Puppet::ResourceApi

This is an implementation of the Resource API specification. Find a working example of a new-style providers in the Palo Alto Firewall module: base provider, type, actual provider with validation and xml processing, and new unit tests for 100% coverage.

Deployment

The puppet-resource_api gem is part of the Puppet 6 Platform. With older versions of Puppet, you can use the puppetlabs-resource_api module to install the gem on your servers and agents.

Getting Started

1. Download the Puppet Development Kit (PDK) appropriate to your operating system and architecture.

2. Create a new module with the PDK, or work with an existing PDK-enabled module. To create a new module, run pdk new module <MODULE_NAME> from the command line, specifying the name of the module. Respond to the dialog questions.

3. To add the puppet-resource_api gem and enable "modern" rspec-style mocking, open the .sync.yml file in your editor, and add the following content:

# .sync.yml
---
Gemfile:
  optional:
    ':development':
      - gem: 'puppet-resource_api'
spec/spec_helper.rb:
  mock_with: ':rspec'

4. Apply these changes by running pdk update

You will get the following response:

$ pdk update
pdk (INFO): Updating david-example using the default template, from 1.4.1 to 1.4.1

----------Files to be modified----------
Gemfile
spec/spec_helper.rb

----------------------------------------

You can find a report of differences in update_report.txt.

Do you want to continue and make these changes to your module? Yes

------------Update completed------------

2 files modified.

$

5. Create the required files for a new type and provider in the module by running pdk new provider <provider_name>

You will get the following response:

$ pdk new provider foo
pdk (INFO): Creating '.../example/lib/puppet/type/foo.rb' from template.
pdk (INFO): Creating '.../example/lib/puppet/provider/foo/foo.rb' from template.
pdk (INFO): Creating '.../example/spec/unit/puppet/provider/foo/foo_spec.rb' from template.
pdk (INFO): Creating '.../example/spec/unit/puppet/type/foo_spec.rb' from template.
$

The four generated files are the type, the implementation, and the unit tests. The default template contains an example that demonstrates the basic workings of the Resource API. This allows the unit tests to run immediately after creating the provider, which will look like this:

$ pdk test unit
[✔] Preparing to run the unit tests.
[✔] Running unit tests.
  Evaluated 5 tests in 0.012065973 seconds: 0 failures, 0 pending.
[✔] Cleaning up after running unit tests.
$

Writing the Type

The type contains the shape of your resources. The template provides the necessary name and ensure attributes. You can modify their description and the name's type to match your resource. Add more attributes as you need.

# lib/puppet/type/foo.rb
require 'puppet/resource_api'

Puppet::ResourceApi.register_type(
  name: 'foo',
  docs: <<-EOS,
      This type provides Puppet with the capabilities to manage ...
    EOS
  attributes: {
    ensure: {
      type:    'Enum[present, absent]',
      desc:    'Whether this apt key should be present or absent on the target system.',
      default: 'present',
    },
    name: {
      type:      'String',
      desc:      'The name of the resource you want to manage.',
      behaviour: :namevar,
    },
  },
)

The following keys are available for defining attributes:

• type: the Puppet 4 data type allowed in this attribute. You can use all data types matching Scalar and Data.
• desc: a string describing this attribute. This is used in creating the automated API docs with puppet-strings.
• default: a default value used by the runtime environment; when the caller does not specify a value for this attribute.
• behaviour/behavior: how the attribute behaves. The current available values include:
  • namevar: marks an attribute as part of the "primary key" or "identity" of the resource. A given set of namevar values needs to distinctively identify an instance.
  • init_only: this attribute can only be set during the creation of the resource. Its value will be reported going forward, but trying to change it later leads to an error. For example, the base image for a VM or the UID of a user.
  • read_only: values for this attribute will be returned by get(), but set() is not able to change them. Values for this should never be specified in a manifest. For example, the checksum of a file, or the MAC address of a network interface.
  • parameter: these attributes influence how the provider behaves, and cannot be read from the target system. For example, the target file on inifile, or the credentials to access an API.

Writing the Provider

The provider is the most important part of your new resource, as it reads and enforces state. Here is the example generated by pdk new provider:

require 'puppet/resource_api'
require 'puppet/resource_api/simple_provider'

# Implementation for the foo type using the Resource API.
class Puppet::Provider::Foo::Foo < Puppet::ResourceApi::SimpleProvider
  def get(_context)
    [
      {
        name: 'foo',
        ensure: 'present',
      },
      {
        name: 'bar',
        ensure: 'present',
      },
    ]
  end

  def create(context, name, should)
    context.notice("Creating '#{name}' with #{should.inspect}")
  end

  def update(context, name, should)
    context.notice("Updating '#{name}' with #{should.inspect}")
  end

  def delete(context, name)
    context.notice("Deleting '#{name}'")
  end
end

The optional initialize method can be used to set up state that is available throughout the execution of the catalog. This is most often used for establishing a connection, when talking to a service (e.g. when managing a database).

The get(context) method returns a list of hashes describing the resources that are currently on the target system. The basic example would always return an empty list. Here is an example of resources that could be returned from this:

[
  {
    name: 'a',
    ensure: 'present',
  },
  {
    name: 'b',
    ensure: 'present',
  },
]

The create/update/delete methods get called by the SimpleProvider base-class to change the system as requested by the catalog. The name argument is the name of the resource that is being processed. should contains the attribute hash - in the same format as get returns - with the values in the catalog.

