This docker image was developed over the course of 2 days and 1 night shift (The night shift with unlimited leftover food was proven very useful to me).
The purpose of the docker image is to make it easier to run Jupyter notebooks on the Orchestra computing cluster in the Harvard Medical School. This docker image was put together in a bit of rush, therefore it is potentially hacky and will need some effort to polish it.
This docker image contains two main aspects:
(1) The most important and useful files are probably the back-end Jupyter spawner script stored in folder /srv/JupyStra/. The most updated version can be found here: https://github.com/msherman997/JupyStra
(2) the front-end web interface, based on LAMP. This docker image was based on greyltc/lamp.
The TODO items that are on our list are:
(1) ensure port collision do not occur, especially in the context of if multiple users are using this script.
(2) ensure that server_spawner is aware of any existing Jupyter session and if there are existing Jupyter sessions, do not launch new ones, but just re-use the existing session.
(3) ensure the security is in accordance with all existing policies within the computing cluster environment.
(4) better STDOUT and STDERR monitoring and exception handling.
PS. For installing packages in greyltc/lamp, instead of using apt-get install, or yum install, this docker will use pacman -S.