summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Add flag to trigger task interface for rackspace0.8.1Monty Taylor2015-04-302-0/+2
| | | | | | | It turns out that the task interface for upload is not implied by v2 glance api. Add a boolean flag that can be consumed elsewhere. Change-Id: I2cdcfe302a73ebfa7f739399c1eeb3bc9f96ab3c
* Document vendor support information0.8.0Monty Taylor2015-04-142-0/+87
| | | | | | | In the future, I'd like for this doc to be generated from the vendors.py file, but for now, this is great. Change-Id: Ifd0c8da5da46ba156c789f05398abcfa689f4f01
* Reset cache default to 0Monty Taylor2015-04-141-1/+1
| | | | | | None breaks the ansible inventory script. Change-Id: Iac30cdcce3a51910e0b373521263b239f7478a15
* Merge "Move region_names out of auth dict"Jenkins2015-04-131-2/+2
|\
| * Move region_names out of auth dictMonty Taylor2015-04-131-2/+2
| | | | | | | | | | | | region_name is not an auth parameter - this is just simply an error. Change-Id: I5cc3847932d7c51288f451b4532b71f95d8c823d
* | Merge "Add runabove to vendors"Jenkins2015-04-131-0/+8
|\ \ | |/
| * Add runabove to vendorsMonty Taylor2015-04-131-0/+8
| | | | | | | | Change-Id: I319365aeb3a5a00498b37128c5c9fbaf018d88f4
* | Merge "add .venv to gitignore"Jenkins2015-04-131-0/+1
|\ \ | |/ |/|
| * add .venv to gitignoreTerryHowe2015-04-131-0/+1
| | | | | | | | Change-Id: I36856a31f7a68280f9b787e5ab9ce5ac3aa0dc60
* | Add image information to vexxhost accountMonty Taylor2015-04-131-0/+2
|/ | | | Change-Id: I0e39c2a9828fb4fa73403158c3e58fb346ac9a10
* Merge "Add vexxhost"Jenkins2015-04-131-0/+6
|\
| * Add vexxhostMonty Taylor2015-04-131-0/+6
| | | | | | | | | | | | vexxhost has a public cloud - we should list them in our vendors file. Change-Id: Icf1276d59dadf50cabca3a7c2540121fb1cf7057
* | Merge "Add DreamCompute to vendors list"Jenkins2015-04-131-1/+11
|\ \ | |/
| * Add DreamCompute to vendors listMonty Taylor2015-04-131-1/+11
| | | | | | | | Change-Id: I4d81e221c8105d796dcd29fcd7628738486e4b00
* | Allow overriding envvars as the name of the cloudMonty Taylor2015-04-114-17/+40
| | | | | | | | | | | | | | | | | | | | For environment variables created cloud objects, it's possible someone may not want it to be called envvars. I mean, let's be honest, I cannot imagine why this would be important ... but people get emotional about things. Let them name their cloud "bunnyrabbit" because that makes people happy. Change-Id: I0c232de7d93080ea632fb66a82b9e6d3e925c901
* | Put env vars into their own cloud configMonty Taylor2015-04-117-109/+181
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The semantics around mixing environment variables and config file values are confusing at best and no reasonable usecase has been expressed as to why doing so is desirable. Move the logic around environment variable processing to always provide an "envvars" cloud if any envvars are set. The cloud will only exist in the presence of OS_ env vars. get_one_cloud() will default to returning the envvars cloud if it exists. Change-Id: I6c3a54997c3278feedfdf93cc4d1e74b6235700a Closes-Bug: #1439927
* | Add keystoneclient to test-requirementsMonty Taylor2015-04-112-6/+7
| | | | | | | | | | | | | | | | | | | | The auth parameter name validation requires keystoneclient and can't be tested if it's not there. While we're at it - update the current requirements to be inline with global requirements. Change-Id: I6da62476f3851670545143184f9f29479f1caaca
* | Actually put some content into our sphinx docsMonty Taylor2015-04-114-22/+4
|/ | | | | | | | The README is pretty good - get it into the main index page for the sphinx docs so that we can publish them. Partial-Bug: #1440814 Change-Id: Ic72b81964cab1f939f08b957dec3be969c47a32e
* Update .gitreview for git section rename0.7.0Monty Taylor2015-03-281-1/+1
| | | | Change-Id: I4a19b251c5fdf76ffb10387002a13336203a7b72
* Merge "Change dogpile cache defaults"Jenkins2015-03-132-5/+15
|\
| * Change dogpile cache defaultsMonty Taylor2015-03-132-5/+15
