Accessing JupyterHub data

Unlike many JupyterHub deployments, your data is yours and have many different ways to access it. Thus, we don’t just have jupyter.cs, but a whole constellation of ways to access and do your work, depending on what suits you best for each part.

Your data (and as an instructor, your course’s data) can be accessed many ways:

  • On jupyter.cs.
  • Via network drive on your own computer as local files.
  • On Aalto shell servers (such as kosh.aalto.fi).
  • On other department/university workstations.

On Paniikki and Aalto computers

On Paniikki, and the Aalto servers kosh.aalto.fi, lyta.aalto.fi, brute.aalto.fi, and force.aalto.fi (and possibly more), the JupyterHub is available automatically. You can, for example, use the Paniikki GPUs.

Data is available within the paths /m/jhnas/jupyter. The path on Linux servers is also available on the hub, if you want to write portable files.

Name Path on hub Path on Linux servers
personal notebooks /notebooks /m/jhnas/jupyter/u/$nn/$username/
course data /coursedata /m/jhnas/jupyter/course/$course_slug/data/
course instructor files /course /m/jhnas/jupyter/course/$course_slug/files/
shared data /m/jhnas/jupyter/shareddata/ /m/jhnas/jupyter/shareddata/
Variable seen above Meaning
$username Your Aalto username
$nn The two numbers you see in echo $HOME (the last two digits of your Aalto uid, id)
$course_slug The short name of your course.

You can change directly to your notebook directory by using cd /m/jhnas/jupyter/${HOME%/unix}.

You can link it to your home directory so that it’s easily available. In a terminal, run /m/jhnas/u/makedir.sh and you will automatically get a link from ~/jupyter in your home directory to your user data.

Permission denied? Run kinit in the shell - this authenticates yourself to the Aalto server and is required for secure access. If you log in with ssh keys, you may need to do this.

Remote access via network drive

Basic info

Name Network drive path
personal notebooks smb://jhnas.org.aalto.fi/$username/
course data smb://jhnas.org.aalto.fi/course/$course_slug/data/
course instructor files smb://jhnas.org.aalto.fi/course/$course_slug/files/
shared data smb://jhnas.org.aalto.fi/shareddata/

You can do a SMB mount, which makes the data available as a network drive. You will have the same copy of the data as on the hub - actually, same data, so edits immediately take effect on both places, just like your home directory. You must be on an Aalto network, which for students practically means you must be connected to the Aalto VPN (see vpn instructions) or use an Aalto computer. The “aalto” wifi network does not work unless you have an Aalto computer.

  • Linux: use “Connect to Server” from the file browser. The path is smb://jhnas.org.aalto.fi/$username. You may need to use AALTO\username as your username. If there is separate “domain” option, use AALTO for domain and just your username for the username.
  • Mac: same path as Linux above, “Connect to Server”. Use AALTO\your_username as the username.
  • Windows: \\jhnas.org.aalto.fi\$username, and use username AALTO\your_username. Windows sometimes caches the username/password for a long time, so if it does not work try rebooting.

You can also access course data and shared data by using jhnas.org.aalto.fi/course/ or jhnas.org.aalto.fi/shareddata/.

See also

Mounting network drives in Windows is the same instructions, but for Aalto home directories. Anything there should apply here, too.

Using GPUs

One problem with our JupyterHub so far is that we don’t have GPUs available. But, because our data is available to other computers, you can use the Paniikki: Computer Lab For Students GPUs (quite good ones) to get all the power you need. To do this, you just need to access the Jupyter data on these classroom computers.

Terminal: First, start a terminal. You can navigate to your data following the instructions above: cd /m/jhnas/jupyter/${HOME%/unix}. From there, navigate to the right directories and do what is needed.

File browser: Navigate to the path /m/jhnas/jupyter/u/$nn/$username, where $nn is the two numbers you see when you do echo $HOME in a terminal. To open a terminal from a location, right click and select “Open in Terminal”.

Now that you have the terminal and the data, you can do whatever you want with it. Presumably, you will start Jupyter here - but first you want to make the right software available. If you course tells you how to do that using an Anaconda environment, go ahead and do it. (Please don’t go installing large amounts of software like anaconda in the Jupyter data directories - they are for notebooks and small-medium data.)

Using the built-in anaconda, you can load the Python modules with module load anaconda and start Jupyter with jupyter notebook:

Start jupyter with the anaconda module.

Note that now, you need to module load anaconda, not anaconda3 like the image shows.