Configuring GPU cluster access permissions

Written by

Yandex Cloud

Updated at October 17, 2024

Assigning a role
Assigning multiple roles
Revoking a role

To grant a user, group, or service account access to a GPU cluster, assign a role for it.

Assigning a role

Management console

CLI

Terraform

API

In the management console, select the folder containing a GPU cluster.
Select Compute Cloud.
In the left-hand panel, select → GPU clusters.
Select the GPU cluster you need.
Go to the Access bindings tab.
Click Assign bindings.
In the window that opens, select the group, user, or service account to grant access to the GPU cluster.
Click Add role and select the required role.
Click Save.

If you do not have the Yandex Cloud command line interface yet, install and initialize it.

The folder specified in the CLI profile is used by default. You can specify a different folder using the --folder-name or --folder-id parameter.

See the CLI command description for assigning a role for a GPU cluster:
```
yc compute gpu-cluster add-access-binding --help
```
Get a list of GPU clusters in the default folder:
```
yc compute gpu-cluster list
```

View the list of roles already assigned for the resource:

yc compute gpu-cluster list-access-bindings <GPU_cluster_ID>

Assign the role using the command:
- To a user:
```
yc compute gpu-cluster add-access-binding <GPU_cluster_ID> \
  --user-account-id <user_ID> \
  --role <role>
```
  Where:
  - --user-account-id: User ID. To assign a role to all authenticated users, use the --all-authenticated-users flag.
  - --role: Role to assign.
- To a service account:
```
yc compute gpu-cluster add-access-binding <GPU_cluster_ID> \
  --service-account-id <service_account_ID> \
  --role <role>
```
  Where:
  - --service-account-id: Service account ID.
  - --role: Role to assign.

Terraform allows you to quickly create a cloud infrastructure in Yandex Cloud and manage it using configuration files. Configuration files store the infrastructure description in the HashiCorp Configuration Language (HCL). Terraform and its providers are distributed under the Business Source License.

For more information about the provider resources, see the documentation on the Terraform website or the mirror.

If you change the configuration files, Terraform automatically detects which part of your configuration is already deployed, and what should be added or removed.

If you don't have Terraform, install it and configure the Yandex Cloud provider.

To assign a role to access a GPU cluster using Terraform:

In the Terraform configuration file, describe the parameters of the resources you want to create:
```
resource "yandex_compute_gpu_cluster_iam_binding" "sa-access" {
  gpu_cluster_id = "<GPU_cluster_ID>"
  role           = "<role>"
  members        = ["<subject_type>:<subject_ID>"]
}
```
Where:
- gpu_cluster_id: GPU cluster ID.
- role: Role to assign.
- members: List of types and IDs of subjects the role is assigned to. Specify it as userAccount:<user_ID> or serviceAccount:<service_account_ID>.
For more information about the yandex_compute_gpu_cluster_iam_binding resource parameters, see the provider documentation.
Apply the changes:
1. In the terminal, change to the folder where you edited the configuration file.
2. Make sure the configuration file is correct using the command:
```
terraform validate
```
  If the configuration is correct, the following message is returned:
```
Success! The configuration is valid.
```
3. Run the command:
```
terraform plan
```
  The terminal will display a list of resources with parameters. No changes are made at this step. If the configuration contains errors, Terraform will point them out.
4. Apply the configuration changes:
```
terraform apply
```
5. Confirm the changes: type yes in the terminal and press Enter.
Terraform will create all the required resources. You can check the update using the management console or this CLI command:
```
yc compute gpu-cluster list-access-bindings <GPU_cluster_ID>
```

To assign a role, use the updateAccessBindings REST API method for the GpuCluster resource or the GpuClusterService/UpdateAccessBindings gRPC API call. In the request body, set the action property to ADD and specify the user type and ID under subject.

Assigning multiple roles

Management console

CLI

Terraform

API

In the management console, select the folder containing a GPU cluster.
Select Compute Cloud.
In the left-hand panel, select → GPU clusters.
Select the GPU cluster you need.
Go to the Access bindings tab.
Click Assign bindings.
In the window that opens, select the group, user, or service account to grant access to the GPU cluster.
Click Add role and select the required role.
To add another role, click Add role.
Click Save.

If you do not have the Yandex Cloud command line interface yet, install and initialize it.

The folder specified in the CLI profile is used by default. You can specify a different folder using the --folder-name or --folder-id parameter.

You can assign multiple roles using the set-access-bindings command.

Alert

The set-access-bindings command completely rewrites the access permissions for the resource. All current resource roles will be deleted.

Make sure the resource has no roles assigned that you would not want to lose:
```
yc compute gpu-cluster list-access-bindings <GPU_cluster_ID>
```
See the CLI command description for assigning roles for a GPU cluster:
```
yc compute gpu-cluster set-access-bindings --help
```

Assign roles:

yc compute gpu-cluster set-access-bindings <GPU_cluster_ID> \
  --access-binding role=<role>,subject=<subject_type>:<subject_ID> \
  --access-binding role=<role>,subject=<subject_type>:<subject_ID>

Where:

--access-binding: Parameters for setting access permissions:
- role: Role to assign.
- subject: Type and ID of the subject getting the role.

