azure.azcollection.azure_rm_ml_compute module – Create, Update or Delete an Azure Machine Learning Compute

Note

This module is part of the azure.azcollection collection (version 3.15.0).

You might already have this collection installed if you are using the ansible package. It is not included in ansible-core. To check whether it is installed, run ansible-galaxy collection list.

To install it, use: ansible-galaxy collection install azure.azcollection. You need further requirements to be able to use this module, see Requirements for details.

To use it in a playbook, specify: azure.azcollection.azure_rm_ml_compute.

New in azure.azcollection 3.15.0

Synopsis

  • Create, Update or Delete an Azure Machine Learning Compute.

Requirements

The below requirements are needed on the host that executes this module.

  • python >= 2.7

  • The host that executes this module must have the azure.azcollection collection installed via galaxy

  • All python packages listed in collection’s requirements.txt must be installed via pip on the host that executes modules from azure.azcollection

  • Full installation instructions may be found https://galaxy.ansible.com/azure/azcollection

Parameters

Parameter

Comments

ad_user

string

Active Directory username. Use when authenticating with an Active Directory user rather than service principal.

adfs_authority_url

string

added in azure.azcollection 0.0.1

Azure AD authority url. Use when authenticating with Username/password, and has your own ADFS authority.

admin_password

string

Password for the administrator user account.

admin_username

string

Name of the administrator user account that can be used to SSH into the node(s).

api_profile

string

added in azure.azcollection 0.0.1

Selects an API profile to use when communicating with Azure services. Default value of latest is appropriate for public clouds; future values will allow use with Azure Stack.

Default: "latest"

append_tags

boolean

Use to control if tags field is canonical or just appends to existing tags.

When canonical, any tags not found in the tags parameter will be removed from the object’s metadata.

Choices:

  • false

  • true ← (default)

auth_source

string

added in azure.azcollection 0.0.1

Controls the source of the credentials to use for authentication.

Can also be set via the ANSIBLE_AZURE_AUTH_SOURCE environment variable.

When set to auto (the default) the precedence is module parameters -> env -> credential_file -> cli.

When set to env, the credentials will be read from the environment variables

When set to credential_file, it will read the profile from ~/.azure/credentials.

When set to cli, the credentials will be sources from the Azure CLI profile. subscription_id or the environment variable AZURE_SUBSCRIPTION_ID can be used to identify the subscription ID if more than one is present otherwise the default az cli subscription is used.

When set to msi, the host machine must be an azure resource with an enabled MSI extension. subscription_id or the environment variable AZURE_SUBSCRIPTION_ID can be used to identify the subscription ID if the resource is granted access to more than one subscription, otherwise the first subscription is chosen.

The msi was added in Ansible 2.6.

Choices:

  • "auto" ← (default)

  • "cli"

  • "credential_file"

  • "env"

  • "msi"

cert_validation_mode

string

added in azure.azcollection 0.0.1

Controls the certificate validation behavior for Azure endpoints. By default, all modules will validate the server certificate, but when an HTTPS proxy is in use, or against Azure Stack, it may be necessary to disable this behavior by passing ignore. Can also be set via credential file profile or the AZURE_CERT_VALIDATION environment variable.

Choices:

  • "ignore"

  • "validate"

client_id

string

Azure client ID. Use when authenticating with a Service Principal or Managed Identity (msi).

Can also be set via the AZURE_CLIENT_ID environment variable.

cloud_environment

string

added in azure.azcollection 0.0.1

For cloud environments other than the US public cloud, the environment name (as defined by Azure Python SDK, eg, AzureChinaCloud, AzureUSGovernment), or a metadata discovery endpoint URL (required for Azure Stack). Can also be set via credential file profile or the AZURE_CLOUD_ENVIRONMENT environment variable.

Default: "AzureCloud"

description

string

Description of the compute target.

disable_instance_discovery

boolean

added in azure.azcollection 2.3.0

Determines whether or not instance discovery is performed when attempting to authenticate. Setting this to true will completely disable both instance discovery and authority validation. This functionality is intended for use in scenarios where the metadata endpoint cannot be reached such as in private clouds or Azure Stack. The process of instance discovery entails retrieving authority metadata from https://login.microsoft.com/ to validate the authority. By setting this to **True**, the validation of the authority is disabled. As a result, it is crucial to ensure that the configured authority host is valid and trustworthy.

Set via credential file profile or the AZURE_DISABLE_INSTANCE_DISCOVERY environment variable.

Choices:

  • false ← (default)

  • true

identity_type

string

The type of managed identity.

