| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2018-07-11 | |||
| 13:20:56 | mordred | gtema: if you feel like thinking about *really* fun things ... in https://review.openstack.org/#/c/574285/ dtantsur|afk mentions that TaskManager is very similar to the futurist library - and that maybe we should just add the ratelimiting support to futurist and switch shade/sdk/nodepool to using it | |
| 13:21:21 | mordred | gtema: I haven't dug in on that yet- as I should probably finish some of my hundred other tasks people are waiting on me for | |
| 13:22:09 | gtema | mordred: you'd bet. | |
| 13:23:28 | gtema | https://review.openstack.org/#/c/574285/ is nice | |
| 13:25:41 | mordred | gtema: :) I realized that shade/sdk support some logic for nodepool that's impossible to completely test in shade/sdk because the multi-threaded/rate-limiting taskmanager is not in tree | |
| 13:26:09 | mordred | gtema: so every time I try to talk to someone about improving/finishing the caching tier in shade/sdk it goes badly, since the whole problem space isn't in front of people | |
| 13:26:14 | mordred | whoops | |
| 13:27:33 | dhellmann | mordred , dtroyer : I've had someone ask me whether they should use shade or openstacksdk for a python app and in the course of trying to explain that things are merging I realized I don't actually know the state of that. What tool should someone use today? | |
| 13:30:41 | mordred | dhellmann: sdk | |
| 13:30:54 | mordred | dhellmann: they can use shade if they feel like it - it will continue to work indefinitely | |
| 13:31:20 | mordred | dhellmann: but the shade code is in sdk, and https://review.openstack.org/577647 (above, should land today) makdes the shade main class a subclass of the sdk class | |
| 13:31:39 | dhellmann | mordred : ok, cool, I couldn't even remember which direction things were merging :-) | |
| 13:31:44 | mordred | dhellmann: the "Connection" class in sdk contains all of the shade methods | |
| 13:31:46 | mordred | dhellmann: :) | |
| 13:31:48 | dhellmann | ETOOMANYPROJECTS | |
| 13:31:59 | mordred | dhellmann: I'm *very* much looking forward to finishing off collapsing the codebasees | |
| 13:32:40 | dhellmann | I'm sure | |
| 13:56:24 | openstackgerrit | Merged openstack/python-openstackclient master: Fix lower-constraints.txt https://review.openstack.org/581603 | |
| 14:13:46 | openstackgerrit | Yuki Nishiwaki proposed openstack/osc-lib master: Check auth_url as well as identity version api https://review.openstack.org/526413 | |
| 14:17:32 | openstackgerrit | Merged openstack/openstacksdk master: Fix Magnum cluster update https://review.openstack.org/581157 | |
| 14:22:52 | openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Use the openstacksdk task_manager constructor argument https://review.openstack.org/580475 | |
| 14:22:52 | openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Make OpenStackCloud a subclass of Connection https://review.openstack.org/577647 | |
| 14:29:12 | openstackgerrit | Merged openstack/openstacksdk master: Add Magnum /certificates support https://review.openstack.org/581598 | |
| 14:54:23 | openstackgerrit | Monty Taylor proposed openstack/openstacksdk master: Add task manager parameter to Connection https://review.openstack.org/577616 | |
| 14:57:34 | openstackgerrit | Monty Taylor proposed openstack/openstacksdk master: Add task manager parameter to Connection https://review.openstack.org/577616 | |
| 16:02:39 | openstackgerrit | Merged openstack/python-openstackclient master: Replace pbr autodoc with sphinxcontrib-apidoc https://review.openstack.org/581415 | |
| 16:02:40 | openstackgerrit | Merged openstack/python-openstackclient master: Skip calls to glance and nova when got no servers https://review.openstack.org/568344 | |
| 16:08:52 | openstackgerrit | Merged openstack/openstacksdk master: Support to wait for load balancer to be ACTIVE https://review.openstack.org/581198 | |
| 17:16:29 | Shrews | mordred: re: 577647 and list_flavors() comment, could we undecorate the shade method and depend on sdk's caching? | |
| 17:17:25 | Shrews | seems weird that shade has to know that sdk decorates that method and we need to manipulate the call | |
| 17:18:42 | Shrews | and fragile, to boot | |
| 17:24:28 | Shrews | also posted a comment with that question | |
| 18:24:05 | mordred | Shrews: yes - I think we can - but I think we need to unwind more first - like - I think we should probably unwind the caching completely so that there are no cache decorators in shade at all | |
