Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2019-02-26
10:15:08 openstackgerrit Dmitry Tantsur proposed openstack/openstacksdk master: baremetal: support network_data when building configdrive https://review.openstack.org/639149
10:33:17 openstackgerrit Surya Seetharaman proposed openstack/python-openstackclient master: API microversion 2.69: Handles Down Cells https://review.openstack.org/633579
10:39:02 openstackgerrit Dmitry Tantsur proposed openstack/openstacksdk master: baremetal: support network_data when building configdrive https://review.openstack.org/639149
10:51:30 gtema dtantsur: a chocolate is definitely well deserved:D
10:51:39 dtantsur nice :)
10:52:07 gtema I will pass it you when we meet in person ;-)
10:52:21 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/openstacksdk master: Add "name" filter in "list" call when retrieving a single register https://review.openstack.org/637238
10:52:33 dtantsur if I get a US visa (the interview is tomorrow, wish me luck), I may see you in Denver
10:52:43 dtantsur gtema: are you going to Denver?
10:53:12 gtema unfortunately not - company is a platin sponsor, but since no talks accepted only top management goes there
10:53:33 dtantsur even to the PTG?
10:53:35 gtema but good luck with the interview
10:53:37 gtema nope
10:53:44 dtantsur wow
10:54:01 dtantsur well, depending on your location, it will spare you a long or not so long flight
10:54:09 gtema gernamy
10:54:12 gtema Germany
10:54:23 gtema so not that short
10:54:23 dtantsur which part of it?
10:54:30 gtema central (NRW)
10:54:56 dtantsur nice part :) I'm in Berlin nowadays.
10:55:13 gtema cool, but to crowded for me
10:55:17 dtantsur are you coming for the ops meetup in March?
10:55:30 dtantsur heh, yeah. but I was born in Moscow, so not too crowded for me :)
10:55:41 gtema ha-ha-ha. Even being in T-systems - no.
10:56:21 dtantsur I see..
10:56:22 gtema I'm born in UA, lived in Kiev, so not too crowded for sure, but compared with Paderborn (150t) is already
10:56:48 dtantsur oh, so you also travelled to the West :)
10:56:54 gtema yupp
10:56:59 dtantsur cool
10:57:04 gtema "Go West"!!!
11:06:41 gtema mordred: do you want to cover comments from dtantsur for a choco, or should I do that?
11:07:39 mordred gtema: I'd like a choco ... lemme take a quick pass
11:08:06 gtema mordred: come to Germany and you will get one too
11:08:45 dtantsur ++
11:09:13 mordred I feel like february isn't the MOST fun time to come to germany
11:09:32 dtantsur depending on your definition of fun :D
11:09:33 gtema it's around 20°C now
11:09:47 dtantsur yeah, this spring came much earlier
11:09:47 gtema it's nearly spring here
11:10:17 mordred oh nice!
11:10:41 mordred well then maybe I should come to germany to get some nice weather
11:11:04 gtema yeah, heard already of weather in US :D
11:16:35 mordred dtantsur: oh goodie. your second comment opens up a whole giant can of worms
11:16:55 dtantsur which one was the second? :)
11:17:03 mordred dtantsur: the one after the first ;)
11:17:08 dtantsur ah, got it
11:17:10 dtantsur :D
11:17:16 mordred the one about avoiding the race by returning image_id
11:17:34 mordred I don't think we can do that - because the method should also work well for people calling it directly on the Proxy - which means we need to return an Image
11:17:36 mordred BUT
11:18:23 dtantsur well, for example, baremetal.list_node_vifs return port IDs without trying to convert them to ports (largely because somewhere on the Earth there may be a downstream ironic plugin where VIF IDs are not port IDs)
11:18:29 mordred we can't just return Image(image_id) - because we don't know what version of the image service we need
11:19:17 mordred so we'd need, I guess, to use the image proxy class to get an Image resource only populated with the image_id
11:19:37 mordred and then the shade layer - and other callers - can call wait_for_image on that Image if they want to
11:20:20 gtema image_id and probably name (it shouldn't be changed anyway)
11:21:18 mordred yeah
11:59:36 dtantsur lazy loading of other attributes anyone? :)
11:59:50 dtantsur I think having a half-populated Resource is even more confusing than returning image_id tbh
12:18:55 openstackgerrit Artem Goncharov proposed openstack/python-openstackclient master: add multiple attributes to image from APIv2.7 https://review.openstack.org/637528
13:37:29 holser__ dtantsur I reworked https://review.openstack.org/637169
13:37:46 dtantsur cool. I'm leaving now, can review on Thu
13:37:56 holser__ ok, thanks
15:52:00 tssurya I have a client side chane here which is sort of the last piece of a blueprint: https://review.openstack.org/#/c/633579/ , reviews would be really appreciated, thanks in advance!
