| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2017-12-06 | |||
| 08:26:24 | openstack | Launchpad bug 1736161 in python-designateclient "The records option gobbles up all trailing parameters" [Undecided,In progress] - Assigned to Dr. Jens Harbott (j-harbott) | |
| 12:13:59 | openstackgerrit | Jake Yip proposed openstack/osc-lib master: Fix sorting in Python 3 https://review.openstack.org/526046 | |
| 12:55:46 | openstackgerrit | Jake Yip proposed openstack/osc-lib master: Fix sorting in Python 3 https://review.openstack.org/526046 | |
| 12:57:56 | openstackgerrit | Jake Yip proposed openstack/osc-lib master: Fix sorting in Python 3 https://review.openstack.org/526046 | |
| 13:54:18 | dtroyer | frickler: we've done it "by hand" in the past, ie set help to SUPPRESS and log a warning in take_action() if it is used | |
| 14:01:35 | frickler | dtroyer: thx, I did the second part, do you have a sample for the help change? https://review.openstack.org/526008 | |
| 14:05:33 | dtroyer | it's just using help=argparse.SUPPRESS in the option declaration | |
| 14:06:05 | dtroyer | so the old option doesn't appear any more… | |
| 14:11:37 | frickler | ah, that's easy, thanks again | |
| 15:59:28 | openstackgerrit | Merged openstack/os-service-types master: Updated from global requirements https://review.openstack.org/525388 | |
| 16:05:32 | openstackgerrit | Mark Goddard proposed openstack-infra/shade master: Baremetal NIC list should return a list https://review.openstack.org/526105 | |
| 16:21:16 | openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Use devstack functional test base job https://review.openstack.org/526116 | |
| 16:22:39 | openstackgerrit | Monty Taylor proposed openstack/python-openstackclient master: Use devstack functional base job https://review.openstack.org/526117 | |
| 16:24:15 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Fix SDK Connection creation alternative to Profile https://review.openstack.org/523938 | |
| 17:13:08 | openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Fix batching for floating ips and ports https://review.openstack.org/526127 | |
| 17:36:20 | dmsimard | mordred: ohai | |
| 17:41:57 | dmsimard | supplemental is being called from here https://github.com/openstack-infra/shade/blob/master/shade/meta.py#L414 and add_server_interfaces is used in a couple places | |
| 17:43:23 | dmsimard | like in these two blocks https://github.com/openstack-infra/shade/blob/master/shade/openstackcloud.py#L3144-L3155 | |
| 17:44:39 | dmsimard | so 'addresses' is not populated until _get_supplemental_addresses runs at least once ? | |
| 17:45:25 | dmsimard | so we're falling in that block and then recursing https://github.com/openstack-infra/shade/blob/6b613f1a387f0c49ce834b0396ac70efb848d565/shade/meta.py#L372 | |
| 17:45:35 | dmsimard | which ends up doing all those floating IP calls ? /me adds debug | |
| 17:47:01 | dmsimard | for fip in cloud.search_floating_ips( filters=dict(port_id=port['id'])): | |
| 17:47:21 | dmsimard | sounds very much like GET call to network for https://phx2.cloud.rdoproject.org:13696/v2.0/floatingips.json?port_id=2fde8e86-ba92-448b-bbee-29c16f635cc3 used request id req-2d6942ca-2eab-41d1-a411-188aa6c5305e | |
| 17:48:40 | dmsimard | but we have the mapping of port id <-> floating ip in list_floating_ips(), let me try that | |
| 17:50:41 | mordred | dmsimard: ah - hrm. | |
| 17:52:03 | dmsimard | hang on, I might have a patch | |
| 17:54:04 | openstackgerrit | Dean Troyer proposed openstack/osc-lib master: Use converged SDK insead of os-client-config https://review.openstack.org/518240 | |
| 18:24:44 | peter___ | hello all, does anybody know why after making connection: conn = connection.Connection(**auth_args) and I want display project projects = conn.identity.list_projects() in ossdk doesnt working? | |
| 18:25:31 | peter___ | looks like object conn doesnt have identity.list_projects() method. | |
| 18:47:47 | dmsimard | mordred: yeah... so I'm seeing the problem but I'm not sure what's the best way to fix it. Here's a summary of what's happening for a server with 5 ports: http://paste.openstack.org/raw/628306/ | |
| 18:48:11 | dmsimard | mordred: We're doing 5 calls to list_floating_ips to check if any of those ports are a floating ip. | |
| 18:48:31 | dmsimard | mordred: It sounds like we should be filtering client side -- do one list_ports, one list_floating_ips, make them friends | |
| 18:49:06 | dmsimard | That's the two for loops from here: https://github.com/openstack-infra/shade/blob/6b613f1a387f0c49ce834b0396ac70efb848d565/shade/meta.py#L377-L380 | |
| 18:50:30 | dmsimard | If you're thinking "why does this server have like 5 ports", it's a tenant in which tripleo runs heat stacks for nested tripleo testing with isolated networks | |