Unit testing

The generated unit tests in spec/unit/puppet/provider/foo_spec.rb get automatically evaluated with pdk test unit.

Remote resources

Support for remote resources is enabled through a transport class. A transport class contains the code for managing connections and processing information to and from the remote resource. For information on how to create a transport class, see the Resource API specification.

An example of a transport class:

# lib/puppet/transport/device_type.rb
module Puppet::Transport
  # The main connection class to a PAN-OS API endpoint
  class DeviceType
    def initialize(context, connection_info)
    # Initialization eg. validate connection_info
    end

    def verify(context)
    # Test that transport can talk to the remote target
    end

    def facts(context)
    # Access target, return a Facter facts hash
    end

    def close(context)
    # Close connection, free up resources
    end
  end
end

An example of a corresponding schema:

# lib/puppet/transport/device_type.rb
Puppet::ResourceAPI.register_transport(
  name: 'device_type', # points at class Puppet::Transport::DeviceType
  desc: 'Connects to a device_type',
  # features: [], # future extension points
  connection_info: {
    host: {
      type: 'String',
      desc: 'The host to connect to.',
    },
    user: {
      type: 'String',
      desc: 'The user.',
    },
    password: {
      type: 'String',
      sensitive: true,
      desc: 'The password to connect.',
    },
    enable_password: {
      type: 'String',
      sensitive: true,
      desc: 'The password escalate to enable access.',
    },
    port: {
      type: 'Integer',
      desc: 'The port to connect to.',
    },
  },
)

Transport Schema keywords

Please note that within the transport schema, the following keywords are reserved words:

Usable within the schema

The following keywords are encouraged within the Transport schema:

• uri - Use when you need to specify a specific URL to connect to. All of the following keys will be computed from the uri if possible. In the future more url parts may be computed from the URI as well.
• host - Use to specify and IP or address to connect to.
• protocol - Use to specify which protocol the transport should use for example http, https, ssh or tcp
• user - The user the transport should connect as.
• port - The port the transport should connect to.

Non-Usable within the schema

The following keywords are keywords that must not be used by the transport schema:

• name - transports should use uri instead of name.
• path - reserved as a uri part
• query - reserved as a uri part
• run-on - This is used by bolt to determine which target to proxy to. Transports should not rely on this key.
• remote-transport - This is used to determine which transport to load. It should always be the transport class name "declassified".
• remote-* Any key starting with remote- is reserved for future use.

Note: Currently bolt inventory requires that a name be set for every target and always uses that name as the URI. This means there is no way to specify host separately from the host section of the name when parsed as a URI.

After the device class, transport class and transport schema have been implemented, puppet device will be able to use the new provider, and supply it (through the device class) with the URL specified in the device.conf.

Transport/device specific providers

To allow modules to deal with different backends independently, the Resource API implements a mechanism to use different API providers side by side. For a given transport/device class (see above), the Resource API will first try to load a Puppet::Provider::TypeName::<DeviceType> class from lib/puppet/provider/type_name/device_type.rb, before falling back to the regular provider at Puppet::Provider::TypeName::TypeName.

Puppet backwards compatibility

To connect to a remote resource through puppet device, you must provide a device shim to maintain compatibility with Puppet. The device shim needs to interface the transport to puppet's config and runtime expectations.

In the simplest case you can use the provided Puppet::ResourceApi::Transport::Wrapper like this:

# lib/puppet/util/network_device/device_type/device.rb

require 'puppet'
require 'puppet/resource_api/transport/wrapper'
# force registering the transport schema
require 'puppet/transport/schema/device_type'

module Puppet::Util::NetworkDevice::Device_type
  class Device < Puppet::ResourceApi::Transport::Wrapper
    def initialize(url_or_config, _options = {})
      super('device_type', url_or_config)
    end
  end
end

Further reading

The Resource API describes details of all the capabilities of this gem.

The hue_rsapi module is a very simple example for using the Resource API for remote resources.

The meraki module is a full example for using the Resource API for remote resources.

This Introduction to Testing Puppet Modules talk describes rspec usage in more detail.

The RSpec docs provide an overview of the capabilities of rspec.

Read betterspecs for general guidelines on what is considered good specs.

Known Issues

This gem is still under heavy development. This section is a living document of what is already done, and what items are still outstanding.

Currently working:

• Basic type and provider definition, using name, desc, and attributes.
• Scalar puppet 4 data types:
  • String, Enum, Pattern
  • Integer, Float, Numeric
  • Boolean
  • Array
  • Optional
  • Variant
• The canonicalize, simple_get_filter, and remote_resource features.
• All the logging facilities.
• Executing the new provider under the following commands:
  • puppet apply
  • puppet resource
  • puppet agent
  • puppet device (if applicable)

There are still a few notable gaps between the implementation and the specification:

• Complex data types, like Hash, Tuple or Struct are not yet implemented.
• Only a single runtime environment (the Puppet commands) is currently implemented.

Restrictions of puppet:

• supports_noop is not effective, as puppet doesn't call into the type under noop at all.
• Attributes cannot be called title, provider, or any of the metaparameters, as those are reserved by puppet itself.

Future possibilities:

• Multiple Providers
• Commands API

Contributing

Bug reports and pull requests are welcome on GitHub at https://github.com/puppetlabs/puppet-resource_api.

Cutting a release

To cut a new release, follow the "Special case: the resource API" instructions on confluence.