| *** yolanda has quit IRC | 00:48 | |
| *** slaweq has joined #openstack-shade | 01:31 | |
| *** slaweq has quit IRC | 01:35 | |
| *** yolanda has joined #openstack-shade | 03:11 | |
| *** slaweq has joined #openstack-shade | 03:46 | |
| *** slaweq has quit IRC | 03:51 | |
| *** yfried has joined #openstack-shade | 03:53 | |
| *** yfried has quit IRC | 03:58 | |
| *** gkadam has joined #openstack-shade | 04:00 | |
| *** gkadam has quit IRC | 04:01 | |
| *** gkadam has joined #openstack-shade | 04:03 | |
| *** yolanda has quit IRC | 04:35 | |
| *** yolanda has joined #openstack-shade | 06:28 | |
| *** yolanda has quit IRC | 07:22 | |
| *** yolanda has joined #openstack-shade | 07:43 | |
| *** slaweq has joined #openstack-shade | 07:47 | |
| *** slaweq has quit IRC | 07:51 | |
| *** ioggstream has joined #openstack-shade | 08:02 | |
| *** yolanda has quit IRC | 08:03 | |
| *** yolanda has joined #openstack-shade | 09:23 | |
| *** cdent has joined #openstack-shade | 09:45 | |
| *** yolanda has quit IRC | 10:48 | |
| *** cdent_ has joined #openstack-shade | 10:59 | |
| *** cdent has quit IRC | 11:01 | |
| *** cdent_ is now known as cdent | 11:01 | |
| *** nokes has joined #openstack-shade | 11:31 | |
| *** noshankus has quit IRC | 11:32 | |
| *** nokes is now known as noshankus | 11:32 | |
| *** yolanda has joined #openstack-shade | 12:24 | |
| *** nokes has joined #openstack-shade | 12:26 | |
| *** noshankus has quit IRC | 12:27 | |
| *** nokes is now known as noshankus | 12:27 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Ops https://review.openstack.org/473394 | 12:39 |
|---|---|---|
| samueldmq | morning shade | 12:39 |
| mordred | samueldmq: morning! comment left for you | 12:43 |
| samueldmq | mordred: hey | 12:44 |
| samueldmq | mordred: thanks, that's useful | 12:44 |
| mordred | \o/ | 12:44 |
| samueldmq | there are a lot of things there I'd like to figure out | 12:44 |
| samueldmq | see my comments (TODO,NOTE), as even gerrit review comments | 12:45 |
| mordred | kk | 12:45 |
| samueldmq | I might need to check the ksc code closer | 12:45 |
| samueldmq | gotta go afk for a bit, brb | 12:45 |
| mordred | kk | 12:49 |
| *** gouthamr has joined #openstack-shade | 12:55 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: DNM Test patch to find bug with nics https://review.openstack.org/473452 | 13:02 |
| *** cdent has quit IRC | 13:20 | |
| *** cdent has joined #openstack-shade | 13:26 | |
| rods | https://developer.openstack.org/api-ref/dns/?expanded=create-zone-detail#create-zone doesn't return a top-level container, yay for consistency | 13:27 |
| mordred | rods: \o/ | 13:36 |
| rods | mordred not a big deal just annoying because now I have to invert the params order here to make `key` optional https://github.com/openstack-infra/shade/blob/master/shade/meta.py#L579 | 13:39 |
| rods | unless we don't want to pass '' or None there when a top-level container is not needed to make it more explicit | 13:40 |
| rods | which I think wouldn't be a bad idea | 13:40 |
| rods | s/needed/present/g | 13:40 |
| mordred | rods: yah - maybe that - get_and_munchify(key=None, data=data) | 13:40 |
| rods | ya | 13:43 |
| *** ioggstream is now known as ioggstrea | 14:31 | |
| *** ioggstrea is now known as _iogg | 14:31 | |
| *** _iogg is now known as ioggstrea | 14:31 | |
| *** ioggstrea is now known as ioggstream | 14:31 | |
| *** gkadam has quit IRC | 14:35 | |
| *** yolanda has quit IRC | 14:45 | |
| samueldmq | mordred: shade functional tests is running on identity v3, correct? | 15:48 |
| samueldmq | are | 15:48 |
| samueldmq | I mean in the gate job gate-shade-functional | 15:48 |
| samueldmq | I guess we'd want to test mulitple versions in that | 15:51 |
| mordred | samueldmq: we have jobs for both | 15:57 |
| mordred | samueldmq: there is an env var that you can check in the functional tests - which you should find some examples of already | 15:57 |
| openstackgerrit | Rosario Di Somma proposed openstack-infra/shade master: Don't remove top-container for stack and zone REST API calls https://review.openstack.org/473517 | 15:58 |
| *** slaweq has joined #openstack-shade | 16:15 | |
| *** slaweq has quit IRC | 16:19 | |
| samueldmq | mordred: where do we test v2? gate-shade-functional-legacy? | 16:31 |
| samueldmq | yes, however it'd be great if we disabled identity v3 in those tests, to make sure only v2.0 is in use | 16:37 |
