19:03:30 <dtroyer> #startmeeting openstackclient
19:03:31 <openstack> Meeting started Thu Mar 16 19:03:30 2017 UTC and is due to finish in 60 minutes.  The chair is dtroyer. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:03:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:03:35 <openstack> The meeting name has been set to 'openstackclient'
19:03:47 <dtroyer> Who is here for an OSC meeting?
19:03:59 <rabel> i am :)
19:04:27 <rabel> but more or less as a listener, since i just started contributing lately
19:04:52 <dtroyer> welcome rabel, glad to have you here, both meeting-wise and with the project
19:05:36 <rabel> thanks :)
19:06:10 <dtroyer> This may be short, I don't expect stevemar to be here much anymore, but he is still hitting reviews it seems
19:07:09 <dtroyer> These meetings are fairly free-form since there isn't usually a lot of folks here
19:07:14 <dtroyer> So I'll drop a couple of status things here for the logs then we'll see if anything else comes up
19:07:24 <ankur-gupta-f4> alrighty
19:07:24 <dtroyer> #topic releases
19:08:00 <ankur-gupta-f4> welcome rabel
19:08:08 <dtroyer> We released 3.9.0 last week to release fixes for the SDK 0.9.14 release that made some API changes to network
19:08:32 <ankur-gupta-f4> any known priority bugs?
19:08:59 <dtroyer> it was also the first rlease after the freeze so a lot of stuff had accumilated anyway
19:09:30 <dtroyer> ankur-gupta-f1: do you mean outstanding or part of the release?
19:09:36 <ankur-gupta-f4> from release
19:09:55 * dtroyer glances through the release notes
19:10:21 <dtroyer> the major ones are the fixes for the network session changes, that's what triggered the sudden quick release
19:10:33 <ankur-gupta-f4> the v1 image tests got fixed but still seeing scattered functional failures. Wondering if its due to known issue or just race conditions. Haven't gotten chance to dig in yet
19:10:46 <dtroyer> I'll get to that in a minute
19:13:20 <stevemar> dtroyer: i randomly have a gap in my calendar, i'll tune in :)
19:14:18 <dtroyer> I would like to get through some of the long-running command additions soon and cut 3.10.0 and try to get caught up.  Then there is a series of things around changes to os-client-config and auth and argument handling coming up
19:15:08 <dtroyer> It is quite possible that those will trigger 4.0 as I expect some of the details of how options are parsed may change and we'll need to signal a compaibility change
19:15:48 <ankur-gupta-f4> dtroyer: probably the best candidate to fit that category is this patch https://review.openstack.org/#/c/383025/ .
19:15:56 <dtroyer> When we start that, I am planning to do a feature branch for the 4.0 work so we don't block so long on bug fixes while waiting for that to complete
19:16:05 <ankur-gupta-f4> is there a spec/bp in place that discusses the potential changes to option parsing
19:16:46 <dtroyer> no, it's all in my, mordred and jamielennox's heads still.  Jamie has started with the simplification of occ and a chunk of that needs to get completed first anyway.
19:17:08 <ankur-gupta-f4> hah okay
19:17:16 <ankur-gupta-f4> will keep eye out
19:17:17 <dtroyer> my intention is to not change anything, but at some point we need to get rid of some old behaviour that is causing us pain and to be different than other tools using occ (shade, etc)
19:18:08 <dtroyer> another high-priority thing is to get the microversion handling sorted out for compute and volume
19:19:06 <dtroyer> those are the high-level plans that are not "implement more missing commands"
19:19:47 <dtroyer> oh, forgot one, needing to build the hooks for plugins to add to common commands like quotas
19:20:09 <dtroyer> there are some other implications there too that may be best timed for 4.0
19:20:51 <dtroyer> any questions so far?
19:20:54 <rabel> is "implement more missing commands" the kind of task that is also suitable new at osc development? like me?
