| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2018-07-24 | |||
| 22:48:09 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Support --community in openstack image list https://review.openstack.org/565152 | |
| 23:10:52 | mordred | dtroyer: https://review.openstack.org/#/c/582334 <-- I left a novel of a comment there - I'm guessing we should probably just land it in that form since it's definitely an error it's fixing and leave making it more better to next cycle ... | |
| 23:11:05 | mordred | dtroyer: but whatever you think | |
| 23:17:07 | corvus | mordred: you said i could pass a dict to the connection constructor... how's that go again? | |
| 23:18:05 | corvus | just openstack.connect(**args) ? | |
| 23:18:57 | corvus | hrm, that gets me: keystoneauth1.exceptions.auth_plugins.MissingRequiredOptions: Auth plugin requires parameters which were not given: auth_url | |
| 23:19:48 | dtroyer | mordred: agreed, there are a number of microversion-affected changes to compute outstanding, some merged already, all of them are likely in this situation, this gets things working to some extent, we'll clean it up when we SDK-enable things and that may influence the order we do that work in... | |
| 23:26:54 | corvus | mordred: oh apparently that's because profile is ignored, so i need to add auth_url in there | |
| 23:28:41 | corvus | mordred: reading https://docs.openstack.org/openstacksdk/latest/user/transition_from_profile.html it's not clear how to accomplish what i want (construct a connection from a dict, using profile=vexxhost and not specifying auth_url) | |
| #openstack-sdks - 2018-07-25 | |||
| 00:42:32 | openstackgerrit | Merged openstack/python-openstackclient master: Support filtering port with IP address substring https://review.openstack.org/532324 | |
| 00:52:39 | openstackgerrit | Merged openstack/python-openstackclient master: compute: host: expand kwargs in host_set() call https://review.openstack.org/583225 | |
| 00:52:41 | openstackgerrit | Merged openstack/python-openstackclient master: Implement support for registered limits https://review.openstack.org/574475 | |
| 00:52:42 | openstackgerrit | Merged openstack/python-openstackclient master: Implement support for project limits https://review.openstack.org/575211 | |
| 01:19:08 | openstackgerrit | Artom Lifshitz proposed openstack/openstacksdk master: Send disk_over_commit if nova api < 2.25 https://review.openstack.org/585579 | |
| 02:49:03 | openstackgerrit | Merged openstack/python-openstackclient master: Support --community in openstack image list https://review.openstack.org/565152 | |
| 13:20:15 | openstackgerrit | Ryan Brady proposed openstack-infra/shade master: Add support for static routes https://review.openstack.org/580661 | |
| 14:32:36 | mordred | corvus: sorry about that ... profile is getting consumed by the old deprecated soon-to-be-removed old sdk profile | |
| 14:33:10 | mordred | corvus: the easiest way around it would be to construct a CloudRegion from a dict and pass that to the connection ctr | |
| 14:34:02 | corvus | mordred: it doesn't look like cloudregion would accept the clouds.yaml form of arguments: https://docs.openstack.org/openstacksdk/latest/user/config/reference.html#openstack.config.cloud_region.CloudRegion | |
| 14:34:10 | corvus | mordred: i don't see 'profile' or 'auth' there...? | |
| 14:35:22 | mordred | corvus: you are right. this is a bug | |
| 14:35:37 | corvus | mordred: what's the "openstack_config" argument for that? | |
| 14:36:31 | mordred | well, that's the bug - I do not think we support passing profile as an argument- itlooks like it's only procesed during config file reading | |
| 14:39:24 | mordred | corvus: oh no, it should work | |
| 14:39:28 | mordred | corvus: do this: | |
| 14:41:35 | mordred | corvus: from openstack.config import loader ; config = loader.OpenStackConfig().get_one(**args) ; conn = openstack.connection.Connection(config=config) | |
| 14:42:48 | corvus | mordred: still wants auth_url | |
| 14:43:32 | corvus | File \"/home/corvus/git/openstack-infra/zuul-jobs/.tox/py35/lib/python3.5/site-packages/keystoneauth1/loading/base.py\", line 162, in load_from_options\n raise exceptions.MissingRequiredOptions(missing_required)\nkeystoneauth1.exceptions.auth_plugins.MissingRequiredOptions: Auth plugin requires parameters which were not given: auth_url\n", | |
