Skip to content

Latest commit

 

History

History
61 lines (34 loc) · 1.93 KB

work_on_remote_server.md

File metadata and controls

61 lines (34 loc) · 1.93 KB
Raw

Work on Remote Server

Tips and cookbooks related to working on a remote server. Modified from this file.

Run graphical Linux applications on Windows 10

https://seanthegeek.net/234/graphical-linux-applications-bash-ubuntu-windows/

ssh login

    ssh -Y username@remote_server

e.g.

    ssh -Y [email protected]

Inquire information from netCDF files

    ncdump -h xxxx.nc

Quick visualization of netCDF files

    ncview xxxx.nc

ssh login without typing password

  1. generate the keys (private + public)

     ssh-keygen -t rsa

all use the defaults settings.

  1. copy the public key to the remote server

     ssh-copy-id username@remote_server

Open Jupyter notebook on remote server and access it at local machine

  1. Remote server terminal

     jupyter notebook --no-browser --port=8899 --ip=127.0.0.1

port 8899 might be unavailable, in which case a different port # may be used.

In case you want to use Jupyter lab, you can run the following command instead:

    jupyter lab --no-browser --port=8899 --ip=127.0.0.1

  1. Local machine terminal

     ssh -NL localhost:8899:localhost:8899 username@remote_server

use the assigned port number in step 1 to replace 8899 when applicable.

In case you are not allowed to log in the remote server (e.g. tigressdata.princeton.edu) directly from an outside network, a jump server (e.g. nobel.princeton.edu) is needed if you do not want to use VPN:

    ssh -N -J username@jump_server -L localhost:8899:localhost8899 username@remote_server

  1. Local machine browser

     http://localhost:8899

Or copy the full link (containing token info) seen in the remote server in step 1.

The jupyterhub server at Princeton

https://github.com/Resplandy/climate-hpc/blob/master/jupyterhub.md