| 18:33:40 | Shrews | mordred: i guess i don't see why we can't do it as we convert the methods (list_flavors here) | |
| 18:36:25 | Shrews | wait... if we just delete shade's list_flavors, wouldn't that work? | |
| 21:17:28 | openstackgerrit | Daniel Speichert proposed openstack/osc-lib master: Fixes a bug with issubclass() being called on an instance https://review.openstack.org/581886 | |
| 23:41:14 | mordred | Shrews: yes- except - the issue is that the list_flavors in shade needs to have different default values for the kwarg get_extra | |
| 23:41:53 | mordred | Shrews: what we probably want to do is make a patch that deletes all the methods except for the ones where we want to change arguments | |
| 23:42:09 | mordred | but I keep avoiding that because it hurts my head | |
| 23:42:28 | mordred | Shrews: once we do that though, I think we can also rip out the shade-side caching and the code will be *much* nicer | |
| #openstack-sdks - 2018-07-12 | |||
| 02:29:28 | openstackgerrit | Merged openstack/openstacksdk master: Add task manager parameter to Connection https://review.openstack.org/577616 | |
| 07:26:33 | openstackgerrit | Benoît Knecht proposed openstack/openstacksdk master: openstackcloud: properly handle scheduler_hints https://review.openstack.org/581984 | |
| 15:02:28 | openstackgerrit | Merged openstack/openstacksdk master: Handle Munch objects in proxies https://review.openstack.org/579908 | |
| 16:17:17 | openstackgerrit | Merged openstack/api-wg master: Expand error code document to expect clarity https://review.openstack.org/577118 | |
| 16:17:58 | openstackgerrit | Merged openstack/api-wg master: Switch to stestr https://review.openstack.org/581323 | |
| 16:20:40 | elmiko | edleafe: ok, ready for editing ;) https://etherpad.openstack.org/p/api-sig-newsletter | |
| 16:29:09 | edleafe | elmiko: looking... | |
| 16:30:35 | edleafe | elmiko: couple of minor nits corrected, and now LGTM | |
| 16:30:45 | elmiko | sweet, thanks! | |
| 17:07:10 | openstackgerrit | Merged openstack/openstacksdk master: openstackcloud: properly handle scheduler_hints https://review.openstack.org/581984 | |
| 17:50:58 | vakuznet | why http://paste.openstack.org/show/725745/ Is this microversion negotiation? | |
| 17:52:57 | vakuznet | drop v2.1 or add -H X-Auth-Token or ... ? | |
| 18:37:42 | openstackgerrit | Artom Lifshitz proposed openstack/python-openstackclient master: Don't sent disk_over_commit if nova api > 2.24 https://review.openstack.org/582334 | |
| 21:17:43 | openstackgerrit | Merged openstack/openstacksdk master: Add vip_qos_policy_id options for loadbalancer https://review.openstack.org/579794 | |
| #openstack-sdks - 2018-07-13 | |||
| 00:20:18 | openstackgerrit | Merged openstack/osc-lib master: Fixes a bug with issubclass() being called on an instance https://review.openstack.org/581886 | |
| 00:20:31 | openstackgerrit | Merged openstack/osc-lib master: Replace autodoc with sphinxcontrib-apidoc https://review.openstack.org/581538 | |
| 01:09:35 | dtroyer | mordred: osc-lib release 1.11.0: https://review.openstack.org/582434, unless you think we should wait a couple of days for https://review.openstack.org/526413 | |
| 07:30:38 | openstackgerrit | Rui Chen proposed openstack/openstacksdk master: Run ansible tests against specific public cloud https://review.openstack.org/582459 | |
| 07:30:52 | openstackgerrit | Benoît Knecht proposed openstack/openstacksdk master: meta: don't throw KeyError on misconfigured floating IPs https://review.openstack.org/582460 | |
| 08:22:05 | openstackgerrit | Rui Chen proposed openstack/openstacksdk master: Run ansible tests against specific public cloud https://review.openstack.org/582459 | |
| 08:59:50 | openstackgerrit | Rui Chen proposed openstack/openstacksdk master: Run ansible tests against specific public cloud https://review.openstack.org/582459 | |
| 11:00:44 | mordred | dtroyer: osc-lib release ++ | |
| 11:01:00 | mordred | dtroyer: I've rechecked https://review.openstack.org/#/c/580901 | |
| 11:18:02 | openstackgerrit | Monty Taylor proposed openstack-infra/shade master: Make OpenStackCloud a subclass of Connection https://review.openstack.org/577647 | |
| 12:30:47 | openstackgerrit | Monty Taylor proposed openstack/python-openstackclient master: Remove local openstack.config overrides https://review.openstack.org/580901 | |
| 12:30:58 | mordred | dtroyer: ^^ that's still not going to work - but we're closer | |
