| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2017-01-13 | |||
| 17:36:59 | stevemar | dtroyer: this should knock one off | |
| 17:37:02 | stevemar | ^ | |
| 17:37:09 | stevemar | well 2 failures | |
| 17:37:44 | dtroyer | my kingdom for a low-level API that doesn't do magic | |
| 17:45:24 | stevemar | dtroyer: http://gifyu.com/images/magicb41e5.gif | |
| 17:45:50 | stevemar | ankur-gupta-f1: are you working on a particular failure? | |
| 17:45:56 | ankur-gupta-f1 | floating ip | |
| 17:46:12 | ankur-gupta-f1 | all actions are being taken on the ip addr rather than the id | |
| 17:47:03 | stevemar | rgr that | |
| 17:47:13 | stevemar | dtroyer: is working on sec rule | |
| 17:51:13 | dtroyer | stevemar: Henning!!! | |
| 17:57:18 | stevemar | dtroyer: oh wow, you know the person, | |
| 17:57:28 | stevemar | dtroyer: i know him as the magician meme | |
| 17:57:50 | dtroyer | well, duh, every kid magic fan in the '70s knows Doug Henning! And on Match Game no less! :) | |
| 18:01:31 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient: Fix security group rule list for NEtwork v2 https://review.openstack.org/420113 | |
| 18:01:39 | dtroyer | stevemar: ^^^ there might be one too | |
| 18:01:50 | stevemar | NETwork | |
| 18:03:10 | openstackgerrit | Steve Martinelli proposed openstack/python-openstackclient: fix functional tests for network agents https://review.openstack.org/420098 | |
| 18:03:20 | stevemar | dtroyer: hehe, forgot to skip my tests | |
| 18:03:25 | stevemar | unskip, even | |
| 18:06:43 | openstackgerrit | Ankur proposed openstack/python-openstackclient: Fixed floating IP functional tests https://review.openstack.org/420117 | |
| 18:07:01 | ankur-gupta-f1 | my env just broken. May have borked it. But ^^^^^ | |
| 18:07:07 | ankur-gupta-f1 | headed to lunch b back in 20 | |
| 18:26:58 | dtroyer | not NETwork, just NEtwork. My pinky finger isn't two keystrokes too slow, just one. jeez… :) | |
| 18:35:49 | openstackgerrit | Nikita Gerasimov proposed openstack/python-openstackclient: Switch server create to block_device_mapping_v2 https://review.openstack.org/407111 | |
| 18:55:22 | sigmavirus | dtroyer: If I point at Requests, will I get your kingdom? | |
| 18:57:08 | dtroyer | sigmavirus: only the sub-basement… we already use that, I'm in need to a just-above-rest+JSON layer for Openstack APIs. Just enough to hide the actual REST nastiness we've inflicted on our consumers | |
| 18:57:37 | dtroyer | SDK is waaaaay too high, much magic and redirection to find anything | |
| 19:09:30 | openstackgerrit | Sindhu Devale proposed openstack/python-openstackclient: OSC Quota List https://review.openstack.org/379813 | |
| 19:22:31 | openstackgerrit | Nikita Gerasimov proposed openstack/python-openstackclient: Switch server create to block_device_mapping_v2 https://review.openstack.org/407111 | |
| 19:23:18 | stevemar | dtroyer: back from lunch, whats the deal | |
| 19:23:38 | stevemar | dtroyer: we should cascade the changed on top of the skip test patch | |
| 19:25:29 | dtroyer | not mine, it is stand-alone and actually fixes a bug | |
| 19:25:44 | dtroyer | the floating IP ones are sdk bugs, working them now | |
| 19:26:02 | dtroyer | some of our code has been bad for a while | |
| 19:28:37 | stevemar | dtroyer: stupid functional tests | |
| 19:28:43 | stevemar | *shakes fist* | |
| 19:28:51 | stevemar | i blame my ocd for needing to move that file | |
| 19:28:54 | openstackgerrit | Ankur proposed openstack/python-openstackclient: Fixed floating IP functional tests https://review.openstack.org/420117 | |
| 19:29:23 | dtroyer | ankur-gupta-f1: the floating Ip issue is an actual SDK bug, not bad tests | |
| 19:30:08 | ankur-gupta-f1 | well the tests them selves were trying to delete floating ip with a IP addr. My change makes it use the FIPs ID | |
| 19:30:31 | dtroyer | that's the bug | |
| 19:30:35 | dtroyer | we document that as working | |
| 19:30:43 | dtroyer | those test are supposed to do that | |
| 19:30:48 | ankur-gupta-f1 | ah. fair enough. my bad then | |
| 19:31:04 | dtroyer | I suspect this code hasn't worked in a long time, if ever | |
| 19:40:41 | stevemar | dtroyer: o/ | |
| 19:40:58 | dtroyer | \o | |
| 19:41:12 | stevemar | so, you have a comment in test_network | |
| 19:41:24 | stevemar | # so make sure it stays that way. | |
| 19:41:24 | stevemar | # NOTE(dtroyer): is_default is not present in the create output | |
| 19:41:31 | stevemar | but apparently its there now | |
| 19:41:50 | stevemar | want me to pop it out from the display or fix the test? | |