| | | | | | | | | | | | | | | | | | | | | | Memory cache can grow unbounded, so it should really be opt in. Change to match shade with the following defaults: - If you specify nothing, you get null cache - If you specify an expiration time and nothing else, you get memory cache - If you specify an explicit cache class, you will get that class Change-Id: I6c9eab71a88a534de7e52ad2a564450c44aacc1d
* | Add cover to .gitignoreGhe Rivero2015-03-101-1/+2
|/ | | | | | Coverage reports in the cover directory should be ignored by git Change-Id: I4cde6b0b8fa2fa04ce6a4c308af3c37dfcdf1b5d
* Flesh out api version defaults0.6.0Monty Taylor2015-03-053-8/+34
| | | | | | | | | | | | openstackclient needs a bit richer support for api version defaults. Namely, it knows what defaults it wants to have - but we need to do defaults processing in os-client-config to get sequencing correct. So provide an API call to set new defaults that can be used before config processing. Also, flesh out the dict of known default values with good defaults to match osc behavior, and add the known v1 default of HP to the vendors.py values. Change-Id: I45e2550af58aee616ca168d20a557077beeab007
* Handle project_name/tenant_name in the auth dictMonty Taylor2015-03-042-11/+37
| | | | | | | | We were doing backwards compat for project/tenant in a way that didn't notice anything in the auth dict - which is there project/tenant info goes. Ooops. Change-Id: I141c1d99f31f381898bf993c4e7fcab1078f40c6
* Add two newlines to the ends of filesMonty Taylor2015-03-042-2/+2
| | | | | | flake8 bites it for me on these locally. Change-Id: I99291c64fafff423aa720da7dba030526ca0cb50
* Rename auth_plugin to auth_typeMonty Taylor2015-03-031-15/+21
| | | | | | | | To match with python-openstackclient is doing, change auth_plugin to auth_type. Since this is out in the wild already, add it to the backwards compat matrix. Change-Id: I36b3f18d57fa827028194f8af91ea309b53b6ee3
* Merge "Add support for configuring dogpile.cache"0.5.0Jenkins2015-03-011-0/+12
|\
| * Add support for configuring dogpile.cacheMonty Taylor2015-03-011-0/+12
| | | | | | | | | | | | | | | | shade has added dogpile.cache support. Since os-client-config already supports processing global cache settings, go ahead and add support for a couple of settings that can be used to feed dogpile.cache. Change-Id: I4d40753b83041c8a48b5c0a6d446f9e0de68220a
* | Merge "Fix coverage report"Jenkins2015-03-011-2/+2
|\ \
| * | Fix coverage reportClint Byrum2015-02-281-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | Coverage can't handle package names that have dashes, as it uses the package name to look for the base module name. So we need to pass in the base module name as it is imported. Change-Id: I2840eea85acaee2d05cab47fb67010e002a14bc0
* | | Merge "Add more testing of vendor yaml loading"Jenkins2015-03-011-22/+54
|\ \ \
| * | | Add more testing of vendor yaml loadingClint Byrum2015-02-281-22/+54
| |/ / | | | | | | | | | | | | | | | | | | Adding coverage for vendor yaml loading and refactoring some tests to make the structure for that sane. Change-Id: I7aca0fcc0b04371f9a71e71c0224897b19cb04af
* | | Merge "Allow keystone validation bypass for noauth usage"Jenkins2015-03-011-0/+4
|\ \ \ | |/ / |/| |
| * | Allow keystone validation bypass for noauth usageJulia Kreger2015-02-271-0/+4
| |/ | | | | | | | | | | | | | | | | | | | | To enable the eventual use of the shade library for communication with ironic in situations where authentication is set to noauth, it is necessary to not attempt to validate a user's credentials. Added a check to disable authentication validation when the auth_plugin is set to '' or 'None' or None. Change-Id: I3807b4724ce5e204b5857c1dbf5325f0e3f4a78d
* | More comprehensive unit tests for os-client-configClint Byrum2015-02-272-0/+43
| | | | | | | | | | | | This includes ensuring that yaml of a sane format can be loaded. Change-Id: I698b3139b7e44f000d2a413d17e79914ef542a22
* | Add basic unit test for configClint Byrum2015-02-262-2/+26
