Unified SDK for OpenStack
Go to file
Stephen Finucane 81d60c7874 README: Add guide on raw HTTP layer
Highlight that one can do e.g. `conn.compute.get('/servers')`. You
almost certainly *don't want* to do this, but it can be helpful (see:
our support for Nova's os-hypervisors API in OSC).

The flow of the README is modified slightly so we go sequentially from
high-level layers to low-level layers. Rubrics (header-like elements
that don't produce anchors or appear in tables of contents) are also
added to produce improve information hierarchy.

Change-Id: Ifd4a5a2c753f6698fa4384a197e81cc5383ef312
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
2024-04-24 11:11:56 +01:00
devstack Finish updating links to point to opendev 2019-04-21 12:31:44 +00:00
doc Merge "Add Binero public cloud to vendor support" 2024-04-22 17:04:43 +00:00
examples Incorrect protocol type in create_security_group_rule() 2024-02-11 08:09:57 +01:00
extras Add ansible stable-2.9 job and run 2.8 and 2.9 2020-03-24 08:43:10 -05:00
openstack Merge "Fix multiple image_id query mappings" 2024-04-22 18:12:13 +00:00
playbooks Prepare acceptance tests for real clouds 2023-03-31 09:28:17 +02:00
releasenotes Merge "Adding SDK support for `glance md-namespace-objects-delete`" 2024-04-22 17:14:34 +00:00
roles/deploy-clouds-config Prepare acceptance tests for real clouds 2023-03-31 09:28:17 +02:00
tools Blackify everything else 2023-05-05 11:31:36 +01:00
zuul.d Add Tap Mirrors to SDK 2024-04-11 09:29:33 +02:00
.coveragerc Fix coverage configuration and execution 2016-03-14 07:34:53 +00:00
.git-blame-ignore-revs Ignore black version bump 2023-05-05 12:49:31 +01:00
.gitignore Merge tox, tests and other support files 2017-10-04 14:51:08 -05:00
.gitreview OpenDev Migration Patch 2019-04-19 19:47:46 +00:00
.mailmap Merge tox, tests and other support files 2017-10-04 14:51:08 -05:00
.pre-commit-config.yaml Merge "mypy: Address issues with remaining service modules" 2024-01-30 09:09:21 +00:00
.stestr.conf Merge shade and os-client-config into the tree 2017-11-15 09:03:23 -06:00
babel.cfg setting up the initial layout; move the api proposals to api_strawman 2014-01-24 22:58:25 -06:00
bindep.txt Remove python-dev from bindep 2022-11-07 11:02:00 +01:00
CONTRIBUTING.rst Switch back to LaunchPad for issue tracking 2023-10-31 11:41:55 +01:00
docs-requirements.txt Add requirements.txt file for readthedocs 2015-05-21 08:16:44 -07:00
HACKING.rst Fix some typos 2019-03-09 17:25:16 +01:00
include-acceptance-regular-user.txt Whitelist cloud functional tests in acceptance 2022-12-02 14:00:32 +01:00
LICENSE setting up the initial layout; move the api proposals to api_strawman 2014-01-24 22:58:25 -06:00
MANIFEST.in setting up the initial layout; move the api proposals to api_strawman 2014-01-24 22:58:25 -06:00
post_test_hook.sh Update load_balancer for v2 API 2017-07-18 18:05:29 -07:00
README.rst README: Add guide on raw HTTP layer 2024-04-24 11:11:56 +01:00
requirements.txt Remove importlib-metadata from requirements 2023-12-01 10:12:09 +09:00
setup.cfg mypy: Address issues with remaining service modules 2024-01-09 10:54:07 +00:00
setup.py Blackify everything else 2023-05-05 11:31:36 +01:00
SHADE-MERGE-TODO.rst Use discovery instead of config to create proxies 2018-10-06 07:44:29 -05:00
test-requirements.txt requirements: Sort alphabetically 2023-11-06 10:47:47 +00:00
tox.ini tox: Correct functional test factors 2024-02-15 14:23:45 +00:00

openstacksdk

openstacksdk is a client library for building applications to work with OpenStack clouds. The project aims to provide a consistent and complete set of interactions with OpenStack's many services, along with complete documentation, examples, and tools.

It also contains an abstraction interface layer. Clouds can do many things, but there are probably only about 10 of them that most people care about with any regularity. If you want to do complicated things, the per-service oriented portions of the SDK are for you. However, if what you want is to be able to write an application that talks to any OpenStack cloud regardless of configuration, then the Cloud Abstraction layer is for you.

More information about the history of openstacksdk can be found at https://docs.openstack.org/openstacksdk/latest/contributor/history.html

Getting started

Authentication and connection management

openstacksdk aims to talk to any OpenStack cloud. To do this, it requires a configuration file. openstacksdk favours clouds.yaml files, but can also use environment variables. The clouds.yaml file should be provided by your cloud provider or deployment tooling. An example:

clouds:
  mordred:
    region_name: Dallas
    auth:
      username: 'mordred'
      password: XXXXXXX
      project_name: 'demo'
      auth_url: 'https://identity.example.com'

openstacksdk will look for clouds.yaml files in the following locations:

  • If set, the path indicated by the OS_CLIENT_CONFIG_FILE environment variable
  • . (the current directory)
  • $HOME/.config/openstack
  • /etc/openstack

You can create a connection using the openstack.connect function. The cloud name can be either passed directly to this function or specified using the OS_CLOUD environment variable. If you don't have a clouds.yaml file and instead use environment variables for configuration then you can use the special envvars cloud name to load configuration from the environment. For example:

import openstack

# Initialize connection from a clouds.yaml by passing a cloud name
conn_from_cloud_name = openstack.connect(cloud='mordred')

# Initialize connection from a clouds.yaml using the OS_CLOUD envvar
conn_from_os_cloud = openstack.connect()

# Initialize connection from environment variables
conn_from_env_vars = openstack.connect(cloud='envvars')

Note

How this is all achieved is described in more detail below.

