4.1 Installing RStudio Connect

In this session#

In this session, you:

Getting from A to B isn't going to be like climbing a mountain, it's like climbing a tree (picking which fork to follow amongst numerous branches).

Refer to the product documentation at:

This course is about teaching you to fish, not giving you fish.

Open up the guide, this will be a useful resource throughout the course and will help you decide which branch to pick.

Typical questions

How big a server?
- It all depends on what your content will be doing in R.
- It is best to learn over time and crawl-walk-run where possible.
- The support article Configuration and sizing recommendations provides some suggestions.
Docker?
- Connect is designed to run on a persistent server (think Sharepoint).
- As you will see, it internally runs and isolates user submitted code, manages processes, sandboxes permissions, etc.
- You will get into this more, but all these aspects make it a bit of an odd fit for Docker. That said, it's possible, but must be run in a privileged container.
- Running in a privileged container may reduce the amount of process isolation that Docker offers. Whether it is still worth dockerizing is a choice you must make.
What operating system?
- Connect is supported on RHEL, Ubuntu and SUSE
- Refer to Requirements for RStudio Connect
- For data science work we recommend Ubuntu

Installation of RStudio Connect requires root permissions.

Where can R be installed?

which R

Rscript -e 'R.home()'

RStudio Connect is all about "deploy and forget". This means you don't want existing content to break when R is upgraded.

To achieve this, RStudio Connect supports multiple versions of R.

Tip

Our recommended best practice is to always install R from pre-compiled binaries

Because Connect supports multiple versions of R, the upgrading decision is easy:

Upgrading R is as simple as installing another version of R side-by-side.

Important: avoid upgrading by using your repository:

Q: What type of server should Connect run on? (Memory, CPU, etc).

A: It depends. You should be able to answer this question by the end of the course.

General guidelines here
It all depends on what your content will be doing in R.
- Best to learn over time and crawl-walk-run where possible.

Download page links to current version:

Next complete the lab exercise.

Signs of success: