I have been working with rstudio from a bioconductor container, e.g.

# Base R image
FROM bioconductor/bioconductor_docker:RELEASE_3_19

and for example I start it like this:

docker run -v .:/home/rstudio --rm -ti -e PASSWORD=example -p 8787:8787 example

It has been working great. However, I think when it starts it tries to reload the cache or something and it can take forever. How can I disable that feature?

The help for docker run indicates that the --rm argument (as one might imagine) removes the docker container after running. I would imagine not removing the docker container would be faster upon repeated use.

Disclaimer: I am a mere (but daily) Docker user, by no means an expert on the internals of Docker.

There is to my knowledge no caching of containers. In fact, the --rm flag immediately deletes the container after termination. There is two concepts to keep in kind with docker, being image and container. When you run docker run you launch a container based on the image. Images are immutable, once created from the Dockerfile. If you make changes to the running container then you have to commit these, creating a new image. What is possible is to start and pause, and then resume a running container, but this does not seem to be the case here.

That having said, I cannot really answer your question. The initial download of the image may take some time depending on internet connection, but the actual run, given the image has been downloaded, should take no time. I use Bioc-based containers daily and cannot say I ever encountered this issue.

Can you provide more details? Is this Linux, or WSL2 something like this?