| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2017-03-20 | |||
| 09:06:44 | reedip | yeah , the patch.object was missing :) | |
| 15:40:53 | openstackgerrit | Dean Troyer proposed openstack/python-openstackclient master: Make MAC address of port updatable https://review.openstack.org/442543 | |
| 15:42:36 | sshank | Can I please get reviews on patch: https://review.openstack.org/#/c/387611/ . I think its ready. Please look into it and comment. | |
| 15:42:45 | sshank | reedip, ^ | |
| 16:19:28 | lbragstad | o/ hey folks - curious if anyone would be able to provide some clarity on a question i have about the guidelines for tags http://specs.openstack.org/openstack/api-wg/guidelines/tags.html | |
| 16:20:17 | edleafe | lbragstad: We can try | |
| 16:20:41 | lbragstad | edleafe the modifying tags section eludes to some set limit of tags that can be modified in a single request - http://specs.openstack.org/openstack/api-wg/guidelines/tags.html#modifying-the-tag-list | |
| 16:21:22 | lbragstad | edleafe my first question is whether or not the API WG provides a baseline as to what that limit should be? | |
| 16:21:49 | lbragstad | edleafe my second question is if the API WG has any recommendations on the total number of tags a thing can have | |
| 16:21:56 | edleafe | lbragstad: it's not a design limit of tags; it's the limit of the web server handling the request | |
| 16:22:13 | edleafe | The number of tags isn't as important as the total size of the tags | |
| 16:22:32 | lbragstad | edleafe ah - sure, ok that makes sense | |
| 16:22:40 | lbragstad | edleafe so that suggestion is to protect against that case | |
| 16:22:53 | edleafe | IOW, if you have 100 10-byte tags, it will be better than 1 50GB tag :) | |
| 16:23:50 | lbragstad | right | |
| 16:24:29 | openstackgerrit | Nakul Dahiwade proposed openstack/python-openstackclient master: OSC Network Flavor Profile https://review.openstack.org/396783 | |
| 16:24:41 | lbragstad | edleafe we have a spec proposed to add tags to projects - https://review.openstack.org/#/c/431785 | |
| 16:24:46 | edleafe | The main point of that suggestion is the response code to return | |
| 16:25:44 | lbragstad | so is the limit referring to the size of the tag, or the total number of tags? | |
| 16:26:05 | lbragstad | because the example provided makes me think its the total size of the tags | |
| 16:26:15 | lbragstad | not limiting the actual number of tags | |
| 16:27:21 | edleafe | My understanding is that while the total size of all the tags is the limiting factory, an API developer might want to limit the number of tags allowed for more practical purposes | |
| 16:28:21 | lbragstad | ok | |
| 16:28:21 | edleafe | IOW, it would be nearly meaningless to assign 10K tags to a server | |
| 16:28:37 | lbragstad | right | |
| 16:28:56 | lbragstad | that kind of what i was getting at with my second question | |
| 16:29:16 | lbragstad | does it make sense to limit the total number of tags a thing can have (not just limiting the request) | |
| 16:29:40 | edleafe | lbragstad: my opinion is that it is dependent on the project, and the resource being tagged | |
| 16:30:05 | edleafe | I don't know of any API-WG stance on that | |
| 16:30:24 | lbragstad | edleafe ok - so for keystone, we could choose to limit the total number of tags a project can have, and we can choose to limi the total number of tags that can be modified in a single request | |
| 16:30:55 | edleafe | lbragstad: well, those should be the same | |
| 16:31:00 | lbragstad | I would think there would be a pretty good correlation between two things | |
| 16:31:03 | lbragstad | edleafe ++ | |
| 16:31:17 | edleafe | since a PUT has to contain the complete list | |
| 16:31:26 | edleafe | (not a PATCH) | |
| 16:31:29 | lbragstad | so my second question doesn't make sense based on the answer from the first :) | |
| 16:31:37 | edleafe | heh | |
| 16:31:52 | lbragstad | but what about the case where single tags are added? | |
| 16:32:51 | lbragstad | if i limit the total number of tags in a single request to 50, that shouldn't stop me from using PUT /servers/1234567890/tags/<random_tag> 50+ times | |
| 16:33:44 | edleafe | lbragstad: it wouldn't, but you have to keep in mind that GET /servers/1234567890/tags is a thing that will return the full list | |
| 16:34:44 | edleafe | So yeah, you can run individual PUTs as many times as you like, but you might have too big a body for the GET | |
| 16:35:07 | lbragstad | edleafe at which point you'd have to start pruning things based upon that list, right? | |
| 16:35:31 | lbragstad | because using PUT to modify the entire list would rewrite the whole thing | |
| 16:36:10 | edleafe | lbragstad: yes, PUT with a body replaces the current tags with the body contents | |
| 16:37:18 | lbragstad | edleafe so what i'm hearing is that it should be possible to exceed the total number of tags limited in a PUT request with individual tag management | |
| 16:38:00 | edleafe | lbragstad: unless you do server-side checking, yes | |
| 16:38:17 | lbragstad | edleafe is there any reason to *not* do server-side checking? | |
| 16:38:46 | lbragstad | (outside of extra validation) | |
| 16:39:02 | edleafe | that would require that the server do a GET to determine the number of current tags with each PUT request | |
| 16:39:27 | lbragstad | yeah - it would have to inspect what's already tagged | |