| 14:44:09 | corvus | that's from the get_one line | |
| 14:45:18 | mordred | poo. lemme poke further 0 sorry about that | |
| 14:48:39 | mordred | corvus: ok. I jave reproduced that lcaly | |
| 14:49:39 | dtantsur | folks, how do I deal with "The service catalog is empty" from session.get_endpoint of the service catalog is NOT empty? | |
| 14:49:50 | dtantsur | in what other cases can this happen? | |
| 14:50:02 | mordred | dtantsur: I've never seen that | |
| 14:50:08 | dtantsur | :( | |
| 14:50:19 | mordred | dtantsur: are you sure you have a token/catalog? | |
| 14:50:48 | dtantsur | mordred: the session is created with keystoneauth1.identity.v3.Token, the token comes from Mistral, no idea how it creates it | |
| 14:51:01 | dtantsur | it's tripleo, there is nothing I can be sure about :D | |
| 14:51:35 | mordred | dtantsur: it's possible the catalog got stripped from the token if it came from middleware | |
| 14:52:00 | dtantsur | mordred: the token - is it a string or an object? I think I get a string | |
| 14:52:01 | mordred | people like doing that to reduce size of the payload for reasons I don't really understand | |
| 14:52:20 | dtantsur | and now a funny thing: if I wrap it in an Adapter, it seems to work Oo | |
| 14:52:29 | mordred | weird | |
| 14:53:54 | dtantsur | extremely weird | |
| 14:57:02 | dtantsur | so, this apparently blows up https://github.com/openstack/keystoneauth/blob/master/keystoneauth1/identity/base.py#L275 | |
| 15:02:16 | openstackgerrit | Monty Taylor proposed openstack/openstacksdk master: Support passing profile to get_one https://review.openstack.org/585780 | |
| 15:02:24 | mordred | corvus: ^^ | |
| 15:04:06 | dtantsur | I guess I short report it against keystoneauth. it should not report EmptyCatalog if it's not empty in reality. | |
| 15:04:24 | corvus | mordred: that works, thanks! | |
| 15:04:38 | dtantsur | s/short/should/ (WUT) | |
| 15:05:13 | mordred | corvus: yay! | |
| 15:13:37 | dtantsur | filed https://bugs.launchpad.net/keystoneauth/+bug/1783590 | |
| 15:13:38 | openstack | Launchpad bug 1783590 in keystoneauth "EmptyCatalog raised when the catalog is not empty" [Undecided,New] | |
| 16:10:59 | dtantsur | mordred: do you know if I can pass the catalog to keystoneauth somehow? given that Token auth is, well, broken apparently in this regard | |
| 16:12:52 | dtantsur | ok, never mind, it just always treats the provided token as unscoped.. | |
| 17:04:26 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Add DNS support to floating IP commands https://review.openstack.org/558824 | |
| 17:06:19 | dtroyer | mordred: if you're happy with the conversation, https://review.openstack.org/#/c/582334/ I think is ready | |
| 17:07:13 | dtroyer | that and https://review.openstack.org/#/c/558824/ are the two I think we want to wait on for the release…I'm doing a last pass through release notes yet | |
| 17:49:57 | mordred | dtroyer: +A on that one | |
| 17:51:37 | dtroyer | mordred: Thanks. did you intend https://review.openstack.org/585030 to go in now too? I just rechecked it... | |
| 17:52:14 | mordred | dtroyer: no - I mean, I wish - but we need a keystoneauth prelease to pick uit up | |
| 17:52:26 | dtroyer | ah, right. ok | |
| 17:52:32 | mordred | silly freezes | |
| 17:52:35 | mordred | ;) | |
| 17:52:57 | mordred | tbh - I've been meaning to write the ksa patch for weeks, so it's totally my fault | |
| 17:57:45 | openstackgerrit | Merged openstack/python-openstackclient master: Add command to show all service versions https://review.openstack.org/584905 | |
| 19:20:27 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Release note cleanup for 3.16.0 release https://review.openstack.org/585853 | |
| 19:31:35 | openstackgerrit | Merged openstack/python-openstackclient master: Don't sent disk_over_commit if nova api > 2.24 https://review.openstack.org/582334 | |
| 19:58:50 | dtroyer | mordred: https://review.openstack.org/#/c/585853/ is the relnote cleanup and likely 3.16.0 release commit, ready to go… | |
| 20:07:57 | mordred | love it | |
| 20:36:34 | openstackgerrit | Merged openstack/python-openstackclient master: Release note cleanup for 3.16.0 release https://review.openstack.org/585853 | |
| 20:56:38 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Add DNS support to floating IP commands https://review.openstack.org/558824 | |