| 18:57:33 | dmsimard | So if you have 100 servers with 5 ports each, we're doing 1*list_port, 5*list_floating_ips per server, that's 600 API calls just for _get_supplemental_addresses | |
| 19:18:10 | peter___ | import openstack.config ModuleNotFoundError: No module named 'openstack.config' | |
| 20:59:28 | openstackgerrit | Julia Kreger proposed openstack-infra/shade master: Baremetal NIC list should return a list https://review.openstack.org/526105 | |
| 21:19:57 | openstackgerrit | Dean Troyer proposed openstack/os-client-config master: Make the get_service_type() overrides tolernat of no defaults https://review.openstack.org/526201 | |
| 21:20:54 | openstackgerrit | Dean Troyer proposed openstack/python-openstacksdk master: Make the get_service_type() overrides tolernat of no defaults https://review.openstack.org/526202 | |
| 21:26:11 | openstackgerrit | Dean Troyer proposed openstack/osc-lib master: Use converged SDK insead of os-client-config https://review.openstack.org/518240 | |
| 21:26:46 | openstackgerrit | Dean Troyer proposed openstack/osc-lib master: Use converged SDK insead of os-client-config https://review.openstack.org/518240 | |
| 21:51:45 | SamYaple | mordred: i just found out there was a netsplit because i couldnt see you in any of the channels. so thanks forthat | |
| 21:52:24 | SamYaple | im wondering if you have any ideas on how to do an introspective valid_kwargs type thing with the saltstack shade modules | |
| 22:47:29 | openstackgerrit | Dean Troyer proposed openstack/python-openstacksdk master: Make the get_service_type() overrides tolernat of no defaults https://review.openstack.org/526202 | |
| 23:44:29 | openstackgerrit | Jake Yip proposed openstack/osc-lib master: Fix sorting in Python 3 https://review.openstack.org/526046 | |
| 23:46:23 | openstackgerrit | Jake Yip proposed openstack/osc-lib master: Fix sorting in Python 3 https://review.openstack.org/526046 | |
| #openstack-sdks - 2017-12-07 | |||
| 02:27:42 | openstackgerrit | Merged openstack-infra/shade master: Add retry logic mechanism https://review.openstack.org/525037 | |
| 11:33:31 | openstackgerrit | Slawek Kaplonski proposed openstack/python-openstackclient master: Add support for get details of Quota https://review.openstack.org/515401 | |
| 14:10:59 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Fix SDK Connection creation alternative to Profile https://review.openstack.org/523938 | |
| 14:27:29 | dtroyer | mordred, Shrews, other SDK fans: I have a depends-on chain that has https://review.openstack.org/#/c/526202/ at the bottom that (finally!!) gets the osc functional-tips job running cleanly. | |
| 14:27:52 | openstackgerrit | Yuki Nishiwaki proposed openstack/osc-lib master: Check auth_url as well as identity version api https://review.openstack.org/526413 | |
| 14:28:24 | dtroyer | The matching o-c-c review is in https://review.openstack.org/#/c/526201/ and is not passing due to Mr Chicken having not yet been introduced to Mr Egg. Actually, it is because -tips is voting in o-c-c and not SDK. Fortunately thisone is not needed to make the job pass... | |
| 15:34:58 | slaweq | dtroyer: hi again | |
| 15:35:10 | slaweq | dtroyer: please check once again https://review.openstack.org/515401 | |
| 15:35:35 | slaweq | I answered to Your comment there and I need Your opinion about way in which I should go with it now | |
| 15:35:38 | slaweq | thx in advance | |
| 17:11:15 | edleafe | cdent: dtantsur: https://etherpad.openstack.org/p/api-sig-newsletter is ready for your review | |
| 17:15:16 | cdent | edleafe: if you’re feeling pedantic, the review you mention in the new text could be listed in guidelines currently under review, but maybe you left that out on purpose since it was already mentioned? | |
| 17:17:05 | cdent | edleafe: so either gtg or add that, whichever you prefer | |
| 17:20:17 | edleafe | cdent: I thought about it, but figured it wasn't quite ready. But since the microversion review is also a WIP, I think I'll add it | |
| 17:20:46 | cdent | edleafe: coooooooool | |
| 17:23:24 | edleafe | sent | |
| 19:44:18 | bar | hello, I've been noticing race-conditions in octaviaclient tests when using fakes for generating resources, as we add more tests. Any way we can avoid that? | |
| 19:47:49 | dtroyer | bar: you have to be careful in test setup/teardown to make sure things are not shared between tests. class variables often are part of the problem, so is allocating things in a setUpClass method. | |
| 19:49:17 | bar | dtroyer, do you suggest to re-allocate our resources in class/function granularity? | |
