Docker containerized apps (for scientists)

The most popular containerization stack.

Containerization is a broad family of technology. Docker is so prominent a member of that family that it’s easiest to define the other in relation to Docker. “Like Docker, but …” is a common way to introduce a new containerization tool.

Which is unfortunate, because Docker is not the most comprehensible or well-explained member of the family, not the one I typically need. Docker is well supported but has awful terminology, is riddled with confusing analogies, and has poor explanations. Fortunately, Julia Evans explains at least the filesystem, overlayfs, by example. The Google best practice page also has good illustrations that make it clear what’s going on. See also the docker cheat sheet, as noted by digithead, who also explains the annoying terminology:

Docker terminology has spawned some confusion. For instance: images vs. containers and registry vs. repository. Luckily, there’s help, for example this stack-overflow post by a brilliant, but under-appreciated, hacker on the difference between images and containers.

• Registry - a service that stores image repositories
• Repository - a set of Docker images, usually versions of the same application
• Image - an immutable snapshot of a running container. An image consists of layers of file system changes stacked up on top of a base image.
• Container - a runtime instance of an image

Essentially, with Docker, we provide a recipe for building a reproducible execution environment, and the infrastructure ensures that environment exists for our program. The recipe is ideally encapsulated in the Dockerfile. The trade-off is that setup is a bit clunkier. The upside is that the second setup (and every one after that) is, in principle, effortless and portable.

It’s often not the best-suited tool for my research-oriented needs, which often means heavy GPU requirements, which Docker handles poorly. The things I do that do not need a GPU are usually tiny, and the massive Docker engine feels inefficient and ungainly to use, so I skip it. In between, there are people trying to scale web apps who love it. There’s an ecosystem targeting researchers called apptainer but since no one cares much about researcher needs, the ecosystem is not as rich. YMMV.

sudo groupadd docker
sudo usermod -aG docker $USER

Error response from daemon: Get https://registry-1.docker.io/v2/:
net/http: request canceled while waiting for connection
(Client.Timeout exceeded while awaiting headers)

{ "dns": [ "8.8.8.8", "8.8.4.4" ]}

## command-line R
docker run --rm -ti rocker/r-base
## Rstudio
docker run -e PASSWORD=yourpassword --rm -p 8787:8787 rocker/rstudio
# now browse to localhost:8787. L