15:52:10 tssurya s/chane/change
16:59:13 gtema dtroyer, amotoki osc question: what is the best way to make utils.find_resource find resource, if `list` should be called with a special flag. When image.os_hidden flag is set to true image can be found by name only with `list(os_hidden=True)` - https://review.openstack.org/#/c/637528/
17:24:32 cdent elmiko, dtantsur|afk, edleafe : I'm looking at my old api-design doc at https://review.openstack.org/#/c/592003/ and wondering where in my head I should prioritize revisiting it on a scale of never to asap. Thoughts?
17:31:46 edleafe cdent: While I think it's a worthwhile effort, it seems to me that the shared storage discussion could be applied here
17:33:06 cdent In my head it's sort of a "thing to do when I need to think about something else". But I'm not certain I will ever think about it well or enough to be useful beyond some very anodyne cliches.
17:33:40 cdent but yeah
17:33:50 cdent I think I'll set it aside
17:35:00 cdent done
17:45:57 edleafe That seems more like a topic for a book than a guideline of a few paragraphs
17:47:35 cdent you send me the huge advance, i'll write the book
17:59:17 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/openstacksdk master: Add "name" filter in "list" call when retrieving a single register https://review.openstack.org/637238
18:07:39 elmiko cdent: i generally agree with edleafe. while i think it would be a lovely doc to have, it could also become a never-ending windmill to tilt at.
18:08:08 elmiko so, i guess unless o'reilly shows up with a check and a bw picture of a cute animal...
18:08:35 elmiko although, i would _love_ to see an o'reilly book with a line art bagpus on it XD
18:09:27 elmiko *bagpuss
18:09:34 fungi so... for you sdk wonks, some of whom may also be storyboard users/fans, there is a feature in progress to store attachments in swift. the bulk of the backend interface implementation is proposed in https://review.openstack.org/633365 and i'm wondering if anyone would like to weigh in on unified python sdk vs swiftclient there
18:10:10 fungi mordred: ^ not sure if you're around but you often have strong opinions on things
18:12:43 fungi this initial version just authenticates directly to swift, so ought to work well with standalone/dedicated swift deployments. i don't know enough about the unified sdk to know whether it complicates standalone swift use cases
18:23:11 openstackgerrit Rajat Dhasmana proposed openstack/python-openstackclient master: Disabling c-backup service for osc-functional-devstack-tips job https://review.openstack.org/639410
18:26:03 openstackgerrit Rajat Dhasmana proposed openstack/python-openstackclient master: Fix: Restore output 'VolumeBackupsRestore' object is not iterable https://review.openstack.org/624860
19:04:47 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/openstacksdk master: Add "name" filter in "list" call when retrieving a single register https://review.openstack.org/637238
19:10:41 mordred fungi: I recommend using sdk - I feel that nobody should ever use the python-*client libraries for any purpose
19:11:11 mordred fungi: that said - I have never used it with standalone swift
19:34:45 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/openstacksdk master: Use mock context in test_fwaas https://review.openstack.org/639421
#openstack-sdks - 2019-02-27
06:35:24 openstackgerrit Rajat Dhasmana proposed openstack/python-openstackclient master: Fix: Restore output 'VolumeBackupsRestore' object is not iterable https://review.openstack.org/624860
06:56:33 openstackgerrit Kailun Qin proposed openstack/python-openstackclient master: Add network segment range command object https://review.openstack.org/625544
08:07:08 openstackgerrit Christian Schneemann proposed openstack/python-openstackclient master: Typo fix https://review.openstack.org/635471
09:33:14 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/openstacksdk master: Add "name" filter in "list" call when retrieving a single register https://review.openstack.org/637238
09:49:54 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Tweak find_image method to search in hidden images https://review.openstack.org/639613
09:57:48 openstackgerrit Slawek Kaplonski proposed openstack/openstacksdk master: Add agent property: resources-synced https://review.openstack.org/635143
12:37:49 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/openstacksdk master: Add "name" filter in "list" call when retrieving a single register https://review.openstack.org/637238
12:47:47 openstackgerrit Pavlo Shchelokovskyy proposed openstack/python-openstackclient master: Paginate over usage list to return all usages https://review.openstack.org/639652
13:28:50 openstackgerrit Merged openstack/openstacksdk master: Add agent property: resources-synced https://review.openstack.org/635143
14:31:53 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Add glance image import support https://review.openstack.org/639122
14:31:53 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Move image methods to sdk image proxy https://review.openstack.org/609684
14:31:56 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Use retriable_status_codes in image upload https://review.openstack.org/639679

Earlier   Later