Dockerfiles for R and Python sessions and jobs to be used RStudio Server Pro, Launcher, and Kubernetes.
The Docker images built from these Dockerfiles are intended to be used for R and Jupyter sessions and jobs with RStudio Server Pro (RSP), Launcher, and Kubernetes.
Note: These Docker images are not equipped or intended to be used to run RStudio Server Pro within a Docker container. Visit the rstudio/rstudio-docker-products repository for images built for that purpose.
For more information about RStudio Server Pro and Launcher, refer to the Launcher Overview on the RStudio Solutions website.
For more information about how to use these images with RStudio Server Pro and Launcher, refer to the RStudio support article on Using Docker images with RStudio Server Pro, Launcher, and Kubernetes.
Built images are available from the rstudio/r-session-complete repository on Docker Hub.
These images include the following layers:
- Base OS
- RSP session components
- System packages required for R, R packages, and RStudio Professional Drivers
- One version of R
- R packages (devtools, tidyverse, shiny, and rmarkdown)
- One version of Python
- Jupyter Notebooks, JupyterLab, and RSP/RSC notebook extensions
- Python packages
- RStudio Professional Drivers
When builds on master are passing and you are ready to release a new version of the image on Docker Hub, you can tag a release on master using the following steps:
-
Bump versions of RSP session components in Dockerfiles in a PR, merge when the build passes
-
Run the following commands on the
master
branch, editing the version string as needed (where1.3.XXXX-X
is the RSP version):export VERSION=1.3.XXXX-X git commit -am "Release for image version ${VERSION}" --allow-empty git tag -a ${VERSION} -m "${VERSION}" git push origin ${VERSION} git push
Note: Add the
-preview
suffix to tagged images that correspond to preview release versions of RSP. Images related to preview builds are intended for testing purposes, subject to being removed, do not fall under our support agreement, and are not recommended for use in production. -
The new images will get built on Travis CI and get pushed to the rstudio/r-session-complete repository on Docker Hub.