Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2018-04-25
17:22:57 Shrews pabelanger: yep, look like it. awesome
17:34:23 openstackgerrit Sean McGinnis proposed openstack/python-openstackclient master: Remove invalid 'unlock-volume' migration arg https://review.openstack.org/558910
17:38:12 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Infer version from old versioned service type aliases https://review.openstack.org/564299
#openstack-sdks - 2018-04-26
01:50:15 adriant mordred: it's not that I keep re-solving them, I just get curious how stuff works and then dig into it myself :P I need to get better at checking for existing patches and bug reports :/
01:52:43 adriant although, mordred, I'm tempted to look at adding swift object chunking support to the SDK itself. We probably want to share that code, so either we find a common place for it, or the proxy uses the shade code, or we move that code into the service code. No clue what's nicer.
02:12:41 adriant that said, there is also the issue that shade doesn't let you pass in an open file stream so if the file NEEDS to be opened as binary to send it to swift... you can't: https://pastebin.com/BHMNUPF4
02:13:06 adriant and doing it with the sdk is 3 lines...
02:13:42 adriant trying to upload a gzip'd tar file with shade throws: UnicodeDecodeError: 'utf-8' codec can't decode byte 0x8b in position 1: invalid start byte
02:17:31 adriant will file a bug with a easy step by step guide to reproduce.
02:36:53 adriant mordred: https://storyboard.openstack.org/#!/story/2001918
11:03:18 mordred adriant: *awesome*
11:04:40 mordred adriant: re: chunking - (and other similar things where we have an advanced version of something in shade layer) - I've been thinking we should move the shade call into the sdk proxy layer and turn the shade call into a passthrough
11:05:32 mordred adriant: but I agree on no clue what's nicer - I figured trying one or two of them and seeing what we like ...
11:12:15 mordred adriant: hrm. looking at your bug - that's really weird - we use the shade code in nodepool to upload qcow images to rackspace ...
11:22:17 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Turn normalize_status into a class https://review.openstack.org/564110
11:22:19 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Infer version from old versioned service type aliases https://review.openstack.org/564299
11:22:20 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Make VersionData class https://review.openstack.org/564469
13:03:55 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Don't assume a full config dict https://review.openstack.org/564493
13:03:56 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Defer all endpoint discovery to keystoneauth https://review.openstack.org/564494
13:04:33 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Infer version from old versioned service type aliases https://review.openstack.org/564299
13:04:34 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Allow tuples and sets in interface list https://review.openstack.org/564495
13:12:55 pabelanger mordred: mind looking at https://review.openstack.org/564247/ again, volume limits for shade
13:16:57 mordred pabelanger: NO! NO REVIEW FOR YOU
13:18:04 mordred pabelanger: jk. +A
13:18:17 pabelanger Yay
13:23:26 mordred pabelanger: thanks for that
13:23:28 umbSublime_ when using the sdk's compute proxy compute.servers() method. Is there a way to get the count returned in the generator object without iterating over it ?
13:24:27 mordred umbSublime_: not really - the generator itself doesn't known a total count because of server-side pagination
13:24:37 umbSublime_ ahh I see
13:24:56 mordred umbSublime_: it would be a nice thing to have if it was possible :)
13:25:31 umbSublime_ indeed!
13:31:35 pabelanger mordred: np, the other issue we might need to do is some process to check for leaked volumes, like we did for FIPs in nodepool. I am sure you might have ideas on how to handle that
13:32:26 mordred pabelanger: I havethe beginnings of some
13:32:44 pabelanger cool
13:33:10 mordred pabelanger: do you know if our leaked volumes in vexxhost are still there?
13:33:58 pabelanger mordred: I think so, I haven't cleaned them up yet
13:37:36 mordred pabelanger: awesome. let me look at them as examples
13:38:03 mordred pabelanger: my biggest question is whetheror not they still show the old server id in their metadata
13:38:31 mordred pabelanger: (figuring out how to identify volumes that are safe to automatically delete is ... terrifying
13:38:45 mordred pabelanger: cause, you know, get that wrong and you could give someone a giant foot-cannon
13:40:28 pabelanger Agree, I too am hoping for some sort of metadata
13:42:33 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Add get_volume_limits() support https://review.openstack.org/564507
13:42:43 mordred pabelanger: ^^ there, that's cherry-picked to sdk
13:45:06 pabelanger mordred: ah, right. Development on openstack-infra/shade is ending?
13:48:04 mordred pabelanger: wel - not quite yet - but yeah, for now we're cherry-picking patches
13:48:35 mordred pabelanger: hopefully we'll be able to get far enough along this cycle to make shade a thin wrapper
13:48:57 pabelanger ack
13:59:19 umbSublime_ I remember talking here a few months ago about making a cutsom api request not supported by the SDK, at the time the work-around was to use the <ressource>._action method and pass the connection.session object. I was discussed that new standard way would be eventualy implemented in the SDK. Was this added since ?
