Skip to content

Creating and Managing Workspaces

In this quick start exercise, you will create one or more interactive Workspaces in the MemVerge AI platform. A Workspace may be assigned to any existing Project. You can even create multiple Workspaces across different Projects to experiment with various environments and resource configurations.

Prerequisites

  • Storage Volume: Ensure at least one storage volume (PVC) is created and available for your project. See Creating Storage Volumes for more information. You need at least one storage volume for each workspace.
  • Project: Have an existing Project (for example, analytics-platform or machine-learning-lab) ready so you can assign your new workspace to it.

Example Projects

If you haven’t already created a Project for your work, consider using one of these Kubernetes-friendly project names:

  • analytics-platform
  • machine-learning-lab
  • devops_infra

You can also reuse an existing Project if you already have one.

Creating Workspaces

  1. Navigate to the Projects or Workspaces Page
  2. In the left navigation bar, go to Projects, then click + Create Workspace (if available).

  3. Alternatively, you may find a Workspaces tab in your platform UI that contains a + Create Workspace button.

  4. Select the Project

  5. In the Project field, pick the project you want to associate with this Workspace (e.g., analytics-platform).

  6. Specify Workspace Details

  7. Name: Give your Workspace a clear name, like jupyter-data-lab or vscode-dev-env.
  8. Environment: Choose a Docker image that fits your use case (e.g., Jupyter with TensorFlow, VS Code + Python).

  9. Compute Resource: Select predefined resource sizes (Small, Medium, Large) or specify a custom GPU, CPU, and memory allocation.

  10. Configure Volumes

  11. Workspace Volume: Pick a storage volume to hold your personal files, notebooks, and code.

  12. Data Volumes (optional): Attach additional volumes containing datasets or shared files that your workspace needs to access.

  13. Enable Checkpointing (Optional)

  14. Turn on Checkpointing if you need the environment state saved upon pause or preemption—useful for lengthy experiments.

  15. Create the Workspace

  16. Confirm all the details, resolve any validation errors, and click Create.
  17. Check the Workspaces dashboard to monitor the status. It may appear as “Initializing” or “Pending” before becoming “Ready.”

Creating Multiple Workspaces

  • Same Project: Repeat the above steps, choosing the same project name (e.g., analytics-platform) to keep multiple Workspaces under one organizational umbrella.
  • Different Projects: If you have separate Project names (e.g., machine-learning-lab, devops_infra), select those during creation to isolate and group your Workspace resources differently.

Connecting to Workspaces

  1. Locate Your Workspace
  2. In the left navigation, go to Workspaces.

  3. Ensure the Workspace you want to connect to is in a Running or Ready state.

  4. Click “Connect”

  5. Under the Actions column, select Connect.

  6. A new tab or window opens, displaying your chosen environment (e.g., Jupyter Notebook or VS Code).

  7. Authenticate If Prompted

  8. Some configurations require a username/password or token.

  9. Enter your credentials to access the interactive environment.

  10. Begin Working

  11. For Jupyter, create or open existing notebooks, import data, and run code cells.
  12. For VS Code, open or clone projects, install extensions, and utilize built-in debugging tools.

Tip: Save critical files on persistent volumes (PVCs) to retain them across Workspace restarts or deletions. If you have connection issues or browser pop-up blockers, refer to the Troubleshooting notes in the full documentation.

Deleting Workspaces

When a Workspace is no longer needed, you can remove it to free up cluster resources:

  1. Open the Workspaces Dashboard
  2. In the left navigation menu, select Workspaces.

  3. Locate the Workspace you want to delete.

  4. Initiate Deletion

  5. Click the X or Delete icon under the Actions column for that Workspace.

  6. In the confirmation dialog, verify you have selected the correct Workspace.

  7. Confirm

  8. Click Delete to permanently remove the Workspace.
  9. The Workspace will disappear from the dashboard once deletion is complete.

Note: Always ensure you’ve saved or backed up any important data. Any ephemeral changes not stored in a persistent volume will be lost when the Workspace is deleted.