19:21:06 <ankur-gupta-f4> do we want to do micro versions this release or wait for next. thinking of the resources needed to get both occ and micro version support done for Pike
19:21:18 <rabel> i wanted to say "...for someone new..."
19:21:38 <dtroyer> that is the category that things like the remaining network commands and volume v3 fall in to, those will always be around to some degree and always on the roadmap
19:22:37 <dtroyer> the microversion handling is more connected to the plumbing insetting up compute and volume clients properly than to getting the options from the user (we already have that)
19:23:01 <dtroyer> and also to building the bits we need for branching inside command classes to handle multple versions
19:24:22 <ankur-gupta-f4> ack
19:25:07 <dtroyer> #topic reviews
19:26:00 <ankur-gupta-f4> Flavor Profile: https://review.openstack.org/#/c/396783/
19:26:03 <dtroyer> The current state of the queue is back to operational thanks to frickler and RuiChen and whomever else caught my failed attempt to restore image v1 to our devstack
19:26:40 <dtroyer> FWIW, the last change frickler submitted should also fix the volume backing file size that I believe is the root cause of our volume failures
19:27:13 <dtroyer> That seems random because the order and parallelism in running the tests is non-deterministic, and when enough volume tests run at the same time we run out of volume space
19:27:18 <ankur-gupta-f4> will keep eye on gate
19:27:36 <ankur-gupta-f4> FIP/Port : https://review.openstack.org/#/c/383025/
19:27:45 <dtroyer> so once https://review.openstack.org/#/c/446504/ merges we should see a drop in the volume failures
19:28:21 <rabel> thats nice!
19:30:40 <dtroyer> ankur-gupta-f1: we need to wrap up both of those soon, they have gone on a long time
19:31:16 <ankur-gupta-f4> We can push the 2 agent commands later. but Flavor and the FIP should be top of the networking commands list
19:31:41 <dtroyer> I just glanced at today's comments on 383025, need to address command formats again
19:32:21 <sindhu> dtroyer:  for reviews:   https://review.openstack.org/#/c/379813/   ,   https://review.openstack.org/#/c/382023/  as well
19:33:05 <ankur-gupta-f4> sindhu: address Rui Chens comments on 382023 first please
19:33:53 <sindhu> ankur-gupta-f4: I need dtroyer's comment on that. Whether it is needed or not
19:34:56 <dtroyer> which comment?  the 'do we need a show command' one?
19:35:06 <sindhu> dtroyer: yes
19:35:18 <dtroyer> I don't have a problem with having a show command even if we can get all of the fields via list
19:35:46 <dtroyer> because with list you either need to hand the filters on the command to get just one, or you need to filter it yousrlf (as a user)
19:36:12 <dtroyer> this happens with other resources where —long shows everything
19:36:41 <sindhu> dtroyer: ok great. I will address the rest of the comments today and keep it review ready :)
19:37:15 <ankur-gupta-f4> stevemar: will you get a chance to do some reviews as well today?
19:37:42 <dtroyer> I need to refresh on this overall though, and to make sure that it isn't too network-centric where command changes will be required to add other APIs
19:40:07 <dtroyer> also, I consider bulk changes like the six.iteritems() reviews to be low priority as they are mostly cosmetic and could possible conflict with a large number of other reviews in flight
19:41:41 <dtroyer> I just took a quick look at the agent review (https://review.openstack.org/#/c/387611/), is there still outstanding questions on that regarding the command formats?
19:42:44 <ankur-gupta-f4> I don't think so. Pretty much resolved
19:42:59 <dtroyer> cool
19:43:07 <dtroyer> any other reviews that we need to discuss?
19:43:17 <ankur-gupta-f4> let me check my list
19:44:23 <ankur-gupta-f4> My top items were Flavor Profile, FIP/Port, Quota List, the 2 agent commands.
19:44:47 <ankur-gupta-f4> quota show once those comments are resolved.
