| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2018-01-03 | |||
| 15:26:33 | Shrews | cmurphy: yeah, i had issues with that combo | |
| 16:52:46 | openstackgerrit | Merged openstack/python-openstackclient master: Update new documentation PTI jobs https://review.openstack.org/530378 | |
| 16:55:52 | openstackgerrit | Merged openstack/osc-lib master: Fix sorting in Python 3 https://review.openstack.org/526046 | |
| 17:26:46 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: Update for new docs PTI https://review.openstack.org/530978 | |
| 21:11:03 | openstackgerrit | Hongbin Lu proposed openstack/python-openstackclient master: Allow ports filtering with device_id https://review.openstack.org/526798 | |
| 21:13:34 | mordred | cdent: you know everything, right? | |
| 21:14:19 | mordred | cdent: do you happen to know if there are any openstack services that paginate resources with limit/marker but do not return a {foo}_links dict with next links in it? | |
| 21:17:58 | mordred | cdent: the question comes up because of https://review.openstack.org/#/c/530769 and briancurtin's comments on it - I'm gonna go look and see what I can learn, but I figured if anyone happened to know off the top of their head about outliers that it would be helpful | |
| 21:18:21 | mordred | edleafe, elmiko: ^^ in case you happen to know | |
| 21:19:52 | edleafe | mordred: I confess that I do not know of any | |
| 21:20:07 | edleafe | mordred: I also confess that I haven't researched that at all :) | |
| 21:53:12 | elmiko | mordred: i haven't looked into that, and it looks like we have not done a study across projects about it. if you end up poking around, any info would be appreciated =) | |
| 21:53:15 | elmiko | https://wiki.openstack.org/wiki/API_Working_Group/Current_Design | |
| 22:12:14 | openstackgerrit | Merged openstack/python-openstacksdk master: Start using Connection in openstack.cloud https://review.openstack.org/523259 | |
| 22:14:51 | openstackgerrit | Bar Elharar proposed openstack/osc-lib master: Suppress subTest() expected errors https://review.openstack.org/531025 | |
| 22:18:36 | openstackgerrit | Bar Elharar proposed openstack/osc-lib master: Suppress subTest() expected errors https://review.openstack.org/531025 | |
| 22:26:00 | openstackgerrit | Bar Elharar proposed openstack/osc-lib master: Suppress subTest() expected errors https://review.openstack.org/531025 | |
| 22:28:44 | mordred | edleafe, elmiko: thanks - I think I'm going to start with a mailing list message and see where it gets me | |
| 22:40:54 | cdent | mordred, edleafe, elmiko : I was elsewhere, catching up | |
| 22:42:25 | mordred | cdent, edleafe, elmiko: I've found at least one example of a service that does not return $foo_links with a next link to indicate that there are more results available ... swift | |
| 22:43:32 | mordred | although now the question is - is it *only* swift (and therefore something that could be overridden in the sdk just for swift) or are there more, in which case adding a pagination_type flag to Resource objects may be the right choice | |
| 22:43:47 | cdent | are you using _ as an indicator of object descent, like server.links | |
| 22:44:36 | cdent | because both the api-wg guidance and looking at nova I'm not seeing what you're talking about, mordred, but I'm also not sure I know what you're talking about | |
| 22:46:01 | mordred | cdent: I've found another - glance is also different ... BUT ... for instance, https://developer.openstack.org/api-guide/compute/paginated_collections.html | |
| 22:46:18 | cdent | and https://developer.openstack.org/api-guide/compute/links_and_references.html | |
| 22:46:31 | mordred | cdent: nova returns {"servers": [...], "servers_links": [..]} | |
| 22:46:56 | mordred | so for a server list one can check servers_links for a next link, and if it exists, follow it | |
| 22:47:13 | cdent | sigh | |
| 22:47:44 | cdent | nobody ever follows the guidance ( http://specs.openstack.org/openstack/api-wg/guidelines/links.html) and even the services aren't in sync with themselves | |
| 22:48:10 | mordred | cdent: but for glance, it's {"images": [...], "next": "http:..."} | |
| 22:48:18 | cdent | blargh | |
| 22:48:22 | mordred | cdent: well - these links are a little bit different than the per-object links | |
| 22:48:51 | cdent | the pagination guidance doesn't distinguish, it's suggest generic 'links': http://specs.openstack.org/openstack/api-wg/guidelines/pagination_filter_sort.html#pagination | |