| 21:04:59 | mordred | dtroyer: woot. now I guess for next cycle there's no excuse I can give for not figuring out the remaining osc/sdk bits | |
| #openstack-sdks - 2018-07-26 | |||
| 01:51:08 | openstackgerrit | Nguyen Hai proposed openstack/openstackclient master: add lower-constraints job https://review.openstack.org/556037 | |
| 02:58:41 | openstackgerrit | Ryan Brady proposed openstack-infra/shade master: Add support for static routes https://review.openstack.org/580661 | |
| 05:02:39 | openstackgerrit | Nguyen Hai proposed openstack/openstackclient master: add lower-constraints job https://review.openstack.org/556037 | |
| 05:03:45 | openstackgerrit | Nguyen Hai proposed openstack/openstackclient master: add lower-constraints job https://review.openstack.org/556037 | |
| 07:11:51 | openstackgerrit | yanpuqing proposed openstack/python-openstackclient master: Add error message when network resources use "set" "unset" command https://review.openstack.org/579515 | |
| 07:24:24 | usr2033 | hi, does official openstack-sdk support quota operations? | |
| 08:58:15 | openstackgerrit | OpenStack Release Bot proposed openstack/python-openstackclient master: Update reno for stable/rocky https://review.openstack.org/586006 | |
| 09:01:02 | openstackgerrit | Josephine Seifert proposed openstack/python-openstackclient master: osc-included image signing (using openstacksdk) https://review.openstack.org/580086 | |
| 09:02:57 | openstackgerrit | Josephine Seifert proposed openstack/python-openstackclient master: osc-included image signing (using openstacksdk) https://review.openstack.org/580086 | |
| 11:54:44 | openstackgerrit | Ryan Brady proposed openstack-infra/shade master: Add support for static routes https://review.openstack.org/580661 | |
| 12:00:01 | openstackgerrit | Monty Taylor proposed openstack/openstacksdk master: Update create_object to handled chunked data https://review.openstack.org/585532 | |
| 12:34:05 | openstackgerrit | OpenStack Release Bot proposed openstack/keystoneauth master: Update reno for stable/rocky https://review.openstack.org/586083 | |
| 12:35:07 | openstackgerrit | OpenStack Release Bot proposed openstack/openstacksdk master: Update reno for stable/rocky https://review.openstack.org/586094 | |
| 12:35:37 | openstackgerrit | OpenStack Release Bot proposed openstack/os-client-config master: Update reno for stable/rocky https://review.openstack.org/586100 | |
| 12:36:32 | openstackgerrit | Monty Taylor proposed openstack/os-service-types master: Add flag for returning unofficial types https://review.openstack.org/585411 | |
| 12:36:33 | openstackgerrit | OpenStack Release Bot proposed openstack/osc-lib master: Update reno for stable/rocky https://review.openstack.org/586112 | |
| 12:48:32 | odyssey4me | mordred: I thought you might like to know a new and exciting and totally useful error I got the other day when I tried to do a compute thing with the openstack client. :) http://logs.openstack.org/15/582315/9/check/openstack-ansible-functional-ubuntu-xenial/911cdf7/job-output.txt.gz#_2018-07-21_09_14_00_043030 | |
| 12:49:09 | odyssey4me | It turns out, when you do not have an endpoint for the service, you get the rather oscure "'list' object has no attribute 'get'" error. | |
| 13:18:10 | mordred | odyssey4me: well! that certainly is exciting | |
| 13:24:33 | mordred | odyssey4me: I'm not 100% sure if that's coming from osc or novaclient- how hard would it be to recreate that with --debug ? | |
| 13:25:31 | odyssey4me | mordred: hmm, let me do that and get back to you - I might already have one up that I can break | |
| 13:25:55 | odyssey4me | oh bother, I don't - ok, let me kick off a build | |
| 13:36:14 | mordred | odyssey4me: woot. thanks! | |
| 16:07:40 | odyssey4me | mordred: it turns out that it's far more fun than anticipated - to replicate it you have to setup the compute endpoint with the wrong service backing it | |
| 16:08:09 | odyssey4me | in my test I put the placement service endpoints into the catalog, saying it was the compute service | |
| 16:08:11 | odyssey4me | https://gist.github.com/odyssey4me/fdfb2474bdd6fde4d1368af2af4044c6 | |
| 16:08:45 | odyssey4me | while of course this is not someone one would expect someone to do, it is perhaps better to handle the error somewhat more elegantly | |