BigR

A virtual machine with a largish amount of ram and cpu compared to a desktop computer, hosted in the Campus Data Center next to Farm I/II.

Current specs: 8 cpu, 40 GB ram, 150 GB disk

URL: https://bigr.biogeo.ucdavis.edu

General Usage

Primary usage is via the website which is RStudio Server. It's just like RStudio except it's in the browser and executes code and reads files from the server.

Login with your campus username and password.

Outline

1. Create & Test Code locally (on your computer)
2. Connect to BigR
3. Upload Data
4. Upload Code
5. Run Code
   • How to setup a job
   • Keep tabs on the work
   • Get notified when done
6. Retrieve results

BigR is a little different as you connect to an RStudio session via a web interface.

Login with your campus kerberos/cas username and password (we authenticate directly against the campus servers and never store your password)

Upload Data and Code

• You can upload data directly in the web interface
• If you need to upload a lot of data, or think the connection might break and require a resume use SFTP (e.g. Fileziila)

Run Code

You can run code 2 different ways. For small scripts that won't take very long, running it directly in the RStudio interface with Run or Source should work quite well. You can also interactively edit and debug your scripts in the interface. It's really just like using RStudio on a Desktop/Laptop. Try this method out, on large jobs it might appear to hang but you should wait it out. If you think it's been way to long you can reset the Session or contact Alex via Help to kill the job.

The second method is by calling your code in Batch or Script. You will need to connect using SSH (Git shell in RStudio work for this and may be the best option for windows users in addition to Git Bash).

ssh <user>@bigr.biogeo.ucdavis.edu

The recommended method for being able to (dis/re)connect is byobu

Tips

• If you need to bulk upload or download use SFTP software.
• Need an R package added, request, rather than installing yourself so that it will be available to all users.

RStudio Tips

• If you have a session that crashes, and it keeps trying to reload then you need to wipe it by deleting ~/.rstudio/suspended-session rm -r ~/.rstudio/suspended-session (Or email Alex for help)
• Clear your Environment after running code to free up the resources. See the Environment tab in the upper right pane of RStudio.