| 12:40:44 | gtema | mordred: in openstack-service-broker I have faced interesting case. I have both clouds.yaml and secure.yaml not on FS, but already as string | |
| 12:40:54 | gtema | don't ask me why :-) | |
| 12:41:18 | gtema | but sdk is not able to get it this way. I do not want really to flush files to FS | |
| 12:41:49 | gtema | would you mind if I extend sdk for getting clouds/secure.yaml content from parameters? | |
| 13:03:46 | mordred | hrm. | |
| 13:04:21 | mordred | gtema: what an interesting thing - I guess it's not exposing them as ConfigMap volumes then? | |
| 13:05:02 | gtema | well, that is currently my problem. due to the APB design I can't get them as ConfigMap (at least so far) | |
| 13:05:11 | mordred | gtema: I'm not opposed to it - we added a similar (but different) thing in ansible so thatyou can pass an entire cloud dict to the cloud: parameter | |
| 13:05:14 | gtema | so those can be only retrieved as variables | |
| 13:05:42 | mordred | oh -if they're ansible variables already, you should be able to pass them to the os_ modules now in the cloud: parameter | |
| 13:06:06 | mordred | but - if you need additional support in sdk, I don't think it's a problem | |
| 13:06:07 | gtema | there is `auth` param for ansible modules, but I need to extract correct one from clouds.yaml struct and merge with secure.yaml | |
| 13:06:32 | mordred | nod | |
| 13:07:27 | gtema | might be an option to create ansible module in this project only, that will parse clouds/secure, merge them and return ready `auth`, but this is not nice either | |
| 14:17:25 | Miouge | How can I specify the equivalent of the—compute-api-version flag with the openstacksdk? | |
| 14:28:06 | mordred | Miouge: you can set compute_api_version in clouds.yaml, or you can pass compute_api_version as a kwarg to the constructor | |
| 14:28:26 | mordred | however, you shouldn't really ever need to set it for anything | |
| 14:28:45 | mordred | and if you do, unless you're using the low-level http/rest layer, it's a bug I'd love to see reported | |
| 15:37:54 | openstackgerrit | Merged openstack/python-openstackclient master: Fix docs from I0dc80bee3ba6ff4ec8cc3fc113b6de7807e0bf2a https://review.openstack.org/578952 | |
| 16:52:58 | Miouge | mordred: I’m getting errors without that flag for soft-anti-affinity server group policies, the doc says to pass --os-compute-api-version 2.15 (https://docs.openstack.org/python-openstackclient/latest/cli/command-objects/server-group.html) but that’s not super clear how to do it with os_server_group Ansible module. Terraform has the same problem: https://github.com/terraform-providers/terraform-provider-openstack/issues/118 | |
| 16:54:46 | mordred | Miouge: WELL - those are docs for python-openstackclient. the intent in openstacksdk and os_server_group ansible module is that microversion negotiation not be exposed in that way .... | |
| 16:55:19 | mordred | Miouge: so it sounds like what we need to do it update sdk's server group policies to try to use 2.15 is somoene passes soft-anti-affinity | |
| 16:55:53 | mordred | Miouge: I can probably get that fixed for you reasonably quickly - although I have to go pick someone up from the airport | |
| 16:55:58 | mordred | so it might be a few hours | |
| 16:56:48 | Miouge | Cool, I’de be happy to give it a +1 :) | |
| 17:57:26 | openstackgerrit | Merged openstack/python-openstackclient master: Slow down and retry aggregate create/delete to lessen race https://review.openstack.org/580732 | |
| 18:05:44 | Miouge | mordred: Mmm I think I solved this issue by upgrading to latest (0.16) openstacksdk. My bad! | |
| 18:29:38 | mordred | Miouge: \o/ | |
| 19:11:08 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Fix error with image show when image name is None https://review.openstack.org/529464 | |
| 22:08:03 | flwang1 | mordred: hello, pls help review the ansible PR for Magnum cluster templates https://github.com/ansible/ansible/pull/42654 thanks | |
| #openstack-sdks - 2018-07-14 | |||
| 12:48:42 | mordred | flwang1: done! | |
| 18:44:10 | flwang1 | mordred: cheers | |
| #openstack-sdks - 2018-07-15 | |||
| 22:08:24 | openstackgerrit | Merged openstack-infra/shade master: Improve Magnum cluster templates functions https://review.openstack.org/577342 | |
| #openstack-sdks - 2018-07-16 | |||
| 02:34:11 | openstackgerrit | chenming proposed openstack/python-openstackclient master: test https://review.openstack.org/582821 | |