| 19:42:19 | dtroyer | it should match nova-net if there is an equivalent | |
| 19:42:25 | stevemar | it is set to None atm | |
| 19:42:32 | dtroyer | I don't recall offhand which that is | |
| 19:45:21 | stevemar | i dont see any formatters for it in network.py | |
| 19:46:31 | stevemar | dtroyer: looks like it's not a nova-net option | |
| 19:46:46 | stevemar | since it's in get_attrs and not get_compute_attrs | |
| 19:49:30 | dtroyer | kk | |
| 19:53:32 | stevemar | dtroyer: well i tried passing in --default and --no-default | |
| 19:53:39 | stevemar | and well, it always come back as None | |
| 19:53:42 | stevemar | http://developer.openstack.org/api-ref/networking/v2/?expanded=create-network-provider-network-detail#create-network-provider-network | |
| 19:53:45 | stevemar | looking at that ^ | |
| 19:54:00 | stevemar | theres no "is_default" in the body expected (which is what the SDK passed) | |
| 19:55:49 | stevemar | theres no default anything in the create network APIs related stuff *shrug* | |
| 19:56:59 | stevemar | we could just keep it in there, the SDK returns it, so be it, maybe it's plumbing for something else | |
| 20:03:41 | openstackgerrit | Steve Martinelli proposed openstack/python-openstackclient: fix network functional tests for sdk 0.9.12 https://review.openstack.org/420154 | |
| 20:05:16 | stevemar | dtroyer: i'm good with https://review.openstack.org/#/c/420113/1 going in | |
| 20:07:34 | openstackgerrit | Steve Martinelli proposed openstack/python-openstackclient: fix functional tests for network agents https://review.openstack.org/420098 | |
| 20:08:26 | dtroyer | hmmm -tips isn't happy there | |
| 20:11:29 | stevemar | dtroyer: yep | |
| 20:11:39 | stevemar | mordred dtroyer know why this is failing? http://logs.openstack.org/98/420098/2/check/gate-osc-dsvm-functional-tips-ubuntu-xenial-nv/41e3ce3/console.html#_2017-01-13_18_35_30_737948 | |
| 20:11:45 | stevemar | ERROR: InvocationError: '/opt/stack/new/python-openstackclient/.tox/functional-tips/bin/pip install -q -U -e /opt/stack/new/osc-lib' | |
| 20:11:50 | stevemar | i'm doing it locally and it's fine | |
| 20:12:03 | mordred | stevemar: hrm. looking | |
| 20:12:19 | stevemar | mordred: it was your recent change to osc ;) | |
| 20:14:00 | mordred | yah. ... | |
| 20:15:16 | mordred | stevemar: ooh, interesting | |
| 20:15:19 | mordred | InstallationError: Command "python setup.py egg_info" failed with error code 1 in /opt/stack/new/osc-lib/ | |
| 20:15:49 | mordred | error: could not create 'osc_lib.egg-info': Permission denied | |
| 20:15:56 | mordred | crapballs | |
| 20:16:08 | stevemar | mordred: sudo ness? | |
| 20:16:43 | mordred | stevemar: k. patch coming | |
| 20:19:17 | openstackgerrit | Monty Taylor proposed openstack/python-openstackclient: Use git+file urls instead of directories https://review.openstack.org/420162 | |
| 20:19:25 | mordred | stevemar: let's see if that works a bit better | |
| 20:20:24 | stevemar | mordred: you're in all sorts of merge conflict, i fixed your patch before it merged (you were installing osc instead of the sdk) | |
| 20:20:41 | mordred | oh. bleh | |
| 20:21:02 | mordred | nice! | |
| 20:21:19 | stevemar | hmm | |
| 20:21:26 | stevemar | i think we may run into the same issue, not sure | |
| 20:21:49 | openstackgerrit | Monty Taylor proposed openstack/python-openstackclient: Use git+file urls instead of directories https://review.openstack.org/420162 | |
| 20:21:54 | mordred | stevemar: maybe that'll be better | |
| 20:22:47 | mordred | stevemar: if we get this working right, we should go poke the oslo folks about getting their stuff updated to point at /opt/stack/new instead of http | |
| 20:24:44 | stevemar | mordred: pardon my ignorance, but whats the big advantage? | |
| 20:32:57 | sindhu | stevemar: so where can I find the definitions of all the methods defined under a particular client? for example if I have compute_client = self.app.client_manager.compute and it contains list_extensions which in turn has all the methods like _create, _delete, _get | |
| 20:33:36 | mordred | stevemar: the big advantage is that /opt/stack/new will have the repo set up by zuul with depends-on changes taken in to account | |
| 20:33:42 | mordred | stevemar: so for most patches it'll be identical | |
| 20:34:34 | mordred | stevemar: but if you ever want to have a depends-on patch in occ, sdk or osc-lib and then show things working in the osc patch, you'll need to clone from /opt/stack/new and not from git.o.o | |
| 20:35:04 | stevemar | mmm ok | |