So to respond to my own toot (stack exchange style). Looks like it's Docker desktop for Mac (https://docs.docker.com/docker-for-mac/) and customising a forked rocker image (https://github.com/rocker-org/rocker/wiki/Using-the-RStudio-image). But would still be glad to hear if others have a different approach to recommend - esp. for geospatial R
Anyone know if there's a significant computing performance hit when running containerised and RStudio server?
@kidwellj R Studio Server was going to be my suggestion to an earlier toot of yours, esp if you have a Debian home server already running. I’ve never messed with docker containers on macOS. I set up and use R Studio Server at work (on RHEL) alongside R Shiny Server and have had no issues.
@wj Thanks for confirming my suspicions. The key thing for me is to support reproducible research, so Docker offers a nice way to preserve (and allow meaningful duplication of) a development environment...
@kidwellj I second the RStudio server suggestion by @wj.
Regarding performance, I run several Rstudio server instances each in their own KVM virtual machine on an Ubuntu server with a Xeon E3-1275 CPU and 32GB RAM and for a single user (me) performance has been a non-issue. My guess is Docker uses even less overhead than KVM. But I strongly suggest RStudio server over RStudio desktop. I've a (slightly outdated) write-up of setting up such a system on my blog.
The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!