| 22:48:52 | mordred | cdent: I actually *do* see the reference links pretty frequently and consistently - other than when they're broken because reverse-proxies aren't set up correctly :) | |
| 22:49:36 | cdent | mordred: oh, I didn't mean nobody follows links. I mean projects not developing consistently. | |
| 22:49:44 | mordred | oh - yah. | |
| 22:50:07 | cdent | the pagination guidance example on that last link (heh) uses just 'links' in a relatively sane collection example | |
| 22:50:10 | briancurtin | since approximately day one, and going on until approximately forever, unfortunately | |
| 22:50:12 | mordred | cdent: that api-wg guidance is nice - and at least structurally similar to what I see from cinder and nova (naming notwithstanding) | |
| 22:50:15 | cdent | so yeah, I would guess that briancurtin is right | |
| 22:50:27 | mordred | briancurtin wins | |
| 22:52:14 | mordred | briancurtin: so - since a quick look has shown at least 4 different mechanisms, I'm thinking I should rework that patch to add a resource-level config flag choose one, with the behavior defaulting to the current sdk inferance behavior | |
| 22:53:01 | mordred | briancurtin: so that as we know or understand that a given resource provides links guidance we can opt a resource in to that behavior, but there is a good default that should work in the other cases | |
| 22:53:05 | briancurtin | mordred: that’s probably reasonable. we originally thought to go something like that direction for times when we knew for sure we wouldn’t need that extra call, so go for it | |
| 22:53:12 | briancurtin | yep | |
| 22:53:19 | mordred | woot | |
| 22:53:49 | mordred | briancurtin, cdent: see - if all the API services were consistent, we'd be issing out on all of this fun! | |
| 23:03:01 | cdent | oh look at the time, goodnight | |
| 23:16:37 | mordred | briancurtin: while I'm bugging you - https://review.openstack.org/#/c/530769/1/openstack/resource2.py line 765 in the original code ... seems like a bug - it'll stop iteration if yielded is < limit - shouldn't that be yielded >= query_params['limit'] - or is my brain just too melty? | |
| 23:18:11 | mordred | briancurtin: also - line 748 - yielded is reset in each loop iteration - seems like it should be moved to above line 734 - and that line 767 should be query_params['limit'] = limit - yielded | |
| 23:19:14 | mordred | briancurtin: I'm going to keep it as is in this patch so we can fix it if needed in its own patch - but just wanted to double-check my reading of the logic there | |
| #openstack-sdks - 2018-01-04 | |||
| 01:08:08 | openstackgerrit | Jake Yip proposed openstack/python-openstackclient master: Fix error with image show when image name is None https://review.openstack.org/529464 | |
| 02:06:01 | liuyulong | hello python-openstackclient core team, can we get this https://review.openstack.org/#/c/517845/ merged, since neutron doc is depending on it. | |
| 02:06:29 | liuyulong | neutron doc patch: https://review.openstack.org/#/c/526033/ | |
| 03:07:47 | dtroyer | liuyulong: I'll be able to verify that my concerns were addressed in the morning, I'd like amotoki to check his too… | |
| 03:09:34 | liuyulong | dtroyer, OK, thank you. | |
| 06:56:21 | openstackgerrit | OpenStack Proposal Bot proposed openstack/python-openstackclient master: Updated from global requirements https://review.openstack.org/531089 | |
| 07:54:51 | amotoki | liuyulong: I am in new year holidays till today. I am aware of yours and will check it tomorrow as I mentioned in the neutron team meeting this week. | |
| 07:57:13 | liuyulong | amotoki: Great, thank you. | |
| 08:51:33 | dtantsur | cdent: no, I haven't received any links. also likely to skip today's meeting due to personal stuff | |
| 08:52:32 | cdent | dtantsur: no problem | |
| 17:06:25 | openstackgerrit | Mike Fedosin proposed openstack/python-openstackclient master: Check that Glance returns image data before processing it https://review.openstack.org/531201 | |
| 17:18:24 | elmiko | edleafe: draft ready, https://etherpad.openstack.org/p/api-sig-newsletter | |
| 17:20:24 | edleafe | elmiko: looking... | |
| 17:22:16 | edleafe | elmiko: LGTM | |
| 17:22:33 | cdent | lgtm too | |
| 17:22:45 | elmiko | thanks, sending =) | |
| 20:23:12 | openstackgerrit | Merged openstack/os-api-ref master: Updated from global requirements https://review.openstack.org/520469 | |
