Here are the steps to use Rstudio server on cluster nodes
Note - VPN is necessary for off campus access.
Find the least loaded compute node from here or qstat -f
Use the corresponding compute node from the links below
| Host Name | Rstudio Link |
|---|---|
| Node1 | Click Here |
| Node2 | Click Here |
| Node3 | Click Here |
| Node4 | Click Here |
| Node5 | Click Here |
| Node6 | Click Here |
| Node7 | Click Here |
| Node15 | Click Here |
| Node17 | Click Here |
| Node18 | Click Here |
| Node19 | Click Here |
| Node20 | Click Here |
| Node22 | Click Here |
| Node23 | Click Here |
| Node24 | Click Here |
| Node25 | Click Here |
| Node26 | Click Here |
| Node27 | Click Here |
| Node28 | Click Here |
Note - To open a terminal via rstudio, please set the following -
Tools → Global Options → terminal - Uncheck “Connect with WebSockets”
Note - The steps below are optional. A minimal central R package directory is already included in the package library path that includes packages like tidyverse and arrow. Additional packages could be added from the big central package directory on request.
To add the central package directory - Add the following to your R session -
.libPaths(c("/software/R-packages", .libPaths()))
Note - It is possible to make the above change persistent by adding the following to ~/.Renviron -
R_LIBS_USER="/software/R-packages"
But Rstudio session start up will slow down a lot. The preferred option is to request addition of the package to the minimal package directory.