| 16:44:23 | lbragstad | edleafe thanks for answering my questions - i think i have enough to go do another review of the spec | |
| 16:51:08 | edleafe | lbragstad: yw | |
| 16:54:42 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: Use Sphinx 1.5 warning-is-error https://review.openstack.org/446629 | |
| 16:54:43 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: docs: Don't include ChangeLog https://review.openstack.org/447588 | |
| 16:54:43 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Remove local fork of apidoc https://review.openstack.org/447589 | |
| 16:54:43 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Fix option-describe typos https://review.openstack.org/447590 | |
| 16:54:44 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Correct option, envvar markup https://review.openstack.org/447591 | |
| 16:54:44 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: Don't duplicate commands https://review.openstack.org/447592 | |
| 16:54:45 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Option arguments are not options https://review.openstack.org/447593 | |
| 16:54:45 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Correct docstrings https://review.openstack.org/447594 | |
| 16:54:46 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Correct invalid indentation https://review.openstack.org/447595 | |
| 16:54:46 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Don't use murano plugin https://review.openstack.org/447596 | |
| 16:54:47 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Correct issues with network-topology spec https://review.openstack.org/447597 | |
| 16:54:47 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Include API modules in docs https://review.openstack.org/447598 | |
| 16:56:45 | dtroyer | so Stephen, ^^^^ could have been done as a couple of reviews that batch together similar things | |
| 17:08:42 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Remove local fork of apidoc https://review.openstack.org/447589 | |
| 17:08:43 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Correct option, envvar markup https://review.openstack.org/447591 | |
| 17:08:43 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Fix option-describe typos https://review.openstack.org/447590 | |
| 17:08:44 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: Don't duplicate commands https://review.openstack.org/447592 | |
| 17:57:38 | openstackgerrit | Merged openstack/python-openstackclient master: Adds missing flavor information in the server list long command https://review.openstack.org/444934 | |
| 17:58:32 | openstackgerrit | Merged openstack/python-openstackclient master: Add extra filtering options to qos policy list https://review.openstack.org/401131 | |
| 17:58:38 | openstackgerrit | Merged openstack/python-openstackclient master: Non-Admin can't list own projects https://review.openstack.org/376056 | |
| 18:51:41 | stevemar | dtroyer: i have questions on https://review.openstack.org/#/c/387611/55 -- you +2'ed, can you answer them in the authors absence ? | |
| 18:53:29 | ankur-gupta-f4 | stevemar: it is laying down a common command in case there are more agent commands down the line which add the network to an agent. | |
| 18:53:34 | ankur-gupta-f4 | hence the --dhcp flag | |
| 18:54:00 | stevemar | ankur-gupta-f4: oh okay | |
| 18:54:04 | ankur-gupta-f4 | Your comments about the help text needs to be addressed though | |
| 18:54:26 | ankur-gupta-f4 | I thought it had already been addressed | |
| 18:54:26 | stevemar | ankur-gupta-f4: fix them up and i'll approve it ;) | |
| 18:55:36 | ankur-gupta-f4 | k gimme a sec | |
| 19:02:09 | openstackgerrit | Ankur proposed openstack/python-openstackclient master: Introduce Neutron DHCP agent commands to OSC https://review.openstack.org/387611 | |
| 19:03:16 | ankur-gupta-f4 | stevemar: done. Double check my work though. It was hurried sitting at airport gate | |
| 19:05:28 | sshank | stevemar, The dhcp flag check was added to support any future agents needs to be extended. | |
| 19:45:19 | stevemar | sshank: gotcha | |
| 20:08:25 | reedip | sshank : checking | |
| 20:08:31 | reedip | o/ | |
| 20:11:47 | reedip | stevemar , ankur-gupta-f4, sshank : there is an issue with https://review.openstack.org/#/c/387611/55 | |
| 20:12:23 | reedip | not a big one though\ | |
| 20:12:31 | sshank | reedip, Whats it about? | |
| 20:12:55 | reedip | sshank : change the commit message plz | |
| 20:13:33 | reedip | sshank : if you are moving the DHCP commands to OSC, please add the DHCP flag to network agent add/list etc | |
| 20:13:48 | reedip | as mentioned in the https://review.openstack.org/#/c/387611/56/doc/source/command-objects/network-agent.rst | |
| 20:13:56 | ankur-gupta-f4 | I think it is a non-issue | |
| 20:14:21 | ankur-gupta-f4 | Since it is introducing a general command for adding agents to network | |
| 20:14:48 | ankur-gupta-f4 | Not just dhcp. Even though currently dhcp is the only agent type supported | |
| 20:15:08 | reedip | ankur-gupta-f4 : its not a big issue , but it does cause a bit of a misunderstanding in the inital review ... :) | |
| 20:15:59 | ankur-gupta-f4 | Up to cores. Im out till tmw. Adios amigos. | |
| 20:16:33 | reedip | hehe ... asta la vista ! | |
| 20:18:55 | reedip | sshank: couple of nits are there | |
| 20:19:04 | reedip | posting them up in another 5 min | |