One can install Jupyter notebook on an AWS EC2 instance via the following command
|
|
And then start it via the command
|
|
However, to access it from a client computer may need some tweaking in both AWS EC2 instance and the SSh connection.
AWS EC2 setting
- Make sure the EC2 instance’s inbound port in security groups include the jupyter server’s port, normally 8888
SSH setting
-
use the following command or alike connect to the EC2 instance
1 2# change port number accordingly ssh -i "key.pem" ubuntu@ec2-ip-address -L localhost:8888:localhost:8888This command opens a SSH tunnel, when one tries to access the port 8888 at the client machine, it will be forwarded to the remote EC2 instance in the same port 8888.
-
after successully connecting to it, run the following command:
1 2tmux new -s jupyter jupyter-lab --no-browserThese commands will start a jupyter server on the EC2 instance within the tmux session (this will avoid the stop of the server when connection is disrupted)
-
now in the browser input ‘localhost:8888’ to access the jupyter server. You may need password or token to access it, depending on jupyter server configuration.
This ssh tunnel forwarding is essential when AWS EC2 instance is run in a private network, but may not be needed for the ones with public IPs.
Update 05/30/2023:
Actually, this tunnel forwarding isn’t necessary, when one starts jupyter using the following command:
|
|
By adding the option ‘–ip 0.0.0.0’, jupyter allows connection from any IP address; in default, it only allows localhost connection and that is why it needs tunnel forwarding.
References
-
AWS jupyter tunnel: https://gist.github.com/jakechen/faf0500132d46d83517004bbfedbe5de
-
Jupyter notebook configuration: https://docs.aws.amazon.com/dlami/latest/devguide/setup-jupyter-config.html
-
AWS jupyter setup: https://towardsdatascience.com/setting-up-and-using-jupyter-notebooks-on-aws-61a9648db6c5
Last modified on 2023-05-23