Skip to main content

RStudio

RStudio is an integrated development environment (IDE) for the R programming language, with limited support for other programming languages (including Python, bash, and SQL). RStudio provides a powerful graphical environment for importing data in a number of formats (including CSV, Excel spreadsheets, SAS, and SPSS); manipulating, analyzing, and visualizing data; version control with git or SVN; a graphical R package manager that provides point/click search/installation/uninstallation of R packages from its substantial ecosystem (including the Bioconductor repository, which provides almost 1500 software tools “for the analysis and comprehension of high-throughput genomic data.”); and many other features.

RStudio Server is a client/server version of RStudio that runs on a remote server and is accessed via the client’s web browser. A graphical file manager allows file upload/download from Ceres via web browser.

SCINet Options for RStudio

SCINet users can use RStudio in one of the following ways:

  1. To run RStudio and access data on your local workstation, download the open source RStudio Desktop.
  2. To run an RStudio Server virtual machine on and access data in Amazon Web Services, contact the SCINet VRSC.
  3. NEW: RStudio Server on Ceres in Open OnDemand.
  4. RStudio Server on Ceres through SCINet VPN or via ssh-tunnel.

RStudio Server on Ceres

RStudio Server is currently available on Ceres using a Docker image (imported into Singularity) provided by the Rocker project. The provided geospatial image provides not only geospatial libraries, but also LaTeX / publishing libraries, and Tidyverse data science libraries. Other R packages can be easily installed into your home directory from within RStudio.

Running RStudio Server on Ceres allows SCINet users to access any data on Ceres that they can access from the command line (SSH).

Users can either run RStudio Server in Open OnDemand or by manually submiting a SLURM job script and connecting to the server through SCINet VPN or via ssh-tunnel. The first method (via Open OnDemand) is much easier-to-use and is recommended over the second method.

A few Ceres-specific notes:

  1. RStudio terminal (bash command shell): since RStudio Server is running in a container with a Debian base image, you won’t be able to access software environment modules (e.g., that you would normally see when logging into Ceres and issuing the module list command), as those are installed on the (CentOS) host.
  2. Data access: your home directory is mounted inside the RStudio Server container, and the VRSC has configured Singularity to mount the /project directory. $TMPDIR (which on a compute node is per-job local scratch on the compute node’s direct attached storage that gets deleted at the end of SLURM job) is mounted inside the container at /tmp.
  3. Software installation: The provided SLURM job script creates a ~/.Renviron file in your home directory that allows RStudio to install additional R packages into your home directory (the container image is immutable). Installing a lot of R libraries may contribute to the default 15G soft limit quota on your home directory being surpassed. To overcome this issue you can move R directory from your home directory to your project directory and create a symbolic link to the new location. In RStudio click on “Terminal” and enter the following commands substituting your values:
cd
mkdir /project/<your_project_dir>/<account_name>
mv R /project/<your_project_dir>/<account_name>/
chgrp -R proj-<your_project_dir> /project/<your_project_dir>/<account_name>/R
chmod -R g+s /project/<your_project_dir>/<account_name>/R
ln -s /project/<your_project_dir>/<account_name>/R R

The mv and chgrp commands may take longer time depending on how much data you have in the R directory. You will need to restart R for this new location to take effect. In the “console” window type quit() and click on “Start New Session”.

  1. Default working directory: By default RStudio sets working directory to your home directory. You can change the default by clicking on Tools -> Global Options… In the General tab click on Browse, then on the three dots on the right and enter path to the folder, e.g. /project/. You can then navigate further to a subdirectory. When done click on "Choose" button, and then on "OK". This will be the default working directory that will show in the panel on the right next time you start RStudio or restart R in the RStudio.

RStudio Server in Open OnDemand

In your browser go to https://ceres-ood.scinet.usda.gov/ and log in using the GUI login instructions.

After logging into Open OnDemand, click on “Interactive Apps” in the menu on the top and select “RStudio Server: Ceres”.

Modify default values if needed and click on the “Launch” button at the bottom of the page. It may take up to several minutes for the new session to start. Once it starts a new button “Connect to RStudio Server” will appear. Clicking on the button will open RStudio IDE in a new browser tab. When done using RStudio, return to the previous tab to stop the RStudio session by clicking on the red “Delete” button.

Note that while the RStudio session is running you can connect and disconnect to/from the RStudio Server multiple times. Every click on the “Connect to RStudio Server” button will open a new browser tab.

Also note that once you click on the “Launch” button and a new session starts, requested resources will be allocated to your job and won’t be available to other users even if you don’t run anything in the RStudio. Please be considerate, request only resources that you need for your tasks and remember to delete the session when done.

For tips and tricks refer to the previous section and to the RStudio Tutorial at https://www.dataquest.io/blog/tutorial-getting-started-with-r-and-rstudio/#tve-jump-173bb2584fe.