| mordred | samueldmq: yes. and yes - we do some tricks to make sure shade doesn't use v3 | 16:38 |
| morgan | samueldmq: disabling v3 is likely going to be a bit weird. | 16:38 |
| morgan | but the tricks in shade are easier to use to ensure no v3 is used | 16:38 |
| mordred | morgan: ++ | 16:38 |
| samueldmq | where can I find those tricks? | 16:38 |
| mordred | samueldmq: in the setup, we force v2 in the clouds.yaml - then we have a test that makes sure that shade doesn't detect v3 and use it anyway | 16:39 |
| *** cdent has quit IRC | 16:39 | |
| mordred | as long as the "use v2" env var is set | 16:39 |
| morgan | we could hook into the GETs and actually raise an exception (yay requests adapters) if a uri with /V3/ in identity is used | 16:39 |
| mordred | (so that we don't have to add that test to every test) | 16:39 |
| samueldmq | I mean, disabling v3 would be simple, but *could* make confusion maybe? since what we want to move is to v3 | 16:39 |
| morgan | samueldmq: yeah, i don't want to encourage in any test suite that v3 is optional | 16:39 |
| morgan | v2, is going away, and therefore is. | 16:39 |
| morgan | v3 is *not* optional, v2 is not really but shouldn't be used. | 16:40 |
| samueldmq | that makes sense, for the sake of making the transition process less complicated (than what it already is) | 16:40 |
| morgan | ++ | 16:40 |
| *** cdent has joined #openstack-shade | 16:41 | |
| morgan | mordred: maybe samueldmq can look into an adapter to add to requests/session object that can raise an exception (something we hook into ksa?) for testing that ensures v3 isn't used? | 16:41 |
| mordred | samueldmq, morgan: https://github.com/openstack-infra/shade/blob/master/shade/tests/functional/test_devstack.py | 16:41 |
| samueldmq | I am not seeing anything telling shade it should be using v2 when -legacy | 16:41 |
| morgan | and vice-versa | 16:41 |
| * samueldmq https://github.com/openstack-infra/project-config/blob/a3685b02599da5b9ed1c5228e0c8b54287d64306/jenkins/jobs/shade.yaml#L37-L44 | 16:41 | |
| morgan | might be useful for any API transitions in openstack | 16:41 |
| mordred | totally - that's the place in the tests where we verify that shade is testing what the test framework is requesting | 16:42 |
| morgan | "inspect uri, complain if pattern match shows inappropriate api is called" | 16:42 |
| morgan | mordred: we should get samueldmq to weigh in on the KSA patch chain | 16:42 |
| morgan | as well | 16:42 |
| morgan | since... you know +2 powers and smart dude. | 16:42 |
| samueldmq | the one in the api discovery, correct? | 16:43 |
| morgan | (not that other keystone cores aren't smart, but he's here now!) | 16:43 |
| morgan | samueldmq: yeah. | 16:43 |
| mordred | samueldmq: oh! thank you | 16:43 |
| mordred | samueldmq: there is a bug in that test definition - it should also set SHADE_USE_KEYSTONE_V2=1 in the if legacy section | 16:43 |
| mordred | samueldmq: if you can make a little patch there I can get it landed | 16:43 |
| samueldmq | mordred: exactly | 16:43 |
| samueldmq | sure, on it | 16:44 |
| samueldmq | mordred: thanks for the compliments | 16:44 |
| mordred | I missed setting that in the refactor | 16:44 |
| samueldmq | shade is fun | 16:44 |
| mordred | yay! | 16:44 |
| mordred | just wait til y'all see the patch I'm writing right now based on an issue yolanda found :) | 16:44 |
| * samueldmq makes popcorn | 16:45 | |
| morgan | mordred: i'm scurred | 16:46 |
| morgan | mordred: today's running commentary is powered by Death Wish Coffee(tm). You should get some. | 16:47 |
| mordred | mmm | 16:52 |
| samueldmq | mordred: that coffee brand seems interesting | 17:00 |
| samueldmq | oops morgan ^ | 17:00 |
| samueldmq | I need more chars to autocomplete you guys | 17:01 |
| morgan | samueldmq: hold on i'll make it more difficult | 17:03 |
| *** morgan is now known as mordgan | 17:03 | |
| mordgan | samueldmq, mordred: *shiftyeyes* | 17:03 |
| samueldmq | ehhehe | 17:03 |
| *** mordgan is now known as morgan | 17:03 | |
| samueldmq | he's gonna probably autocomplete himself now | 17:04 |
| rods | after https://review.openstack.org/#/c/473517/1 there are *just* other 72 tests to fix :) | 17:13 |
| rods | mostly nova | 17:14 |
| mordred | rods: I'm doing create_server right now | 17:20 |
| rods | mordred cool, I'm gonna work on the security group calls later in the evening | 17:23 |
