Clusters

Create and manage GPU clusters.

clusters create

tg beta clusters create

Parameters

Flag	Description
--name [string]	Name for the cluster.
--num-gpus [integer]	Number of GPUs to allocate in the cluster
--region [string]	Region to create the cluster in. Valid regions can be found with clusters list-regions
--billing-type [ON_DEMAND|RESERVED]	Specify the cluster reservation approach. ON_DEMAND begins billing the moment the cluster is created. Billing continues until you delete the cluster. RESERVED starts billing immediately. The cluster is automatically torn down after the --duration-days length elapses.
--nvidia-driver-version [string]	Driver version to use for the cluster. Valid driver versions can be found with clusters list-regions
--cuda-version [string]	CUDA version to use for the cluster. Valid CUDA versions can be found with clusters list-regions
--duration-days [number]	Only used with RESERVED billing to specify how many days the clusters will be available to you.
--gpu-type [string]	GPU type to use for the cluster. Find available gpu types for each region with the clusters list-regions command.
--cluster-type [KUBERNETES|SLURM]	Cluster workload managers / orchestrators
--volume [string]	Storage volume ID to use for the cluster. You can reference cluster volumes from clusters storage list.

clusters update

tg beta clusters update [CLUSTER_ID]

Parameters

Flag	Description
--num-gpus [integer]	Number of GPUs to allocate in the cluster
--cluster-type [KUBERNETES|SLURM]	Cluster workload managers / orchestrators

clusters retrieve

tg beta clusters retrieve [CLUSTER_ID]

clusters delete

Usage

tg beta clusters delete [CLUSTER_ID]

clusters list

Usage

tg beta clusters list

clusters list-regions

Get configuration information per region to use in creating a GPU cluster.

tg beta clusters list-regions

Example Output

{
  "regions": [
    {
      "driver_versions": [
        {
          "cuda_version": "12.9",
          "nvidia_driver_version": "575"
        },
        {
          "cuda_version": "12.8",
          "nvidia_driver_version": "570"
        },
        {
          "cuda_version": "13.1",
          "nvidia_driver_version": "590"
        },
        {
          "cuda_version": "13.1",
          "nvidia_driver_version": "580"
        }
      ],
      "name": "us-central-8",
      "supported_instance_types": [
        "H100_SXM",
        "H200_SXM"
      ]
    },
  ]
}

clusters get-credentials

Download the cluster’s configuration and credentials to your local .kube/config file to manage Kubernetes resources.

tg beta clusters get-credentials [CLUSTER_ID]

Parameters

Name	Description
--file [Path|-]	Override the path to write the kubeconfig to. If you pass - it will print the config to stdout instead of writing to a file. default: ~/.kube/config
--context-name [string]	Name of the context to add to the kubeconfig. By default, the context name will be the cluster name.
--overwrite-existing	If there is a conflict with the existing kubeconfig, overwrite the existing kubeconfig instead of raising an error.
--set-default-context	Change the current context for kubectl to the new context.

clusters storage create

tg beta clusters storage create

Parameters

Name	Description
--region [string]	Region to create the storage volume in. required
--size-tib [intger]	Size of the storage volume you need. required
--volume-name [string]	Name for the storage volume. required

clusters storage retrieve

tg beta clusters storage retrieve [VOLUME_ID]

clusters storage list

Usage

tg beta clusters storage list 

clusters storage delete

Usage

tg beta clusters storage delete [VOLUME_ID]