Einstein Toolkit School

Owned by Craig Willis
Last updated: Jul 28, 2017

About the Einstein Toolkit School Workbench

The NDS Labs Workbench is a general-purpose platform for launching applications via Docker images. For the Einstein Toolkit School, the Workbench supports a single Jupyter image with the Einstein Toolkit dependencies.  You can view the Dockerfile or access the image via Dockerhub.

Docker containers

There are a few things you should know about Docker containers:

  • While they behave like virtual machines, they are different in one important way: when you restart a container (or application in Workbench) local changes that are not preserved on a mounted volume are lost. See "Home directories" below for information about mounted directories in the Jupyter containers.
  • This means that other changes (installed packages, python libraries, etc) that are not in the home directory will be lost after restart.

User accounts

Each user will be given an account at the workshop. You should not need to register for an account to access the system. 

Catalog

The Catalog contains the list of applications available on the system.  The File Manager allows you to easily copy data or manage files stored on Workbench.  The Jupyter/Einstein Toolkit application is a customized Jupyter SciPy notebook with the dependencies required to compile and experiment with the Einstein Toolkit software. Select "Add" to add an instance of the application to your dashboard.

Applications

The Applications page contains the list of configured and/or running applications in your account. For the Workshop, you will likely have a single Jupyter application running.  From the Applications page you can:

  1. Access the running application endpoint (authentication is required using the same username/password)
  2. Launch, Shutdown and Remove running applications
  3. Edit or view the application configuration options (none are required for this workshop)
  4. Access the container console (this is not needed for Jupyter, since the application has it's own built-in terminal)

Home directories

In the Workbench system, each user has a home directory. This home directory is mapped to /home/jovyan/work in the Jupyter application and also available under /home/<user>. Files stored in these two directories will be preserved across container restarts and accessible from other applications in Workbench.

Resource limits

  • Each Jupyter container is limited to 2 cores and 2GB RAM.
  • Applications will be stopped if the user hasn't logged into Workbench for 24 hours. Stopped applications can be restarted.

Troubleshooting/FAQ

TBD