For example, assign roles to multiple users and a service account:

yc compute gpu-cluster set-access-bindings my-gpu-cluster \
  --access-binding role=editor,subject=userAccount:gfei8n54hmfh********
  --access-binding role=viewer,subject=userAccount:helj89sfj80a********
  --access-binding role=editor,subject=serviceAccount:ajel6l0jcb9s********

For more information about the provider resources, see the documentation on the Terraform website or the mirror.

If you change the configuration files, Terraform automatically detects which part of your configuration is already deployed, and what should be added or removed.

If you don't have Terraform, install it and configure the Yandex Cloud provider.

To assign multiple roles for file storage using Terraform:

In the Terraform configuration file, describe the parameters of the resources you want to create:
```
resource "yandex_compute_gpu_cluster_iam_binding" "role1" {
  gpu_cluster_id = "<GPU_cluster_ID>"
  role           = "<role_1>"
  members        = ["<subject_type>:<subject_ID>"]
}

resource "yandex_compute_gpu_cluster_iam_binding" "role2" {
  gpu_cluster_id = "<GPU_cluster_ID>"
  role           = "<role_2>"
  members        = ["<subject_type>:<subject_ID>"]
}
```
Where:
- gpu_cluster_id: GPU cluster ID.
- role: Role to assign.
- members: List of types and IDs of subjects the role is assigned to. Specify it as userAccount:<user_ID> or serviceAccount:<service_account_ID>.
For more information about the yandex_compute_gpu_cluster_iam_binding resource parameters, see the provider documentation.
Apply the changes:
1. In the terminal, change to the folder where you edited the configuration file.
2. Make sure the configuration file is correct using the command:
```
terraform validate
```
  If the configuration is correct, the following message is returned:
```
Success! The configuration is valid.
```
3. Run the command:
```
terraform plan
```
  The terminal will display a list of resources with parameters. No changes are made at this step. If the configuration contains errors, Terraform will point them out.
4. Apply the configuration changes:
```
terraform apply
```
5. Confirm the changes: type yes in the terminal and press Enter.
Terraform will create all the required resources. You can check the update using the management console or this CLI command:
```
yc compute gpu-cluster list-access-bindings <GPU_cluster_ID>
```

To assign roles for a resource, use the setAccessBindings REST API method for the GpuCluster resource or the GpuClusterService/SetAccessBindings gRPC API call.

Alert

The setAccessBindings method and the GpuClusterService/SetAccessBindings call completely rewrite the resource access permissions. All current resource roles will be deleted.

Revoking a role

Management console

CLI

Terraform

API

In the management console, select the folder containing a GPU cluster.
Select Compute Cloud.
In the left-hand panel, select → GPU clusters.
Select the GPU cluster you need.
Go to the Access bindings tab.
In the line with the user you need, click and select Edit roles.
Next to the role, click .
Click Save.

If you do not have the Yandex Cloud command line interface yet, install and initialize it.

The folder specified in the CLI profile is used by default. You can specify a different folder using the --folder-name or --folder-id parameter.

See the CLI command description for revoking a role for a GPU cluster:
```
yc compute gpu-cluster remove-access-binding --help
```

View the roles and assignees for the resource:

yc compute gpu-cluster list-access-bindings <GPU_cluster_ID>

To revoke access permissions, run this command:

yc compute gpu-cluster remove-access-binding <GPU_cluster_ID> \
  --role=<role> \
  --subject=<subject_type>:<subject_ID> \

Where:

--role: ID of the role to revoke.
--subject: Type and ID of the subject getting the role.

For example, to revoke the viewer role for a GPU cluster from a user with the ajel6l0jcb9s******** ID:

yc compute gpu-cluster remove-access-binding my-gpu-cluster \
  --role viewer \
  --subject userAccount:ajel6l0jcb9s********

For more information about the provider resources, see the documentation on the Terraform website or the mirror.

If you change the configuration files, Terraform automatically detects which part of your configuration is already deployed, and what should be added or removed.

If you don't have Terraform, install it and configure the Yandex Cloud provider.

To revoke a role assigned for a GPU cluster using Terraform:

Open the Terraform configuration file and delete the fragment with the role description.

resource "yandex_compute_gpu_cluster_iam_binding" "sa-access" {
  gpu_cluster_id = "<GPU_cluster_ID>"
  role           = "<role>"
  members        = ["<subject_type>:<subject_ID>"]
}

Apply the changes:
1. In the terminal, change to the folder where you edited the configuration file.
2. Make sure the configuration file is correct using the command:
```
terraform validate
```
  If the configuration is correct, the following message is returned:
```
Success! The configuration is valid.
```
3. Run the command:
```
terraform plan
```
  The terminal will display a list of resources with parameters. No changes are made at this step. If the configuration contains errors, Terraform will point them out.
4. Apply the configuration changes:
```
terraform apply
```
5. Confirm the changes: type yes in the terminal and press Enter.
You can check the update using the management console or this CLI command:
```
yc compute gpu-cluster list-access-bindings <GPU_cluster_ID>
```

To revoke a role, use the updateAccessBindings REST API method for the GpuCluster resource or the GpuClusterService/UpdateAccessBindings gRPC API call. In the request body, set the action property to REMOVE and specify the user type and ID under subject.

Configuring GPU cluster access permissions

Assigning a roleAssigning a role

Assigning multiple rolesAssigning multiple roles

Revoking a roleRevoking a role

Was the article helpful?

Assigning a role

Assigning multiple roles

Revoking a role