Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2018-01-31
15:38:20 mnaser so trying to do get_volume() in the admin tenant results in no volume found
15:38:43 mnaser is there some switch i'm not turning on?
15:38:59 mordred mnaser: yes - that is the default behavior. there is a constructor flag "use_direct_get" which will cause it to use get instead of list for gets
15:39:15 mnaser mordred: wonderful, that's exactly what i needed, thank you
15:40:01 mnaser hmm
15:40:36 mordred mnaser: I believe we'll flip the value of that default in the openstacksdk version of shade - we filter list calls to support nodepool, which I promise you do not want doing 1000s of get calls ... but for humans who are not nodepool it can be unexpected behavior :)
15:41:06 mnaser mordred: i think that flag might have been lost somehow
15:41:17 openstackgerrit Merged openstack-infra/shade master: Add devel branches and override-checkout for ansible-devel job https://review.openstack.org/536159
15:41:22 mnaser it seems like it is not a constructor value but a thing to set as an attribute to cloud
15:41:28 mnaser (aka cloud.use_direct_get = True)
15:41:38 mordred mnaser: I see it on the constructor?
15:41:54 mnaser note to self don't trust github search
15:42:16 mordred mnaser: http://git.openstack.org/cgit/openstack-infra/shade/tree/shade/openstackcloud.py#n137
15:42:21 mordred mnaser: never trust github
15:42:37 mnaser it doesn't seem to be in the release shipped with rdo pike, maybe i'll ask for a bump sometime soon
15:42:51 mordred oh - yah - it's a new flag
15:43:00 mordred got implemented this cycle
15:43:15 mnaser shade==1.22.2 and this merged in 1.24.0
15:43:21 mnaser ok cool venv it is
15:43:27 mordred \o/
15:43:43 mordred mnaser: luckily 1.24.0 is post-client-ectomy, so the venv should be quite small
15:43:53 mnaser oh awesome
15:45:44 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Shift voting flag and test_matrix_branch for ansible-devel job https://review.openstack.org/539563
15:46:19 tosky client-ectomy++
15:47:40 mordred tosky: ++ seriously. so much better
15:48:39 tosky and it helps packagers
15:51:02 mnaser ok slip up
15:51:11 mnaser shade.openstack_cloud != shade.OpenStackCloud
15:51:37 mnaser that got me confused for a while why it was still false
16:13:14 mordred mnaser: oh - well piddle. we should add that parameter to the shade.openstack_cloud constructor
16:14:24 mnaser mordred: and looks like get_server doesn't respect use_direct_get because it uses the find function all the time, but get_server_by_id also has 2 extra calls that get fips and ports
16:14:36 mnaser (i dont have time to look into it right now just fyi)
16:16:30 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Plumb use-direct-get through factory functions https://review.openstack.org/539586
16:17:53 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Plumb use-direct-get through factory functions https://review.openstack.org/539586
16:17:53 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Switch to providing created_at field for servers https://review.openstack.org/536003
16:18:02 mordred mnaser: ^^ there you go - I'll look at servers real quick
16:44:54 openstackgerrit Monty Taylor proposed openstack/os-client-config master: Updated from global requirements https://review.openstack.org/533991
16:44:55 openstackgerrit Monty Taylor proposed openstack/os-client-config master: Split docs requirements and update tox.ini https://review.openstack.org/539594
17:28:40 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Remove openstack_cloud factory function https://review.openstack.org/539395
17:28:40 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Merge Connection and OpenStackCloud https://review.openstack.org/539394
17:37:05 openstackgerrit Monty Taylor proposed openstack/os-client-config master: Split docs requirements and update tox.ini https://review.openstack.org/539594
17:37:06 openstackgerrit Monty Taylor proposed openstack/os-client-config master: Updated from global requirements https://review.openstack.org/533991
17:52:04 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Split request logging into four different loggers https://review.openstack.org/505764
18:03:35 openstackgerrit Merged openstack-infra/shade master: Shift voting flag and test_matrix_branch for ansible-devel job https://review.openstack.org/539563
20:03:35 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Shift config exceptions to openstack.exceptions https://review.openstack.org/539396
20:03:40 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Cleanup the BaseProxy docs a bit https://review.openstack.org/539270
20:03:45 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Rename BaseProxy to Proxy https://review.openstack.org/539271
20:08:20 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Fix get_server to work with use_direct_get https://review.openstack.org/539663
20:08:20 openstackgerrit Monty Taylor proposed openstack-infra/shade master: Plumb use-direct-get through factory functions https://review.openstack.org/539586
20:08:27 mordred mnaser: ^^ ok. those should work for you now
20:52:07 openstackgerrit Monty Taylor proposed openstack/python-openstacksdk master: Add get_client_config method to CloudRegion https://review.openstack.org/539678
20:59:32 openstackgerrit Merged openstack-infra/shade master: Switch to providing created_at field for servers https://review.openstack.org/536003
20:59:45 mordred dtroyer: feel like merging https://review.openstack.org/#/c/525283/ ?