RStudio Server through VPN or ssh tunnel

To use RStudio Server on Ceres, a user submits a SLURM job script. A default job script that should suffice for most users is provided.

After a user is done using RStudio Server, they should save their work in RStudio, and then stop RStudio Server by cancelling the job with the slurm scancel command.

Starting RStudio Server

The instructions below tell to use RStudio version 3.6.0 . To see other versions available on Ceres issue:

ls -l /reference/containers/RStudio/

If you wish to use a different version, replace 3.6.0 below with the version number you choose.

Note that older versions are not supported. If the version you chose is not working, try the latest one before contacting VRSC.

  1. (If using VPN) Connect to SCINet VPN (see video instructions):
    VPN Using the OpenConnect Client
    VPN Using the Cisco AnyConnect Client

  2. Log into Ceres via SSH (see the Quick Start Guide for instructions).

  3. Submit the RStudio SLURM job script with the following command: 
    sbatch /reference/containers/RStudio/3.6.0/rstudio.job

    (Optional) By default, this SLURM job is limited to a 4 hour time limit, 1 processor core, and 6600 MB memory. To customize, see the section Requesting Additional Compute Resources below.

  4. After the job has started, view the “$HOME/rstudio-JOBID.out” file for login information (where JOBID is the SLURM job ID reported by the sbatch command). Here is an example of such a file. Your file will have different job number, user name, password, compute node name and port number. 
    [jane.user@sn-cn-8-1 ~]$ sbatch /reference/containers/RStudio/3.6.0/rstudio.job
Submitted batch job 214664

[jane.user@sn-cn-8-1 ~]$ cat ~/rstudio-214664.out
VPN Users:

1. Connect to SCINet VPN and point your web browser to http://ceres14-compute-3-eth.scinet.local:44200

2. log in to RStudio Server using the following credentials:

   user: jane.user
   password: 4wjRJfpIvQDtKdDZpmzY

SSH users:

1. SSH tunnel from your workstation using the following command (macOS or Linux only;
   for how to enter this in PuTTY on Windows see the Ceres RStudio User Guide)

   ssh -N -L 8787:ceres14-compute-3-eth.scinet.local:44200 jane.user@ceres.scinet.usda.gov

   and point your web browser to http://localhost:8787

2. log in to RStudio Server using the following credentials:

   user: jane.user
   password: 4wjRJfpIvQDtKdDZpmzY

When done using RStudio Server, terminate the job by:

1. Exit the RStudio Session ("power" button in the top right corner of the RStudio window)
2. On the Ceres command line, issue the command

   scancel -f 214664
  5. Instructions for this step differ for when you use SCINet VPN or SSH Port Forwarding.

Using VPN

If using VPN point your web browser to the listed hostname / port (in the example above, http://ceres14-compute-3-eth.scinet.local:44200), then enter your SCINet user name and the temporary password (valid only for this job only; in this example 4wjRJfpIvQDtKdDZpmzY) screenshot of signing into RStudio in a web browser

SSH Port Forwarding (instead of VPN)

Note Before performing the instructions below, first read and follow the instructions in steps 2-4.

macOS / Linux / Windows + Windows PowerShell users
  1. Open a new macOS/Linux terminal window or a new Windows PowerShell window and enter the SSH command listed in the job script output file. In this example: 
    ssh -N -L 8787:ceres14-compute-3-eth.scinet.local:44200 jane.user@ceres.scinet.usda.gov

    There will be no output after logging in. Keep the window / SSH tunnel open for the duration of the RStudio session.

  2. Point your browser to http://localhost:8787. Enter your SCINet user name, and one-time password listed in the job script output file.

Stopping RStudio Server

  1. Click the Quit Session (“power”) button in the top-right corner of the RStudio window (see picture below), or select “File > Quit Session…” screenshot of the Quit Session power button in RStudio
  2. After the “R Session has Ended” window appears, cancel the SLURM job from the Ceres command line. E.g., if the job ID is 214664: 
    [jane.user@sn-cn-8-1 ~]$ scancel -f 214664

    Be sure to specify the scancel -f / --full option as demonstrated above.

  3. (If using SSH Port Forwarding instead of VPN) Close the terminal / PuTTY window in which the SSH tunnel was established.

Requesting Additional Compute Resources

The default job resources (4 hour time limit, 1 processor core, 6600 MB memory) may be customized by:

  • sbatch command-line options, e.g., to specify an 8-hour wall time limit, 16 G memory, and 2 processor cores (= 4 hardware threads): 
    sbatch --time=08:00:00 --mem=16G --cpus-per-task=4 /reference/containers/RStudio/3.6.0/rstudio.job
  • Copying the job script to a directory one has write access to and modifying the appropriate SLURM #SBATCH directives.
  • Using the Ceres Job Script Generator to create a new job RStudio Server job script
    • In the Job Script Template drop down menu, select RStudio Server