plasmabio / plasma

It seems adding the start script mentioned here has no effect in plasma. I guess because there is a custom tljh_plasma/entrypoint/entrypoint.sh script.
I did change the file and added source ./start here

plasma/tljh-plasma/tljh_plasma/entrypoint/entrypoint.sh

Lines 43 to 46 in 7883a6a

    
           cd ${IMAGE_DIR} 
        
           # execute the notebook process as the given user 
        
           exec su - $NB_USER -m -c '"$0" "$@"' -- "$@"

Did I miss something or is there a proper way to execute the start script. If not maybe it is possible to add a line to entrypoint.sh?

Thank you
Jan

Thanks @jvail for reporting.

I believe this is right. The tljh-plasma plugin uses a custom entrypoint to be able to start the servers as the users on the host machine. Maybe it should try to execute the start script if it exists, and default to the current exec command otherwise?

Thanks for looking into this issue. I am afraid, I am neither a docker nor a jupyter hub pro. But the little 'hack' into the entrypoint.sh did work for us - but I am not sure if it is the best option to enable the start script. I thought it might be a good idea to make the start script available by default in plasma.

The use case was to run a pip install -e . in order to install a lib in dev mode from within the repo every time the server starts for a specific user. Because otherwise the dev install will live in home/jovyan and wont reflect any changes made after the image is built.

	cd ${IMAGE_DIR}

	# execute the notebook process as the given user
	exec su - $NB_USER -m -c '"$0" "$@"' -- "$@"

start script support