| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
| |
keystoneclient bug #1513839 means that even when you proerly pass
a Session (like we do) to the discovery constructor (what you'd be
calling if you were passing "pass_version_arg = True") ksc fails
because you didn't send a URL.
We _HAVE_ an appropriate URL that we can pull from the Session. So
until ksc learns how to pull the URL from the Session itself, do it
for them.
Change-Id: I38eb4cfa750fab5196b86989c2cd498d41bf37ac
|
|
|
|
|
|
|
| |
While we're at it, let's also put in some tests to ensure that we're
processing data types properly.
Change-Id: I0442d234e8422a58738612b2da114f61cc9afc5c
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Excitingly, one of the places where we hard-code a workaround for
a differently behaving client is not needed. The Client in neutron
that takes a version arg is in neutronclient.neutron.client. Also,
pass the version arg as positional when we pass it, as some of the
clients (neutron) require it, and all accept it.
Change-Id: Ifcbaab782173b95a678af0b2792a1194b198b687
|
|/
|
|
|
|
| |
We do it for int values, so floats should be covered, too.
Change-Id: I1b98f7887ad421596b3bfdf2f6ad871a7d7f6bb2
|
|
|
|
|
|
|
|
|
| |
neutronclient expects 2.0, not 2 - but we'd not noticed because all
of the consumption of client constructors were bypasing the version
arg for neutron. In a follow-up patch we'll fix that, but the
value in defaults.json is just straight wrong.
Change-Id: I684c4ef063193355c6cf936d4f18576db919762b
|
|
|
|
|
|
| |
It turns out that requests does not like that.
Change-Id: I206be8107f5cfaaa7dc7f34ab0b0764e0dc3fb0d
|
|
|
|
|
|
|
| |
We have no things that expect actual int values. It's easy to write
ints in yaml though - so normalize them to strings.
Change-Id: I58cfdda697da9b7e030f4165ef8e60f4e9d00b66
|
|\ |
|
| |
| |
| |
| |
| |
| |
| | |
Since we can return a keystoneauth1 Session correctly, disable the
stupid warnings that are evil and never should be emitted.
Change-Id: I128e4587ab07a20c7c5da745b12e4f5d0a3ee5a7
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
The _log.py module is from shade and is just basic logging support
that does not emit warnings if the consumer does not have logging
enabled.
Change-Id: Id4639763cf488eb7cd0c27904be055a7843e287f
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
These come originally from the shade library, but are helpful for
things like the client libs themselves. Once one has a CloudConfig,
there is really one and only one correct way to get both a session
and a Client.
Change-Id: I1b4d4321828864fddab85a127fbf63f4c8384ab9
|
|/
|
|
|
|
| |
Turns out the lon region is not really a thing.
Change-Id: Ib315c301d5f8b589006a61d2a255a6b295b1b9a5
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In preparation for sharing the default and vendor data with other
projects, potentially even non-python ones, move the data into json
format, which is slighly less exciting to read, but has more widespread
standard library support. The user-facing config file will still be in
yaml format, because that's easier on the eyes and it's expected to be
read and edited by humans.
Continue to accept yaml everywhere, because an end user may have dropped
a yaml config file into a dir somewhere, and that's fine.
Change-Id: I269d31e61da433ac20abb39acdde0f9f9fe12837
|
|
|
|
|
|
| |
Auro runs neutron now. Update the config and the docs to reflect that.
Change-Id: I4d2fd12b6f3d93f9e1d201b687145f1fe022e593
|
|
|
|
|
|
|
| |
There is the world's weirdest race condition in the unit test for this
on python 3.4.
Change-Id: Idee8320f03651964600ba0822edfff04f880ea4a
|
|
|
|
|
|
| |
Small typo making one region not useful for customers (my bad!).
Change-Id: If2680b521f92789f2621aa1249b304666fcde95b
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
We say "not supported" when the cloud doens't do something, but there
were two places where we used other words. Fix them. Also, make the
image properties a blockquote.
Change-Id: Id933f7d38d706c6fb3ed9f59ad10c2e1e2d73c43
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| | |
We've got enough of them that the arbitrary sort order was a bit weird.
Sort alphabetically.
Change-Id: I6d4ec563f03d1b25bad7d8337db90007a562e97c
|
|\ \
| |/ |
|
| |
| |
| |
| | |
Change-Id: I05a716bc54c98390a7cbeb352338a7c6cd7e86c3
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Some clouds have auth_urls per region, which means putting a static
auth_url in the config file isn't going to work. Allow {} substitutions
in the auth_url so that other elements of the auth dict can be injected.