13:59:29 mugsie is there any easy way to take a path to openstack-client-config file in openstack-sdk?
13:59:54 mugsie (not using a env var)
14:33:49 mordred mugsie: you want to override the location of the clouds.yaml file?
14:34:23 mugsie yeah - I am writing a certbot plugin, and the location of the clouds.yaml file may be $RANDOM_DIR
14:34:48 mordred mugsie: the config_files parameter to openstack.config.loader.OpenStackConfig is what you want
14:35:20 mordred mugsie: there's some examples of usage in the test suite - one sec, lemme find you a reference
14:35:33 mugsie mordred: sweet - thanks
14:35:51 mordred mugsie: actualy - openstack/cloud/inventory.py and look for config_files=
14:37:07 mordred mugsie: but you;ll need to do config = loader.OpenStackConfig( ... then cloud_region = config.get_one( ... then conn = connection.Connection(config=cloud_region)
14:37:35 mugsie yup, looks doable - thanks :)
14:50:08 openstackgerrit Merged openstack/openstacksdk master: Fix openstack-inventory https://review.openstack.org/563754
14:58:27 mordred samP: remote: https://review.openstack.org/564527 Release openstacksdk 0.13.0
15:00:07 Shrews pabelanger: did you ping someone about the bifrost jobs fix? can't merge your shade change w/o it
15:00:44 mordred Shrews: oh - I thought I saw that was fixed - do we need to get something landed?
15:01:05 Shrews mordred: yeah. https://review.openstack.org/564279
15:05:00 mordred dtantsur, TheJulia: ^^ have a sec to land a bifrost patch to unstick shade?
15:19:17 dtantsur mordred: looking
15:24:00 crunchengine hello mordred: with version 0.12.0, code like server = "conn.compute.get_server("db323e83-8913-4016-9d8e-7d452a3577f2")" then "server.get_console_output(conn.session)" does not work
15:24:11 crunchengine it fails with keystoneauth1.exceptions.catalog.EndpointNotFound: Could not find requested endpoint in Service Catalog.
15:24:39 crunchengine compute URL is not prepended to request URL, so it can't work, do you know why ?
16:03:50 mordred crunchengine: looking
16:04:22 mordred crunchengine: oh - don't use conn.session - use conn.compute
16:04:53 crunchengine already tried :/
16:04:56 mordred crunchengine: server.get_console_output(conn.compute)
16:04:58 mordred yeah?
16:06:47 crunchengine oooh it working!
16:06:55 crunchengine let me try with _action :)
16:06:56 mordred \o/
16:07:48 mordred the session parameter to resource ethods is currently very porrly named
16:22:46 pabelanger Shrews: oh, forgot about that
16:24:25 openstackgerrit Merged openstack/api-wg master: Add guidance on needing cache-control headers https://review.openstack.org/550468
16:41:08 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Infer version from old versioned service type aliases https://review.openstack.org/564299
16:42:25 openstackgerrit Monty Taylor proposed openstack/keystoneauth master: Allow tuples and sets in interface list https://review.openstack.org/564495
16:51:48 elmiko cdent edleafe dtantsur ready for publish, https://etherpad.openstack.org/p/api-sig-newsletter
16:53:16 dtantsur LGTM
16:54:00 cdent elmiko: " elmiko as mentioned" not sure what you're after there
16:54:19 elmiko cdent: fixed, thanks
16:55:39 cdent lgtm
16:57:05 elmiko cool, gonna ship it unless edleafe tells me to pump the brakes
16:57:13 edleafe elmiko: hold on
16:57:23 elmiko yup yup, i see you changing things
16:59:54 edleafe yeah, the wording was a little awkward, so I changed it up a bit
17:00:05 edleafe If you agree with the chages, ship it!
17:00:55 elmiko looks good to me, thanks for the clarity edleafe =)
17:01:44 elmiko sent!
17:06:38 cdent I thought that lack of clarity was just elmiko's inimitable style
17:06:45 cdent which I took as artistry
17:10:48 elmiko cdent: pretty much XD
18:53:46 JPL Hi. I just started using shade to connect to our private cloud instances, and am having some issues switching projects (`connect_as_project`).Ref: https://docs.openstack.org/shade/latest/user/usage.html
18:53:53 JPL https://gist.github.com/jplindquist/acb4ddfb00f9e412563fb26343b138a2
18:55:10 harlowja ya, i've been trying that same thing with JPL
18:55:13 JPL I can connect to my user-project, list servers, etc, all without issue, but if I attempt to connect_as_project to another project, I receive an error: `os_client_config.exceptions.OpenStackConfigException: Cloud defaults was not found.`

Earlier   Later