| 19:51:46 | dtroyer | bar: I have found it to be east to go too far with refactoring/normalizing things here. remember the tests all may run in parallel and you can't guarantee order so each one ultimately has to keep track of its own resources. | |
| 19:52:08 | dtroyer | in OSC, the Identity functional tests are probably in the best shape regarding that balance | |
| 19:53:21 | bar | dtroyer, Identity functional tests? | |
| 19:54:52 | dtroyer | if you are talking about unit tests, those are generally easier to deal with but the basic problem is the same | |
| 19:54:52 | dtroyer | http://git.openstack.org/cgit/openstack/python-openstackclient/tree/openstackclient/tests/functional/identity/v3 | |
| 20:00:13 | bar | dtroyer, just to make sure, there are (A)no guarantees of the order that different functions in the same test file is called, and (B)different functions in the same class could be run in parallel? | |
| 20:08:17 | dtroyer | bar: correct. at least that is what I try to code for, anything less random in the test runner still works that way | |
| 20:09:10 | bar | dtroyer, thanks! | |
| 20:37:35 | mordred | dtroyer: woot! | |
| 21:17:55 | elmiko | edleafe: sorry i missed the meeting today, at a conference and totally spaced on the timezone conversion | |
| 21:24:35 | edleafe | elmiko: we were going to assign all the things to you | |
| 21:24:42 | edleafe | but there weren't any things | |
| 21:25:08 | openstackgerrit | Merged openstack-infra/shade master: Fix batching for floating ips and ports https://review.openstack.org/526127 | |
| 23:41:59 | openstackgerrit | Merged openstack/keystoneauth master: Make none auth usable in CLI https://review.openstack.org/515730 | |
| #openstack-sdks - 2017-12-08 | |||
| 00:52:04 | openstackgerrit | Yuki Nishiwaki proposed openstack/osc-lib master: Check auth_url as well as identity version api https://review.openstack.org/526413 | |
| 01:02:05 | openstackgerrit | Gilles Dubreuil proposed openstack/api-wg master: Adds API-Schema guideline https://review.openstack.org/524467 | |
| 04:54:08 | rm_work | anyone know why on one system when I run OSC with `endpoint list -h` it tells me optional arguments include `--service` and `--interface`, but on another system it doesn't? verified same version of python-openstackclient and python-keystoneclient | |
| 04:55:09 | rm_work | does it go out to keystone and check for version or something before it even returns the help? | |
| 04:55:55 | rm_work | ahhh i see, keystone v2/v3. nevermind | |
| 09:14:30 | openstackgerrit | Slawek Kaplonski proposed openstack/python-openstackclient master: Add support for get details of Quota https://review.openstack.org/515401 | |
| 09:17:32 | openstackgerrit | Slawek Kaplonski proposed openstack/python-openstackclient master: Add support for get details of Quota https://review.openstack.org/515401 | |
| 15:03:33 | TheJulia | mordred: When you have a minute, if you could take a look at https://review.openstack.org/#/c/526105 | |
| 15:05:46 | mordred | TheJulia: done | |
| 15:06:36 | TheJulia | Thanks! | |
| 15:30:39 | openstackgerrit | Monty Taylor proposed openstack/os-client-config master: Remove osc devstack tips jobs https://review.openstack.org/526705 | |
| 15:30:39 | openstackgerrit | Monty Taylor proposed openstack/os-client-config master: Make the get_service_type() overrides tolernat of no defaults https://review.openstack.org/526201 | |
| 15:30:40 | openstackgerrit | Monty Taylor proposed openstack/os-client-config master: Add osc tips jobs back https://review.openstack.org/526706 | |
| 15:30:58 | mordred | dtroyer: ^^ quick remove/re-add wrapped around your fix patch | |
| 15:31:11 | mordred | dtroyer: the re-add depends-on the 'make it voting in osc' patch | |
| 16:00:26 | frickler | dtroyer: the plugin check needs an update, because the zuul restart earlier changed the format of zuul.projects, could you have a look? http://logs.openstack.org/01/515401/10/check/openstackclient-check-plugins/d41b3ab/job-output.txt.gz#_2017-12-08_10_32_51_957615 | |
| 16:00:45 | frickler | it is a hash now like zuul._projects was earlier | |
| 16:01:36 | frickler | the good news is that the check now is really running on plugins, stumbled about that in python-designateclient | |
| 16:33:43 | openstackgerrit | Merged openstack-infra/shade master: Baremetal NIC list should return a list https://review.openstack.org/526105 | |
| 16:46:58 | dtroyer | mordred: kk, we have to work up the food chain to fix -tips, last I looked the SDK review needs another +2. I'll +A it anyway by noon to unblock everything else. | |
| 21:07:24 | openstackgerrit | Joshua Harlow proposed openstack-infra/shade master: Pass through all_projects for get_server https://review.openstack.org/526774 | |