If you are using Moodle Assignments or a git repository to transfer Jupyter content to your students, each student will receive their own copy of the content. This can be a problem if you are working with large data files and resources. In this case, it is possible to host your data on a NAS and directly connect it to your Jupyter.

6.1 – Share Provided by the Jupyter Service

We can provide a share, with a few limitations:

• reasonable data sizes (see below for terabytes sized data)
• the data will be visible for any ETHZ user
• write access is exclusively for Windows access
• the share will be deleted when the JupyterHub is EoL

Ask for a share, and provide us a single user, or preferably an AD group which will get full access to the root directory. You may manage additional access permissions on directories you create.

6.2 – Custom NAS Setup

It is possible to connect a NFS share provided by you. The setup needs to be adjusted to make a connection possible from within the JupyterHub and may involve discussions with your ISG. However, if you plan to access petabytes of data, not duplicating this data on the JupyterHub NAS is likely the preferable way.

6.2.1 – Configuration

Some important points to notice are:

• There needs to be an export for the Cluster-IPs 129.132.110.64/27 on a read-only basis.
• Complex authentication protocols like Kerberos are not supported.
• Access must be granted either read for all, or based on an user ID of 1000.

6.2.2 – Connecting JupyterHub

Connecting your NAS to a JupyterHub has to be done by the Jupyter support team. Usually, we will map your NAS drive to a read-only folder /data within JupyterLab. This way, each person who accesses the same JupyterHub will have access to the shared network files within this folder. This ensures that no data is copied between the instances of JupyterLap for the different students.