Choices:

  • "SystemAssigned"

  • "UserAssigned"

idle_time_before_scale_down

integer

Node idle time in seconds before scaling down the cluster.

location

string

The location to be used for the new compute. If not specified, defaults to the location of the workspace.

log_mode

string

Parent argument.

log_path

string

Parent argument.

max_instances

integer

The maximum number of nodes to use on the cluster.

min_instances

integer

The minimum number of nodes to use on the cluster.

ml_workspace

string / required

Name of the Azure ML workspace.

name

string / required

Name of the Azure ML compute.

password

string

Active Directory user password. Use when authenticating with an Active Directory user rather than service principal.

profile

string

Security profile found in ~/.azure/credentials file.

resource_definition

string

Compute definition in YAML

resource_group

string / required

Name of resource group.

secret

string

Azure client secret. Use when authenticating with a Service Principal.

size

string

VM size to use for the compute target. More details can be found here: https://aka.ms/azureml-vm-details.

ssh_key_value

string

SSH public key of the administrator user account.

ssh_public_access_enabled

boolean

Indicates whether public SSH port is enabled.

Choices:

  • false

  • true

state

string

State of the Compute. Use present to create or update and absent to delete. start to start, stop to stop and restart to restart.

Choices:

  • "absent"

  • "present" ← (default)

  • "start"

  • "stop"

  • "restart"

subnet

string

Name of the subnet. Can also reference a subnet in an existing vnet by ID instead of name. If subnet ID is specified then vnet_name will be ignored. Subnet ID can refer to a vnet/subnet in another RG by specifying the fully qualified subnet ID. Required when vnet name is specified.

subscription_id

string

Your Azure subscription Id.

tags

dictionary

Dictionary of string:string pairs to assign as metadata to the object.

Metadata tags on the object will be updated with any provided values.

To remove tags set append_tags option to false.

Currently, Azure DNS zones and Traffic Manager services also don’t allow the use of spaces in the tag.

Azure Front Door doesn’t support the use of

Azure Automation and Azure CDN only support 15 tags on resources.

tenant

string

Azure tenant ID. Use when authenticating with a Service Principal.

thumbprint

string

added in azure.azcollection 1.14.0

The thumbprint of the private key specified in x509_certificate_path.

Use when authenticating with a Service Principal.

Required if x509_certificate_path is defined.

tier

string

VM priority tier.

Choices:

  • "dedicated"

  • "low_priority"

type

string

The type of compute target.

Choices:

  • "amlcompute"

  • "computeinstance"

user_assigned_identities

list / elements=string

User Assigned Identities.

user_object_id

string

AAD object ID of the assigned user.

user_tenant_id

string

AAD tenant ID of the assigned user.

vnet_name

string

Name of the virtual network.

x509_certificate_path

path

added in azure.azcollection 1.14.0

Path to the X509 certificate used to create the service principal in PEM format.

The certificate must be appended to the private key.

Use when authenticating with a Service Principal.

Notes

Note

  • For authentication with Azure you can pass parameters, set environment variables, use a profile stored in ~/.azure/credentials, or log in before you run your tasks or playbook with az login.

  • Authentication is also possible using a service principal or Active Directory user.

  • To authenticate via service principal, pass subscription_id, client_id, secret and tenant or set environment variables AZURE_SUBSCRIPTION_ID, AZURE_CLIENT_ID, AZURE_SECRET and AZURE_TENANT.

  • To authenticate via Active Directory user, pass ad_user and password, or set AZURE_AD_USER and AZURE_PASSWORD in the environment.

  • Alternatively, credentials can be stored in ~/.azure/credentials. This is an ini file containing a [default] section and the following keys: subscription_id, client_id, secret and tenant or subscription_id, ad_user and password. It is also possible to add additional profiles. Specify the profile by passing profile or setting AZURE_PROFILE in the environment.

See Also

See also

Sign in with Azure CLI

How to authenticate using the az login command.

Examples

- name: Create ML Compute
  azure.azcollection.azure_rm_ml_compute:
    location: eastus
    name: MyCompute
    description: My Compute created by Ansible
    resource_group: myResourceGroup
    ml_workspace: myMLWorkspace
    type: amlcompute
    min_instances: 0
    max_instances: 8
    tags:
      createdByToolkit: Ansible

Return Values

Common return values are documented here, the following are the fields unique to this module:

Key

Description

changed

boolean

Whether the resource is changed.

Returned: always

Sample: false

ml_compute

dictionary

Compute that was just created or updated.

Returned: always

Sample: {"description": "Compute Integration Tests", "enable_node_public_ip": true, "id": "/subscriptions/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/resourceGroups/xxxxx-ml-compute/providers/Microsoft.MachineLearningServices/workspaces/workspace-xxxxxxx-xxx/computes/amlxxxxxxx-xxxcomp", "idle_time_before_scale_down": 120, "location": "eastus", "max_instances": 200, "min_instances": 0, "name": "amlxxxxxxx-xxxcomp", "network_settings": {}, "provisioning_state": "Succeeded", "size": "STANDARD_DS3_V2", "ssh_public_access_enabled": true, "tags": {"created_by": "Ansible"}, "tier": "dedicated", "type": "amlcompute"}

Authors

  • Bill Peck (@p3ck)