RStudio Server

We recommend using RStudio through Open OnDemand as the main GUI interface to R

RServer will change the permissions on the working directory to 777. Do not run from HOME directory or your main blue directory

See the warning above. We recommend only running rserver from a directory 'deep' inside your personal blue storage directory tree to make sure the unavoidable permission change doesn't expose your files to other users e.g. /blue/GROUP/USER/somewhere/in/the/directory/tree/

To run RStudio Server in a job on HiPerGator and connect to the session from a web browser on your local computer do the following:

• Create a job script similar to the following job script, but with your own resource request. E.g. run.sh

#!/bin/bash
#SBATCH --job-name=rserver
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=4
#SBATCH --mem=8gb
#SBATCH --time=48:00:00
#SBATCH --output=rserver_%j.log
module purge; module load R
rserver

• Submit the job

sbatch run.sh

• Once the job starts see the connection information in the job log. E.g.

$ cat rserver_9999999.log

Starting rserver on port 37546 in the /blue/somegroup/someuser/rserver_test directory.
Create an SSH tunnel with:
ssh -N -L 8080:c12345a-s42.ufhpc:37546 albert.gator@hpg.rc.ufl.edu
Then, open in the local browser:
http://localhost:8080

• Use the above information to create an ssh tunnel from your local computer to the rserver instance.
  • Once you enter the ssh line and get connected, the terminal will appear to hang and just sit there.
    This is normal, proceed to the next step, leaving the terminal running in the background.
• Open http://localhost:8080 in the web browser on your local computer.

If you run into any issues open a support ticket.

If you need to access environment variables outside of the small list RStudio Server allows through like SLURM_CPUS_ON_NODE to automate parallelization use the following patter in your job script

module purge; module load R
env > ~/.Renviron
rserver
rm -f .Renviron