Install and configure CLI

This section explains the procedure for installing and configuring the researcher Command Line Interface.

Note

This section refers to CLI v2 only. You can find installation instructions for the CLI v1 here. CLI v1 was deprecated in v2.20.

Prerequisites

• Ensure the Improved command line interface toggle is enabled, under General settings → Workloads

• Only clusters that are version 2.18 or above are supported.

Installing the CLI

1. Click the Help (?) icon in the top right corner

2. Select Researcher Command Line Interface

3. Select the cluster you want the CLI to communicate with

4. Select your computer’s operating system

5. Copy the installer command and run it in the terminal or download the binary file for Windows OS

6. Follow the installation process instructions

7. Click Enter to use the default values (recommended)

Testing the installation

To verify the CLI client was installed properly

1. Open the terminal

2. Run the command runai version

Configuring the CLI

Follow the steps below to configure the CLI.

Setting the control plane URL

The following step is required for Windows users only. Linux and Mac clients are configured via the installation script automatically

Run the command runai config set --cp-url <CONTROL_PLANE_URL>. This will also create the config.json file in the default path.

Authenticating the CLI

After installation, sign in to the NVIDIA Run:ai platform to authenticate the CLI.

1. Open the terminal on your local machine

2. Run runai login

3. Enter your username and password in the NVIDIA Run:ai platform's sign-in page

4. Return to the terminal window to use the CLI

Setting a default cluster

If only one cluster is connected to the account, it is set as the default cluster when you first sign-in. If there are multiple clusters, you must follow the steps below to set your preferred cluster for workload submission:

1. Open the terminal on your local machine

2. Run runai cluster and select the desired cluster from the interactive menu.

Alternatively:

1. Open the terminal on your local machine.

2. Run runai cluster list to find the desired cluster name.

3. Run the following command runai cluster set <CLUSTER_NAME>.

Setting a default project

Set a default working project, to easily submit workloads without mentioning the project name in every command.

1. Open the terminal on your local machine.

2. Run runai project and select the desired cluster from the interactive menu.

Alternatively:

1. Open the terminal on your local machine.

2. Run runai cluster list to find the desired project name.

3. Run the following command runai project set <PROJECT_NAME>

4. If successful, the following message is returned project <PROJECT_NAME> configured successfully.

Validating the configuration

To view the current configuration run runai config generate --json

Installing command auto-completion

Auto-completion assists with completing the command syntax automatically for ease-of-use. Auto-completion is installed automatically. The interfaces below require manual installation:

Installation instructions for ZSH

1. Edit the file ~/.zshrc

2. Add the following code:

   autoload -U compinit; compinit -i
   source <(runai completion zsh)

Installation instructions for Bash

1. Install the bash-completion package

2. Choose your operating system: Mac: brew install bash-completion

   Ubuntu/Debian: sudo apt-get install bash-completion

   Fedora/Centos: sudo yum install bash-completion

3. Edit the file ~/.bashrc and add the following lines:

   [[ -r “/usr/local/etc/profile.d/bash_completion.sh” ]] && . “/usr/local/etc/profile.d/bash_completion.sh”
   source <(runai completion bash)

Installation instructions for Windows

Add the following code in the powershell profile:

runai.exe completion powershell | Out-String | Invoke-Expression
Set-PSReadLineKeyHandler -Key Tab -Function MenuComplete

For more completion modes options, see Powershell completions.