AI Notebooks - Utilisation de l’espace Workspace (EN)

Base de connaissances

AI Notebooks - Utilisation de l’espace Workspace (EN)


Icons/System/eye-open Created with Sketch. 324 vues 17.11.2025 Cloud / AI Notebooks

Objective

This guide details how AI Notebooks workspaces work and the related best practices.

Instructions

What is a notebook's workspace?

AI Notebooks are managed Jupyter or Visual Studio Code notebooks, linked to compute resources (CPUs, GPUs) and storage.

This storage includes both the remote containers you have attached to your notebook, and your notebook's default storage.

Indeed, every notebook has by default an internal object storage volume attached, fully managed by OVHCloud, which is your workspace storage and is located at /workspace.

How do AI Notebooks' workspace work?

When your notebook is created, a /workspace directory is automatically created. It contains configuration information such as the Machine Learning framework you have chosen to launch your notebook (Miniconda, PyTorch, TensorFlow, ...), but also all your installed python libraries. This directory also allows you to store your data (datasets, codes, etc.).

When you stop your notebook, your workspace (/workspace) is automatically saved and pushed to your Object Storage. This means that all files contained in your /workspace directory will be saved for future use.

Therefore, the next time you restart your notebook, your workspace will be pulled directly from your Object Storage into /workspace. You will not need to reinstall your python libraries, re-import your data or anything else you put in the /workspace directory.

This workspace is saved as long as your notebook is in STOPPED state.

There is also a /data folder in which you can store your data. This solution is more efficient but ephemeral, meaning that it will not be persisted if you close your notebook or if it crashes. In other words, the data contained within this directory cannot be recovered if the notebook leaves the RUNNING state.

Here is a graph summarizing your notebook's workspace and ephemeral storage usage during its various states:

image

Deleting workspace files

If you run the ls -a command in your notebook's /workspace directory, you will see all the files and directories that your workspace contains, including hidden ones (whose names begin with a dot (.)).

Hidden files and directories often contain temporary files and configuration files.

It may be worth deleting some of these files, such as cache, to free up storage space in your workspace.

Be careful, some configuration files must not be removed under any circumstances, to prevent workspace initialization errors. This is the case of the .workspace.initialized file. Never delete it and therefore never delete the entire contents of the /workspace directory.

If you delete this file and stop your notebook, you will not be able to restart it in the future, as it will reach the Error status.

Notebook deletion and workspace backup

When you delete a notebook, its workspace data stored in Object Storage is not automatically cleaned up. You can find and delete this data in the notebooks_workspace container of your Object Storage, under the notebook ID directory.

This means that even if a notebook is deleted, its state stored in Object Storage (including your notebook files) remains accessible.

Workspace billing

The first 10GB of the /workspace directory are free during 30 consecutive days. After this period, the price of OVHcloud Object Storage is applied for each GB. To learn how we bill AI Notebooks, refer to the Billing and lifecycle documentation.

Feedback

Please send us your questions, feedback and suggestions to improve the service:

If you need training or technical assistance to implement our solutions, contact your sales representative or click on this link to get a quote and ask our Professional Services experts for a custom analysis of your project.

Articles associés