Accessing the SAFE Research Infrastructure

Accessing the SAFE Research Infrastructure

SAFE Research Data Center

Errikos Melissinos, Research Assistant - May 2022

Documentation:
YouTube video (31 mins). Sections: Specs and Intro Use the Cluster through the Web Browser Get access to your SAFE drive on Windows Access the Cluster on a terminal (via SSH) Run Code through TMUX Get access to your SAFE Drive on Mac Run Cluster on PyCharm Why not use Spyder Run Cluster on VS Code How to find and use shared SAFE Drives

Connect to the Jupyter Hub¶

You can connect in two very easy steps:

Be in campus or have your uni VPN active (https://www.rz.uni-frankfurt.de/45186991/30_Verbinden_AnyC).
Go to http://compute.safe-frankfurt.de and log in using your HRZ credentials.

From the hub:

you can manage your files,
you can write interactive notebooks,
you can access a terminal (aka command prompt),
- in the terminal you can also add your own libraries with "pip install \<library name> --user"
you can create Python, Julia and R files
and you can run the files.

Personal drive on the network file system¶

Windows¶

You can find detailed walthrough in the documentation https://compute.safe-frankfurt.de/docs/map_network_drive.pdf

Right click on "This PC"
click on "Map Network Drive..."
As Folder set \\storage.safe-frankfurt.de\<HRZ user>
then for the credentials enter ntsamba1\<HRZ user> and then your HRZ password.

Mac¶

For mac:

just open a Finder window
go to "Go>Connect to Server" from the menu bar
type smb://storage.safe-frankfurt.de/<HRZ user> as the server (you can also add it to the favourites)
enter your HRZ credentials (not your MAC user name)

Terminal - SSH¶

You can also connect to the cluster by using SSH in the terminal (or command prompt).

In [ ]:

# type
# ssh <HRZ username>@safe-gpu01.safe-frankfurt.de
# and then enter the HRZ Password
# then the terminal is running on the server and you can have access to the 
# SAFE drives.

TMUX¶

Finally you can also run your code directly from your own Terminal (or Command Prompt) application without using the hub. You can also find info on this on the documentation (https://compute.safe-frankfurt.de/docs/tmux.pdf).

In [ ]:

tmux new -s <session_name> # create a session
# press: ctrl+b, d
# to detach from a session
tmux ls # list running sessions
tmux attach -t <session_name> # to attach to running session
tmux kill-session -t <session_name> # to close specific session outside of tmux
exit # to close session from within the session

PyCharm¶

It is also possible to access the cluster directly through an IDE. Here I will go over PyCharm and Visual Studio Code. You can also find resources in the documentation (https://compute.safe-frankfurt.de/docs/presentation.pdf).

The steps are relatively easy:

Open PyCharm
Select ssh from the side bar
Click on "New Connection"
Enter your HRZ user name in "Username" and enter "safe-gpu01.safe-frankfurt.de" in "Host".
Then click on "Check Connection and Continue"
Then enter your HRZ password.
Now select a directory for your project (after you find your folder).
Give your password again. And you are probably done!

Maybe there is still an issue with the interpreter. Then you can select "Add interpreter" at the bottom right. Go to SSH interpreter and then enter the Host and the HRZ credentials one more time.

Visual Studio Code¶

Finally we will look at Visual Studio Code.

The steps are the following:

Make sure that you have the Remote - SSH Extension
Click on the bottom left green corner in an new window,
click on "Connect to Host",
click on "Add New SSH Host...",
type "ssh <HRZ user>@safe-gpu01.safe-frankfurt.de",
finally you will be asked for the password a couple of times.

Also you may need to install the Python extension on the server side.

In Visual Studio Code you can also run Julia and R.