Have a Question?
RStudio on ARCC2
Connecting to RStudio
The RStudio editor is available on ARCC2 via the Open OnDemand web portal under the ‘Interactive Apps’ section (note that you must be on campus wifi or connected to UC’s VPN to access this page):

Selecting ‘RStudio Server’ will bring you to a page with a selection of parameters for your job, including desired partition, total wall time, and resource quantity. Many jobs will only require one CPU; two exceptions to this are jobs that take advantage of parallelism via an R package like doParallel or future (or use a package that takes advantage of these parallelism packages behind-the-scenes) and jobs that need more RAM than one core provides (see the RAM/CPU chart at the bottom of the job parameters page). Users requiring >64 cores should consider submitting their R script as a batch job through the command line instead, as our RStudio implementation does not support multi-node jobs. For more information on this, refer to our R Software on ARCC2 knowledge base article.

Clicking the ‘Launch’ button will initiate the job with the requested resources and bring you to a new page displaying the job status on an interactive session card; once this changes from ‘Queued’ to ‘Running’, click the ‘Connect to RStudio Server’ button to connect to the session (this opens a new browser tab and may take a few minutes to load).

When finished with your session, terminate your RStudio instance using the power icon in the top-right corner of the pane to prevent warning messages (or possible errors) during the startup of your next RStudio job, and click the ‘Delete’ button on the OnDemand session card.

Why am I getting Status Code 502 errors?
Status code 502 errors usually result from timeouts while connecting to the RStudio session. This is often caused by one of two things: attempting to load a command-intensive ~/.bashrc file, or restoring/loading a large environment or dataset from a previous session. We recommend using a minimal ~/.bashrc when loading RStudio via OnDemand and not saving large workspace files upon session closure; in fact, we have disabled RStudio’s ‘autosave workspace on exit’ behavior in an attempt to prevent these errors. If you suspect you need to clear an old cached workspace, you can try the following shell command (note that you may lose any unsaved work):
mv ~/.local/share/rstudio ~/.local/share/rstudio.backup