Creating a Workspace

A Workspace is an interactive environment—often running Jupyter Notebook or VS Code—where users can develop and run code in real time. This section walks you through creating a new Workspace, including selecting an environment image, configuring resources, and attaching storage volumes.

---

Open the Workspace Creation Dialog

1. Navigate to the Workspaces Dashboard

   • In the left navigation bar, select the Workspaces Icon .
   • You will land on the Workspaces dashboard, which shows all current Workspaces.

   

2. Launch the “Create Workspace” Dialog

   • Click on the aqua "+Create Workspace" button on the right of the dashboard.
   • A pop-up form will appear, prompting you to specify key details such as the Workspace Name, Project, Environment, Compute Resource, Workspace Volume, Dataspace Volume, and Checkpointing.

   

Fill in the Workspace Details

1. Basic Information

   • Name: Provide a descriptive name (e.g., data-science-lab or ai-model-dev).
   • Project: Select the project under which this workspace will be organized and billed.

2. Environment

   • Choose from predefined Environments (e.g., Jupyter + PyTorch, Jupyter + TensorFlow, VS Code + Python) or create your own Custom Environment, where you can upload your own image:
   • If Creating a Custom Environment, you will need to choose if you wish to enable nginx proxy or shared memory.
   • The Environment you select determines which libraries and tools are pre-installed.

3. Compute Resource

   • Pick a predefined resource profile (e.g., CPU Only, Small, Medium, Large) or define a Custom configuration by specifying the exact number of GPUs, CPU cores, and memory (MiB).
   • Ensure you match the resource profile to your workload’s requirements.

4. Workspace Volume

   • Configure a persistent Workspace Volume to store your personal files, notebooks, and code.
     • Select an existing storage volume or create a new one by specifying Size and Access Mode.
     • Indicate the Mount Path (e.g., /home/jovyan) where these files will be accessible inside the container.

5. Data Volumes (Optional)

   • If your project requires additional datasets, you can attach Data Volumes.
   • Provide the Size, Access Mode, and a Mount Path where the data should be mounted inside the workspace.

6. Checkpointing (Optional)

   • Toggle Checkpointing to On if you want the workload’s state to be saved whenever it is preempted or paused.
   • This is particularly useful for long-running computations or training sessions.

7. Resolve Any Form Validation Errors

   • If required fields (e.g., Name, Project, or Volume details) are missing or invalid, the form will highlight errors in red.
   • Provide correct values to clear the errors.

8. Click “Create”

   • Once all fields are valid, click Create to provision your new Workspace.
   • The platform will schedule the Workspace with the specified resources and environment image.

9. Verify Deployment

   • Return to the Workspaces dashboard to see your newly created Workspace in the list.
   • Status indicators may show “Initializing” or “Pending” until it becomes fully ready.
   • Note: It make take several minutes for the workspace to become "Ready", especially if the image has to be downloaded before the workload can be fully provisioned. To monitor the provisioning progress, click the workspace to see the details page and navigate to "Conditions". Any errors will be shown here. You will need to look at the pod logs using kubectl for further debugging.