| samueldmq | morgan: mordred: ksa adapters are just like things we put in the services pipelines | 17:24 |
| samueldmq | is that correct? | 17:24 |
| *** ioggstream has quit IRC | 17:24 | |
| samueldmq | that intercept the calls (in the order they're defined) and let them continue or not (or modify), whatever | 17:26 |
| mordred | samueldmq: I'm not sure I fully understand your question ... but a ksa adapater has a session and then information about the service that wants to be used. so it knows how to find the base URL for a given service | 17:35 |
| mordred | samueldmq: so if you have an adapter with service_type='compute', you can say get('/servers') and the adapter will translated that to GET https://compute.example.com/v2.1/servers | 17:36 |
| samueldmq | mordred: ah I thought you could combine adapters | 17:36 |
| mordred | (assuming https://compute.example.com/v2.1/servers is what's in the catalog and/or the resultof version discovery) | 17:36 |
| samueldmq | like in a paste pipeline | 17:36 |
| *** slaweq has joined #openstack-shade | 17:36 | |
| mordred | not really, no - it's all about encoding the base endpoint of a given service | 17:36 |
| samueldmq | ah okay, so shade's raw_clients are adapters | 17:37 |
| samueldmq | correct? | 17:37 |
| rods | samueldmq that's correct https://github.com/openstack-infra/shade/blob/master/shade/openstackcloud.py#L354 | 17:39 |
| samueldmq | rods: nice, thanks | 17:39 |
| *** slaweq has quit IRC | 17:41 | |
| samueldmq | in https://review.openstack.org/#/c/473394/ | 18:35 |
| samueldmq | I am mostly getting No mock address for the URL calls in our tests | 18:36 |
| samueldmq | does anyone have a tip on how to quickly fix that? (seems to be something simple I am missing) | 18:36 |
| samueldmq | do I need to add all the URLs to the uri_registry one by one ? D: | 18:39 |
| morgan | samueldmq: you need to make sure everything is mcked out in the register_uris | 18:44 |
| morgan | so.. yes | 18:45 |
| morgan | if you look at other tests it models the entire process | 18:45 |
| morgan | and they need to be in order in register_uris | 18:45 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Convert create_server mocks to request_mock https://review.openstack.org/473554 | 18:45 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: RESTify create_server https://review.openstack.org/473555 | 18:45 |
| morgan | so that you can do the assert at the end of the test | 18:45 |
| mordred | samueldmq, morgan, rods, Shrews: ^^ enjoy :) | 18:46 |
| morgan | mordred: oh god | 18:46 |
| mordred | morgan: actually wans't as terrible as I feared | 18:46 |
| samueldmq | oh wait, 433 lines touched in that test | 18:47 |
| mordred | samueldmq: yah - step 1- convert from mocking novaclient to doing requests_mock but with novaclient still in place - that proves the http mocks in the test are correct | 18:48 |
| mordred | samueldmq: step 2 - change the call from novaclient to compute adapter - preferrably not changing any of the tests | 18:49 |
| mordred | in this case, we do change the tests to remove an extra call novaclient injects that we don't need | 18:49 |
| mordred | but we try to avoid that as much as possible | 18:49 |
| mordred | we _could_ have that patch add the extra GET call itself and then do a third patch to remove the GET call and change the tests - but that seemed a little extra pedantic | 18:50 |
| samueldmq | yeah, could've been done in a third patch so it's clearer | 18:50 |
| samueldmq | ++ but that seems to be fine (now that you explained | 18:50 |
| samueldmq | why do we have tests using requests_mock in shade if we were using the python-*clients | 19:00 |
| samueldmq | why did we have them before the restification work? did we check on the python-*client adapters? | 19:01 |
| mordred | samueldmq: yah - before we started this, we were just mocking the calls to python-*client | 19:01 |
| mordred | samueldmq: but the nice thing about moving to requests_mock is that we can express the thing we want to test in terms of the actual rest interactions | 19:01 |
| mordred | and it _should_ work whether we're using python-*client to do the REST for us, or doing it directly | 19:02 |
| * mordred afks for a bit | 19:02 | |
| samueldmq | kk gotcha | 19:03 |
| samueldmq | I was wondering why https://github.com/openstack-infra/shade/blob/master/shade/tests/unit/test_project.py#L142 was working | 19:03 |