21:01:50 dtroyer mordred: ugh, I didn't do that yet???
21:02:02 dtroyer done
21:02:19 adriant mordred: I'm look at https://bugs.launchpad.net/python-openstacksdk/+bug/1746535
21:02:20 openstack Launchpad bug 1746535 in OpenStack SDK "`openstack network list --external` breaks with openstacksdk 0.11.0 + openstackclient 3.14.0" [Undecided,New]
21:02:52 adriant it was introduced by my patch, but... what worries me is this likely means that that never actually worked.
21:03:10 adriant my patch just exposes the error rather than drops the param on the floor
21:03:28 adriant although, it is also possible my patch's logic is busted, so that is an option
21:03:28 openstackgerrit Merged openstack/os-client-config master: Add betacloud to the vendors https://review.openstack.org/527248
21:03:31 openstackgerrit Tim Burke proposed openstack/python-openstacksdk master: Let enforcer.py work under both py2 and py3 https://review.openstack.org/539685
21:05:43 mordred timburke: ^^ thanks. I could have sworn I'd done that already, but clearly I'm smoking something
21:06:29 mordred adriant: oh - piddle - I said this morning I'd look at that and totally forgot
21:06:34 timburke mordred: np. i saw https://review.openstack.org/#/c/525642/ but for some reason it got abandoned
21:07:08 mordred timburke: oh. it was stacked on another thing that got abandoned ... who knows.
21:09:08 adriant mordred: looks my patch is at fault. will play with it and submit a fix.
21:10:35 mordred adriant: cool. there's a followup to that that we do in shade that I've been meaning to port over - I'll do that on top of your patch
21:10:46 mordred adriant: luckily it's easy enough to reproduce ...
21:10:48 adriant oh...
21:10:54 adriant I think I see the issue
21:10:59 adriant it isn't technically my code
21:11:18 mordred can we blame you anyway thought? :)
21:11:20 adriant mordred: https://github.com/openstack/python-openstackclient/blob/master/openstackclient/network/v2/network.py#L500-L501
21:11:33 adriant the client supplies both
21:11:38 adriant but only the second does anything
21:12:11 mordred adriant: AH
21:12:12 adriant my code simply ensures that you're supplying it as: "is_router_external" since it is dropped otherwise
21:12:40 adriant lemme double check
21:12:53 mordred so - yah- introducing the strictness broke a thing that was just passing multiple options and counting on unsupported things falling on the floor
21:13:45 adriant we can potentially introduce a "strict_list_params" boolean if we want, and default it to false, with a deprecation warning?
21:14:38 mordred yah. I'd love it if we could introduce that on the Connection and not on each call ...
21:14:59 adriant oh yeah, no on each call...
21:15:19 adriant not*
21:15:53 mordred also - I may be overthinking this ... but at least in the openstackclient case, the extra argument is the 'real' name for something we have a mapping for
21:16:12 adriant patch for openstackclient?
21:16:15 mordred adriant: we *could* have it do something similar to what your swift patch did a while back (look at the keys and then also the values)
21:16:26 adriant yeah, I can add that in
21:16:50 adriant but I bet you someone somewhere is passing in bogus values... but my worry is they are doing so expecting it to work
21:16:53 mordred thatwould let people use the REST-side parameters or the sdk parameters
21:17:18 adriant oh, so we do the remapping if they supply the values?
21:17:30 adriant yeah, could work
21:18:07 mordred yah. we've already got is_router_external='router:external' in sdk
21:18:31 adriant ok, lemme play with that
21:18:43 mordred so if we're just add a friendly mapping for folks before the set - then router:external would turn in to is_router_external twice- and then set would collapse it and OSC would work again
21:18:45 mordred cool
21:19:20 adriant and anyone beyond that is actually supplying bogus values does get an error, but... that's probably a good thing
21:19:46 mordred agree
21:22:31 adriant mordred: ok, now I'm confused... what branch are we working against? because it looks like you've merged stuff to stable queens but not master
21:22:52 mordred both - go ahead and write it for master and we can then cherry-pick it back to stable queens

Earlier   Later