MEDS Computational Server Guide

What is a Computational Server?

Purpose: Shares vast amounts of computing resources, especially CPU and random-access memory, over a network. You can think of it as a multi-user computer that people can utilize simultaneously.

Why Use R-Studio Desktop vs. Workbench / Server?

R-Studio Desktop is an R IDE that works with the version of R you have installed on your local Windows, Mac OS X, or Linux workstation.  R-Studio Desktop is a standalone desktop application that in no way requires R-Studio Workbench or RStudio Server.

R-Studio Workbench and R-Studio Server are Linux server applications that provide a web-browser-based interface to the version of R running on the server. You may want R-Studio Workbench or RStudio Server to:

• Access larger CPU and memory footprints

• Leverage compute resources more efficiently

• Control access to data in a centralized manner

• Ensure the R package / Python library versions used are standardized

• R-Studio Workbench provides Enterprise-ready tools like authentication, high-availability and load balancing, multiple versions of R, multiple R sessions per user, security features, and more.

https://support.rstudio.com/hc/en-us/articles/217799198-What-is-the-difference-between-RStudio-Desktop-and-RStudio-Server-

Computing servers that are available to MEDS students

MEDS students have access to two computing servers that have been configured for MEDS coursework: 

1. 1. workbench-1.bren.ucsb.edu (this server will primarily be used for coursework)
   2. workbench-2.bren.ucsb.edu (this server will be used for capstone work)

The process of connecting to these servers is the same. Below you will find directions for connecting to workbench-1, however you can apply the same concept to connecting to workbench-2. 

_

Connecting to Workbench-1 (Connection Methods) 

Like channels on a walkie talkie, TCP / UDP ports are different 'channels' that computers can talk to each other over the internet. There are general standards / guidelines for which applications / services should use which ports.

HTTPS - https://workbench-1.bren.ucsb.edu (using secure webpage - default port 443)

SFTP - workbench-1.bren.ucsb.edu (using CyberDuck - default port 22)

SSH - {username}@workbench-1.bren.ucsb.edu (secure shell - default port 22)

Bren Credentials

Username: Bren Desktop Account Username (same as ucsbnetid)

Password: Bren Desktop Account Password (specific to Bren, assigned during orientation, cannot be changed)

VPN not required if on UCSB networks (NCEAS, Bren, Campus Housing, etc.)

_

How to access IDE for workbench-1.bren.ucsb.edu

1. Connect to campus VPN. 
2. Navigate to https://workbench-1.bren.ucsb.edu
3. Log in with Bren credentials.
4. You can start an RStudio, JupyterLab or Jupyter Notebooks session by selecting "New Session"
5. Select whichever IDE you'd like to use to launch a session. You can have as many sessions running as you need.

_

How to transfer files using a file manager (like Finder, File Explorer, etc) for workbench-1.bren.ucsb.edu

1. Download CyberDuck (sftp client).
2. Connect to campus VPN. 
3. Open up CyberDuck
4. Select "Open Connection"
5. Choose "SFTP (SSH File Transfer Protocol)" and input the following: Note: Use your Bren credentials for username/password.
6. Once you're signed in, you can drag and drop files or use the "Actions" button for more options.
7. To navigate to /courses/ folder, select root directory from drop-down menu then courses:

_

You also have the ability to SSH if you prefer. Here's how you can SSH into workbench-1.bren.ucsb.edu

1. Connect to campus VPN. 
2. Open up Terminal (mac) or GitBash (windows)
3. Run the following command: ssh {username}@workbench-1.bren.ucsb.edu
4. Type in your Bren password
5. You'll know if you connected successfully if the last line says "{username}@workbench-1" like the picture below.

_

General structure

The directories on the servers that students will mostly use are the /Users/, /courses/, and /capstone/ folders. 

• /Users/{username} is a private directory that is unique to the user and no other user can access. It is also the user's default working directory when interacting with the server.
• /courses/ is a public directory that some instructors will use to upload datasets, documents, etc. for their courses.
• /capstone/ is a directory that will be used for capstone projects. When students are assigned their capstone groups, they will have a folder created for their team. They can use this folder to collaborate on any projects related to their capstone.

Example of how to read in data from courses folder on server:

IMPORTANT!

The server is a completely new machine! You'll need to set up your GitHub credentials like you did on your local machine. Please refer to steps 6 and 7 of the MEDS installation guide to set up your PAT on the server: https://ucsb-meds.github.io/MEDS-installation-guide/

As always, please feel free to reach out at [e-mail] or on the #compute channel on Slack with any questions.