|/ | | | | | Adding this exposed some python3 compatibility issues with iteritems. Change-Id: Ia78bd8edd17c7d2360ad958b3de734503d400774
* Prefer dest value when option is depricated0.4.3Julia Kreger2015-02-111-1/+7
| | | | | | | | Changed logic that re-assigns depricated key names while preserving their values, to prefer the value stored in the dest key if it exists, instead of attempting to generate the new key name. Change-Id: Ibe961688cdb6bd4c9b2dbd27b08c722c3c741586
* Allow region_name to be NoneJulia Kreger2015-02-111-1/+1
| | | | | | | | | | | | | | | | | When using the shade library with ansible playbooks and classical environment variables for authentication, it is possible for a region_name value to be passed in with the value of None. When the value is set to and preserved as None, the logic later on in the method will fail to create a config key/value entry for region_name which is required by code later on in the method. Modified the region_name check to not only check for the presence of a missing region_name value, but to check to see if that value is set to None. This allows the value to be reset preventing the method from failing on the return call due to a missing key. Change-Id: Id8a3edf53ac751f0c6ee4d71405a926ba90c0602
* Don't return the auth dict inside the loop0.4.2Monty Taylor2015-02-111-1/+1
| | | | | | | | It turns out that when you're looping over a set of params to move them into an internal dict, returning inside the loop results in processing exactly one of them. This is not, it turns out, what you wanted. Change-Id: If1bf0c22b758e7238846b08991f4b0d25c841583
* Make sure we're deep-copying the auth dictMonty Taylor2015-02-111-2/+6
| | | | | | | | Having a defaults dict that has an empty dict and then does a bunch of updates means you have ONE instance of a dict that all of the other instances have references to. Change-Id: Id008f7ec98ff7b392553cebca5a5b301330e67a3
* Remove runtime depend on pbr0.4.1Monty Taylor2015-01-271-3/+0
| | | | | | | os-client-config is not using pbr during runtime, so the runtime requirement for it was a lie. Change-Id: I3ed57ec5c2b0fdf4060ef17d6df1fa801cfa14cd
* Provide Rackspace service_name overrideMonty Taylor2015-01-271-0/+1
| | | | | | | | Rackspace puts two compute services in their catalog. This means that keystone session code cannot find the right one without a name match override. Change-Id: I1bc06b97261341ad01bf84ebf5a12294cd0d383c
* Merge "Workflow documentation is now in infra-manual"0.4.0Jenkins2015-01-211-4/+3
|\
| * Workflow documentation is now in infra-manualJeremy Stanley2014-12-051-4/+3
| | | | | | | | | | | | | | Replace URLs for workflow documentation to appropriate parts of the OpenStack Project Infrastructure Manual. Change-Id: I97d11d3a24a374c9212cd29f49472e199b1c8bc0
* | Merge "Start keeping default versions for all services"Jenkins2015-01-211-5/+9
|\ \
| * | Start keeping default versions for all servicesMonty Taylor2015-01-211-5/+9
| | | | | | | | | | | | | | | | | | It turns out we need to do the evil glance dance for almost everything. Change-Id: Ic0ad77ba0627bd4be88bdf0136aa04c2ba43afe6
* | | Merge "Support keystone auth plugins in a generic way."Jenkins2015-01-213-34/+104
|\ \ \ | |/ /
| * | Support keystone auth plugins in a generic way.Monty Taylor2015-01-213-34/+104
| | | | | | | | | | | | | | | | | | | | | | | | | | | Basically, if keystoneclient is available, validate arguments into the auth dict. If it's not - we should probably be talking about what possible use this library has - but it should degrade cleanly and treat everything as passthrough. Change-Id: Ia31039d5c724eba22d053a004eefeaf6857f500d
* | | Merge "Corrections to readme"Jenkins2015-01-211-3/+1
|\ \ \ | |/ / |/| |
| * | Corrections to readmeJeremy Stanley2014-11-241-3/+1
| |/ | | | | | | | | | | | | | | * README.rst: Remove docs and bugs links since the project is not currently publishing/consuming these anywhere that I can find. Also correct the source URL. Change-Id: I2b7002f7ea301a2c03a2e41e904de0c91a14831a
View Lorry Controller Status