| samueldmq | since update_* in keystone are PATCH calls, and tehre is no PATCH registry in that test | 19:03 |
| samueldmq | hmm I like request_mock, it fails when the URLs called dont match exactly how you registered the URIs | 19:10 |
| *** slaweq has joined #openstack-shade | 19:30 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Ops https://review.openstack.org/473394 | 19:55 |
| mordred | samueldmq: https://developer.openstack.org/api-ref/identity/v2-admin/index.html?expanded=update-tenant-detail#update-tenant in v2, update tenant seems to be POST | 20:23 |
| samueldmq | mordred: just to make it more interesting, thanks | 20:24 |
| mordred | samueldmq: :) | 20:28 |
| samueldmq | mordred: in grant_role https://github.com/openstack-infra/shade/blob/master/shade/operatorcloud.py#L1679 | 20:42 |
| samueldmq | what is that domain for? for the project? for the user? or are we assuming project and user are always in the same domain? | 20:42 |
| mordred | samueldmq: it's for both at the moment if either user or project are passed by name | 20:44 |
| samueldmq | gotcha, just checking as that question appeared when I was fixing tests | 20:45 |
| mordred | samueldmq: in the current api, if you want to do user and domain in different domains, you'd need to do a get_user and get_project call and pass those objects in as objects not names | 20:45 |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project List https://review.openstack.org/473581 | 20:45 |
| mordred | samueldmq: yah - it's a good question - and worth considering whether we should add support for other things | 20:45 |
| samueldmq | mordred: ^ let's see how the functional tests like that | 20:46 |
| mordred | samueldmq: and/or update the docs to be clearer | 20:46 |
| mordred | samueldmq: woot | 20:46 |
| samueldmq | mordred: ++ | 20:46 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: RESTify create_server https://review.openstack.org/473555 | 20:47 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Don't fetch extra_specs in functional tests https://review.openstack.org/473583 | 20:47 |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: Improve grant docs on when and how use domain arg https://review.openstack.org/473589 | 21:03 |
| samueldmq | mordred: ^ I need to do it right away or I forget about it :) | 21:03 |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project List https://review.openstack.org/473581 | 21:08 |
| *** gouthamr has quit IRC | 21:19 | |
| *** gouthamr has joined #openstack-shade | 21:39 | |
| *** slaweq has quit IRC | 21:47 | |
| *** slaweq has joined #openstack-shade | 21:47 | |
| *** gouthamr has quit IRC | 21:47 | |
| *** gouthamr has joined #openstack-shade | 21:48 | |
| *** gouthamr has quit IRC | 21:48 | |
| *** slaweq has quit IRC | 21:51 | |
| *** slaweq has joined #openstack-shade | 21:55 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Create https://review.openstack.org/473609 | 22:05 |
| *** yolanda has joined #openstack-shade | 22:34 | |
| *** slaweq has quit IRC | 22:38 | |
| *** cdent has quit IRC | 22:38 | |
| *** slaweq has joined #openstack-shade | 22:38 | |
| *** cdent has joined #openstack-shade | 22:39 | |
| *** slaweq has quit IRC | 22:40 | |
| *** slaweq has joined #openstack-shade | 22:41 | |
| *** slaweq has quit IRC | 22:43 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Don't fetch extra_specs in functional tests https://review.openstack.org/473583 | 22:44 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Convert create_server mocks to request_mock https://review.openstack.org/473554 | 22:44 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: RESTify create_server https://review.openstack.org/473555 | 22:44 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Add boot from volume unit tests https://review.openstack.org/473616 | 22:44 |
| *** cmurphy has joined #openstack-shade | 22:46 | |
| *** gouthamr has joined #openstack-shade | 22:52 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project List https://review.openstack.org/473581 | 23:13 |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Create https://review.openstack.org/473609 | 23:38 |
| *** cdent has quit IRC | 23:39 | |
| openstackgerrit | Samuel de Medeiros Queiroz proposed openstack-infra/shade master: De-client-ify Project Delete https://review.openstack.org/473628 | 23:52 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!