| 21:14:53 | openstackgerrit | Samuel Pilla proposed openstack/python-openstackclient master: Add project tags functionality https://review.openstack.org/481284 | |
| 22:58:06 | openstackgerrit | Merged openstack/python-openstacksdk master: Update the shade-merge document https://review.openstack.org/524391 | |
| #openstack-sdks - 2018-01-05 | |||
| 00:17:47 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: Update for new docs PTI https://review.openstack.org/530978 | |
| 00:17:47 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: Prefer links dicts for pagination https://review.openstack.org/530769 | |
| 00:17:48 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: Use sdk for list_servers https://review.openstack.org/530770 | |
| 00:17:48 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: WIP Make resource a dict subclass usable by shade layer https://review.openstack.org/530835 | |
| 00:17:49 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: Remove name from zuul project stanza https://review.openstack.org/531267 | |
| 00:17:49 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: Port wait_for_ methods to use iterate_timeout https://review.openstack.org/531268 | |
| 00:29:35 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: Use sdk for list_servers https://review.openstack.org/530770 | |
| 00:29:36 | openstackgerrit | Monty Taylor proposed openstack/python-openstacksdk master: WIP Make resource a dict subclass usable by shade layer https://review.openstack.org/530835 | |
| 03:46:14 | ldl | Is there any friend use the CentOS7.2 to develop openstack-sdk or openstack ? the official site docs is use Ubuntu as example. | |
| 09:41:21 | openstackgerrit | Merged openstack/python-openstackclient master: Allow ports filtering with device_id https://review.openstack.org/526798 | |
| 11:12:53 | openstackgerrit | Akihiro Motoki proposed openstack/osc-lib master: Relocate utils tests to match the main code https://review.openstack.org/529956 | |
| 13:24:11 | amotoki | dtroyer: dhellmann: can we cut a new release of cliff? | |
| 13:24:37 | amotoki | I would like to use the new sphinxext in neutronclient doc | |
| 16:09:55 | dhellmann | amotoki : good idea. do you want to propose it in openstack/releases? | |
| 16:10:19 | amotoki | dhellmann: sure. I will propose it | |
| 16:10:50 | dhellmann | we'll get that csv fix in this release, too | |
| 16:14:27 | jpich | Nice | |
| 18:18:31 | cdent | elmiko: that dog face is not the face of gabbi excitement. I think the dog just has resting tripping balls face | |
| 18:18:41 | cdent | he was perfectly relaxewd and chilled out | |
| 18:43:53 | openstackgerrit | Mike Fedosin proposed openstack/python-openstackclient master: Check that Glance returns image data before processing it https://review.openstack.org/531201 | |
| 18:49:24 | johnsom | Do we have a timeline on a new openstacksdk release? octavia-dashboard has some patches waiting on a release with https://review.openstack.org/#/c/522962/ in it. | |
| 22:40:21 | openstackgerrit | Michael Johnson proposed openstack/python-openstacksdk master: Add an octavia functional test gate https://review.openstack.org/531514 | |
| #openstack-sdks - 2018-01-06 | |||
| 00:28:04 | johnsom | mordred Yee master of all things zuul, http://git.openstack.org/cgit/openstack/python-openstacksdk/tree/.zuul.yaml#n69 and http://logs.openstack.org/14/531514/1/check/openstacksdk-functional-devstack-octavia/14a627c/job-output.txt.gz#_2018-01-05_23_58_10_164433 | |
| 00:28:40 | johnsom | We usually disable cinder as it's not needed and saves gate time. Not sure how to override that in the parent zuul config | |
| 00:32:13 | mordred | johnsom: looking - quick question ... any reason to not just add the octavia config to the other functional jobs? | |
| 00:32:33 | mordred | like - does it increase the memory footprint a bunch and/or is enabling it flaky in any way? | |
| 00:32:50 | johnsom | mordred Run time is the biggest reason we usually don't, but I'm open to it. It was before | |
| 00:33:02 | johnsom | It boots VMs | |
| 00:34:51 | mordred | johnsom: nod. well - I think we can get it going with a second job just to see - that might also make it nicer/easier for octavia to add the sdk functional test to octavia patches | |
| 00:34:56 | johnsom | My thought on this would just push that config down to jobs that need it | |