name" [Undecided,New]
19:28:14 <stevemar> image v2 has a few patches now
19:28:27 <terrylhowe> I’ll be back on network once this next sdk release goes out (this week soemtime)
19:28:33 <stevemar> dtroyer_zz: oh maybe
19:28:50 <stevemar> might have been...
19:28:58 <stevemar> err, 'might be'
19:29:02 <dtroyer_zz> that's the only reason I can think of why it might not be a server-side problem…
19:29:25 <dtroyer_zz> I'm not sure how to corss domains from the API intentionally
19:31:33 <dtroyer_zz> it looks like there are a couple of bugs connected to o-c-c
19:31:54 <dtroyer_zz> whoever mentioned doing a release there soon is right...
19:32:08 <stevemar> yep!
19:32:18 * dtroyer_zz too lazy to scroll back
19:32:26 <stevemar> we should do one soonish
19:32:30 <dtroyer_zz> any other bugs?
19:32:56 <stevemar> once the gate is fixed, and we push a few bug fixes
19:33:37 <dtroyer_zz> #topic open discussion
19:33:45 <dtroyer_zz> so, speaking of releases…
19:33:57 <dtroyer_zz> a) occ gets one soon, then…
19:34:07 <dtroyer_zz> b) osc gets one following
19:34:26 <dhellmann> cliff should get one when we land that second ux patch
19:34:27 <stevemar> err not quite
19:34:30 <stevemar> cliff gets one too
19:34:48 <stevemar> dhellmann: knows whats good
19:34:59 <dtroyer_zz> I think it would be goood to get the image work finished now before doing another release
19:35:05 <dtroyer_zz> re cliff: agreed
19:36:15 <dtroyer_zz> and we have a new process for doing the releases.  dhellmann, would it be helpful if we do depends-on in those reviews to set order?
19:36:24 <stevemar> dtroyer_zz: i think we can say we have volume v2 support now: https://github.com/openstack/python-openstackclient/blob/master/setup.cfg#L378-L405
19:36:49 <dhellmann> dtroyer_zz: some of the repos from the team aren't being managed, so we should talk about that, too
19:37:04 <dhellmann> does the order matter?
19:37:05 <stevemar> dhellmann: hmm? which repos?
19:37:08 <dtroyer_zz> hmmm…. I saw cliff in there, so occ?
19:37:17 <dhellmann> yeah, occ is release:independent
19:37:25 <dtroyer_zz> order only in that cliff and occ need to come before osc
19:37:40 <dhellmann> dtroyer_zz: ok, you could just submit them as a series of patches
19:38:00 <dtroyer_zz> ah, right, that would be easier to follow
19:38:10 <dhellmann> yeah, depends-on is only needed for cross-repo things
19:38:16 <stevemar> oh more open discussion
19:38:29 <dhellmann> and it's possible we'll disallow those for releases, to avoid having people try to ask for a release of something that hasn't merged
19:38:33 <stevemar> abhide asked me if i had a bp he could work on
19:38:53 <stevemar> i told him to do this: https://blueprints.launchpad.net/python-openstackclient/+spec/global-limits
19:39:26 <shaleh> ooo fun
19:40:24 <dtroyer_zz> shaleh: looking through the BP's, there is https://blueprints.launchpad.net/python-openstackclient/+spec/image-member that applies to the image stuff you are working on…go ahead and reference that and we can close it when the image create is finished
19:40:34 <terrylhowe> should that bp have —marker?
19:40:45 <shaleh> dtroyer_zz: shall do
19:40:47 <stevemar> dtroyer_zz: i was going to say, i plan on marking some BPs are obsolete, ceilometer/heat/trove/sahara/magnum/manila plugin bps
19:40:51 <dtroyer_zz> err, image member or whatever it is called
19:41:08 <dtroyer_zz> stevemar: go for it
19:41:11 <stevemar> ty
19:42:48 <dtroyer_zz> any other things to discuss?
19:43:02 <stevemar> dtroyer_zz: there we go, we lost some weight there: https://blueprints.launchpad.net/python-openstackclient
19:43:31 <stevemar> dtroyer_zz: i'm in a `go-through-bps` mood
19:44:09 <shaleh> dtroyer_zz: https://blueprints.launchpad.net/python-openstackclient/+spec/api-list
19:44:16 <shaleh> dtroyer_zz: want help on that?
19:45:20 <dtroyer_zz> shaleh: I did that three times already and none of them really worked right due to the wide variances of server response.
19:45:47 <dtroyer_zz> OSC has a hack in it for mis-configured keystone but the rest are still sitting in a branch somewhere
19:46:00 <stevemar> oh oh, before i forget...
19:46:14 <stevemar> dtroyer_zz: really need your input on this one
19:46:14 <dtroyer_zz> I think the ideal solution is to generalize it into the SDK
19:46:32 <dtroyer_zz> and handle the project-specific hacks there
19:46:41 <stevemar> we are nuking the keystone CLI very soon, when we release keystoneclient 2.0
19:46:48 <dtroyer_zz> \o/
19:46:56 <stevemar> bknudson had some remarks about morgans work: https://review.openstack.org/#/c/196297/3
19:47:15 <stevemar> if you could check PS3, and brants comments... that would be great
19:47:15 <dtroyer_zz> I left a comment there this morning
19:47:26 <stevemar> i think brant it crazy, but that's not unusual
19:47:45 <dtroyer_zz> on the specific one you posted…
19:48:42 <dtroyer_zz> in general, OSC has tried to avoid using the oslo-config based arg handling, so I'm of the opinion that removing that should hurt us.  it might affect others...
19:48:53 <dtroyer_zz> er, shouldn't hurt us
19:48:55 <stevemar> should or shoudln't
19:48:59 <stevemar> ok cool
19:49:45 <dtroyer_zz> do you have plans to set any version caps anywhere before 2.0 is released?
19:49:53 <dtroyer_zz> or find the breakage then?
19:50:10 <stevemar> not sure, gotta ask morgan
19:51:35 <dtroyer_zz> ok sports fans, anything else today?
19:51:55 <terrylhowe> nada
19:52:09 <terrylhowe> just a whole bunch of patches to push through
19:52:45 <stevemar> terrylhowe: yeah
19:53:02 <stevemar> once we get our open patch queue to something reasonable we should push for a release
19:53:22 <terrylhowe> all these gate problems have made things tough
19:53:41 <stevemar> terrylhowe: yeah, at least we have fixes in place now
19:53:43 <dtroyer_zz> yes.  I may not have time to do that next week, but with the current processes that is basically pulling together release notes in the docs
19:53:48 <terrylhowe> yep
19:54:16 <stevemar> dtroyer_zz: that's all i got
19:54:20 <stevemar> dtroyer_zz: wrap it up!
19:54:29 <dtroyer_zz> ok, we'll call it a day then
19:54:33 <dtroyer_zz> #endmeeting