This will not solve _everything, but it will solve all of the currently
identified issues, and should be upwardly open if we need something
more complex later.
Change-Id: I6107d0669734647cfa0bef118bdf7739949991ad
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
Just using equals is not a consistent test, because dict ordering can
change. assertDictEqual tests that the contents of the two dicts are the
same.
Change-Id: I1b084a3fee91d73d57a6b9e6a2d0ab9c186e5572
|
|/
|
|
|
|
|
|
|
| |
Regions are special, since they're a driver for creating a cloud
config object in the first place. We weren't syncing them from
the vendor profile, so if the region wasn't set in the clouds.yaml,
region name validation (and get_all_clouds) would fail.
Change-Id: I5b8d6807a86b87a7f69f523f2fee2784389fab0a
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We just added an unreleased interface method to the CloudConfig object -
but maybe that should have been more aligned with dogpile words.
SO - change the docs to reference the dogpile words and add support for
that, while keeping backwards compat support for people using max_age.
Also, do the -/_ transform on the cache config like elsewhere.
Then, while we're in there, add support for per-service cache timings.
We need this in nodepool and shade is adding support, so ability to
configure it will be important.
Change-Id: I31190a31ab0b79fc080db3611c0cd584076387d4
|
|\ |
|
| |
| |
| |
| | |
Change-Id: I01df6e7f0cf09d52a50e1af14cdc76db90b44320
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
We have lists of valid regions for clouds, but specifying a region
incorrectly is a common mistake (specifying dfw instead of DFW,
for instance) Throw an error early with a helpful error message.
Change-Id: I55edf20c6dddde4a4d71dad33a41d4e10448ddac
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| | |
It turns out region_name is an important parameter - it's not
just another kwarg that will get passed through.
Change-Id: I5cca8d324a1dcd1355991df793fe29eedfa15dc0
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The unit of consumption is the CloudConfig object, but the cache
settings are found only globally in the OpenStackConfig object. Pass
the OpenStackConfig object in so that a user consuming settings from
a CloudConfig can find out what the cache settings are.
Change-Id: I633e35b1d7f295581d7abed9c3957e802690614e
|
|/
|
|
| |
Change-Id: I2c40965378dedd67808875eb9453ab8664f0fee8
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Since os_client_config started pulling in ksa plugins we have some
duplicate behaviour with OSC. If o-c-c is going to be loading auth
plugins OSC needs a way to hook into the notch between setting config
values and actually loading the plugins in order to maintain existing
behaviours for compatibility.
It may also be desirable at some point to defer the plugin loading,
OSC did that orignally due to ksc's poor load times, ksa shouldn't
have the same problem.
Change-Id: I1e8a50e5467fd4f9151a64aa7ae140d2654f3186
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Remove identity api version 2 support (default) and some cleanup
in the doc (tables properly formated)
Change-Id: I5d90723dd54905c8a67c52c6111bdf089f4346a0
|
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently the vendor support list is mis-formatted for the last
two entries on the page. It appears that the tables don't have
lines underneath them completing the formatting.
This patch adds lines under the last two vendor entries fixing
the formatting of the tables for those entries.
Change-Id: I6327eef99059291fcce5c9493a622bec33da3824
|
|/
|
|
|
|
| |
os-client-config gates on python3.
Change-Id: I887ab2f2f2436e7423eab8abc23655423ee7b226
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Rackspace puts neutron in the service catalog but it does not work (404s
the endpoint) While it's easy enough to try/except around network calls,
it's also a known fact and can be communicated clearly. Allow things
that want to short-circuit read the value from the config and react
accordingly.
Note: This is not a value going in to the defaults, because the fact
that this is happening at all is highly strange and hopefully something
that can be solved by collaborating on future iterations of DefCore.
However, communicating clearly that it's a known issue is helpful for a
class of users.
Change-Id: I506f4dd397608f2feb6fbda48d297d283312a169
|
|/
|
|
| |
Change-Id: Idf3df94ea039f07fa958a3afadf3388221ffa2ff
|
|
|
|
|
|
|
|
|
| |
IPv6 support is detectable, so rather than having a user opt-in to it,
provide a flag that can be provided to tell it that detected IPv6
support is lying. This should have to be set for far fewer people and
should result in transparent opt-in to IPv6 where available.
Change-Id: Ib0c4c4e8b3b7b4bcee5fa3414719969274929b9a
|