The Ohio Supercomputer Center is providing no-cost access to cloud computing resources for classroom instruction. OSC has a dedicated web-browser interface for its substantial Linux computer systems designed for classroom use. The interface - class.osc.edu- provides access to OSC clusters with both shell terminal and interactive applications such as RStudio and Jupyter Notebook through a customized installation of Open OnDemand. These applications can be tailored-made with class-specific application environments and packages, thereby removing the need for professors and students to manage different versions of software and packages.
Other resources provided to classroom projects that can be requested include: shared project storage space for classroom material; reservation for compute nodes to be used during class and/or lab sessions; and OSC introduction presentations. Please contact OSCHelp@osc.edu to talk to an OSC staff member about classroom support options that are available to you.
Wilbur, we are big fans of OOD and would like to expand ours to class usage. Do you have a github or similar, or would you consider releasing some of your setup?
Most of the /etc/ood/config files are in https://github.com/OSC/osc-ood-config/tree/master/class.osc.edu. We customize the dashboard using the locales file by providing an override for the welcome_html key https://github.com/OSC/osc-ood-config/blob/13b89d6c877130b8d18febd813f8829f5189aabe/class.osc.edu/locales/en.yml#L24-L81 and use the <h2 class="apps-section-header-blue"> - those headers are the same styling that the OnDemand dashboard in app sharing mode uses. We add large icons for Jupyter and RStudio and hardcode the links to the corresponding web forms for each app. This is a kinda hack-solution but it gets around not having a nicer iconified dashboard (something we will work on in OnDemand 2.0). We also add helpful links for each class and a link to Canvas at the top (OSU’s Carmen).
Our RStudio and Jupyter apps provide dropdown to select the class you are in so that, for example, Jupyter notebooks the professor adds to project space are copied to the student’s home directory during launch, so they can have a copy of the notebook to work with.
Each class has its own project/group so that we can assign a reservation of node(s) to that class to ensure there are no queue times and student’s RStudio/Jupyter sessions start immediately.
The configuration for the RStudio and Jupyter apps are split between