19:45:05 <ankur-gupta-f4> been keeping an eye on the v3 volume as it comes in
19:46:08 <dtroyer> volume v3 is special, in that we lat it point to the v2 commands rather than copy them.
19:46:32 <dtroyer> we MUST not let any v3 changes in to the v2 tree, those classes need to be copied as changes are made
19:47:04 <dtroyer> having to remember that is one reason I've insisted prior to this to just copy the entire tree :)
19:47:22 <ankur-gupta-f4> correct just changing the env api-version
19:48:02 <ankur-gupta-f4> first time i saw https://github.com/openstack/python-openstackclient/blob/master/openstackclient/tests/functional/volume/v3/test_volume.py#L17
19:48:09 <ankur-gupta-f4> i was like ... why....
19:49:04 <dtroyer> so that will force the tests top be copied to actually work, as they will need to use the right superclass
19:49:27 <dtroyer> so yeah, that chould just copy the v2 file to /v3 and go from there
19:49:55 <dtroyer> ATM that is totally unused, and mostly harmless
19:50:55 <dtroyer> really, that should be the first thing we do, is copy all of the unit and functional tests before copying any code
19:51:46 <dtroyer> we're under 10 minutes left, so …
19:51:48 <dtroyer> #topic open discussion
19:51:48 <ankur-gupta-f4> Well thats all i got. Gotta quickly eat before next meeting in 10. If you remember the plugin client that is implemented just as plugin it would be really helpful otherwise will reference the cookie cutter. Thanks dtroyer:
19:51:58 <dtroyer> what else is on your minds?
19:52:02 <ankur-gupta-f4> rabel: welcome to the team.
19:52:23 <rabel> thanks ankur-gupta-f4. good to be here!
19:52:29 <dtroyer> ankur-gupta-f1: the osc-debug repo is exactly that, or am I misunderstanding you?
19:53:04 <ankur-gupta-f4> rabel: feel free to add me to any reviews.
19:53:35 <ankur-gupta-f4> dtroyer: oh yea. thats good enough.
19:53:50 <dtroyer> good deal
19:54:07 <rabel> ankur-gupta-f4: thank you. i'll probably do that ;)
19:54:23 <ankur-gupta-f4> rabel: don't fear the -1. embrace it :P
19:54:28 <rabel> :D
19:54:29 <dtroyer> on a client-related note I've been chasing some issues in cliff's stable gates…
19:54:46 * dhellmann perks up his ears
19:55:01 <dtroyer> we got into a bind with a combination of events that are requiring multiple things in a single review to unblock
19:55:26 <dtroyer> dhellmann: yeah, I was about to ask you about those, two are the appdirs problem that I thought I knew how to unblock
19:55:35 <dhellmann> which ones?
19:55:36 <dtroyer> clearly I'm not getting something right
19:55:59 <dtroyer> my current attempts:  newton https://review.openstack.org/#/c/445187/
19:56:14 <dtroyer> ocata: https://review.openstack.org/445186
19:56:37 <dtroyer> mitaka is passed but hung up in merge, I need to poke in -infra on that one I think
19:56:59 <dtroyer> unless it needs the other two to pass too for some reason I don't see
19:56:59 <dhellmann> oh, I removed my workflow bit because I thought we would want the changes in the newer branches to merge first
19:57:13 <dtroyer> ok, np
19:57:17 <dhellmann> only because I thought that was policy
19:57:28 <dhellmann> I'll poke at the other 2 patches this afternoon for a bit
19:57:40 <dtroyer> it probably is, I ended up poking at all three to see if they all had the same problems
19:57:43 <dtroyer> they didn't
19:57:55 <dhellmann> ok, I'll start with ocata then
19:58:14 <dtroyer> thanks
19:58:54 <dtroyer> Two minutes… anything else?
19:59:52 <dtroyer> so let's call it a meeting a minute early
19:59:55 <dtroyer> thanks everyone!
19:59:59 <dtroyer> #endmeeting