The cloud layer

openstacksdk consists of four layers which all build on top of each other. The highest level layer is the cloud layer. Cloud layer methods are available via the top level Connection object returned by openstack.connect. For example:

import openstack

# Initialize and turn on debug logging
openstack.enable_logging(debug=True)

# Initialize connection
conn = openstack.connect(cloud='mordred')

# List the servers
for server in conn.list_servers():
    print(server.to_dict())

The cloud layer is based on logical operations that can potentially touch multiple services. The benefit of this layer is mostly seen in more complicated operations that take multiple steps and where the steps vary across providers. For example:

import openstack

# Initialize and turn on debug logging
openstack.enable_logging(debug=True)

# Initialize connection
conn = openstack.connect(cloud='mordred')

# Upload an image to the cloud
image = conn.create_image(
    'ubuntu-trusty', filename='ubuntu-trusty.qcow2', wait=True)

# Find a flavor with at least 512M of RAM
flavor = conn.get_flavor_by_ram(512)

# Boot a server, wait for it to boot, and then do whatever is needed
# to get a public IP address for it.
conn.create_server(
    'my-server', image=image, flavor=flavor, wait=True, auto_ip=True)

The proxy layer

The next layer is the proxy layer. Most users will make use of this layer. The proxy layer is service-specific, so methods will be available under service-specific connection attributes of the Connection object such as compute, block_storage, image etc. For example:

import openstack

# Initialize and turn on debug logging
openstack.enable_logging(debug=True)

# Initialize connection
conn = openstack.connect(cloud='mordred')

# List the servers
for server in conn.compute.servers():
    print(server.to_dict())

Note

A list of supported services is given below.

The resource layer

Below this there is the resource layer. This provides support for the basic CRUD operations supported by REST APIs and is the base building block for the other layers. You typically will not need to use this directly but it can be helpful for operations where you already have a Resource object to hand. For example:

import openstack
import openstack.config.loader
import openstack.compute.v2.server

# Initialize and turn on debug logging
openstack.enable_logging(debug=True)

# Initialize connection
conn = openstack.connect(cloud='mordred')

# List the servers
for server in openstack.compute.v2.server.Server.list(session=conn.compute):
    print(server.to_dict())

The raw HTTP layer

Finally, there is the raw HTTP layer. This exposes raw HTTP semantics and is effectively a wrapper around the requests API with added smarts to handle stuff like authentication and version management. As such, you can use the requests API methods you know and love, like get, post and put, and expect to receive a requests.Response object in response (unlike the other layers, which mostly all return objects that subclass openstack.resource.Resource). Like the resource layer, you will typically not need to use this directly but it can be helpful to interact with APIs that have not or will not be supported by openstacksdk. For example:

import openstack

# Initialize and turn on debug logging
openstack.enable_logging(debug=True)

# Initialize connection
conn = openstack.connect(cloud='mordred')

# List servers
for server in openstack.compute.get('/servers').json():
    print(server)

Configuration

openstacksdk uses the openstack.config module to parse configuration. openstack.config will find cloud configuration for as few as one cloud and as many as you want to put in a config file. It will read environment variables and config files, and it also contains some vendor specific default values so that you don't have to know extra info to use OpenStack

  • If you have a config file, you will get the clouds listed in it
  • If you have environment variables, you will get a cloud named envvars
  • If you have neither, you will get a cloud named defaults with base defaults

You can view the configuration identified by openstacksdk in your current environment by running openstack.config.loader. For example:

$ python -m openstack.config.loader

More information at https://docs.openstack.org/openstacksdk/latest/user/config/configuration.html

Supported services

The following services are currently supported. A full list of all available OpenStack service can be found in the Project Navigator.

Note

Support here does not guarantee full-support for all APIs. It simply means some aspect of the project is supported.

Supported services
Service Description Cloud Layer Proxy & Resource Layer
Compute
Nova Compute ✔ (openstack.compute)
Hardware Lifecycle
Ironic Bare metal provisioning ✔ (openstack.baremetal, openstack.baremetal_introspection)
Cyborg Lifecycle management of accelerators ✔ (openstack.accelerator)
Storage
Cinder Block storage ✔ (openstack.block_storage)
Swift Object store ✔ (openstack.object_store)
Cinder Shared filesystems ✔ (openstack.shared_file_system)
Networking
Neutron Networking ✔ (openstack.network)
Octavia Load balancing ✔ (openstack.load_balancer)
Designate DNS ✔ (openstack.dns)
Shared services
Keystone Identity ✔ (openstack.identity)
Placement Placement ✔ (openstack.placement)
Glance Image storage ✔ (openstack.image)
Barbican Key management ✔ (openstack.key_manager)
Workload provisioning
Magnum Container orchestration engine provisioning ✔ (openstack.container_infrastructure_management)
Orchestration
Heat Orchestration ✔ (openstack.orchestration)
Senlin Clustering ✔ (openstack.clustering)
Mistral Workflow ✔ (openstack.workflow)
Zaqar Messaging ✔ (openstack.message)
Application lifecycle
Masakari Instances high availability service ✔ (openstack.instance_ha)