Advanced: Deploy repositories in your own VLab

Author

VLab Course Team

This optional guide is for advanced users who want to understand how the VLab5 RStudio environment itself is deployed and how to reproduce something similar in their own VLab.

1) Deployment of RStudio Server in a VLab

The deployment process is documented in this D4Science support ticket:
https://support.d4science.org/issues/23536#note-15

The RStudio server setup for VLab5 is managed in the following repository:
https://code-repo.d4science.org/gCubeSystem/rstudio-d4science/src/branch/fisheriesatlas-vlab5

Key files include:

  • 07_vlab5_github_repositories.sh - defines which GitHub repositories (stable/dev) are preloaded into the environment.
    View script
  • Dockerfile - defines the RStudio server Docker image and how repositories and dependencies are installed.
    View Dockerfile

⚠️ This is more technical and system-level. If you are interested in customizing or deploying your own RStudio VLab instance, we can schedule a dedicated walkthrough.

2) Managing a Global renv cache in Docker images

When building custom RStudio Docker images for your own VLab, it is possible to preconfigure a shared global cache for renv. This ensures that all projects within the VLab use the same package cache, saving time and storage.

This is done by adding a line to .Renviron.site during Docker build:

RUN echo 'RENV_PATHS_CACHE="~/blue-cloud-dataspace/GlobalFisheriesAtlas/cacheRenv"' \
  >> "${R_HOME}/etc/Renviron.site"

Reference: https://support.d4science.org/issues/29739

Key points:

  • You can set RENV_PATHS_CACHE to any directory, depending on your VLab setup.
  • Using ~/blue-cloud-dataspace/ is recommended: it is accessible across all users and all VLab instances (though slightly slower than local storage).
  • .Renviron.site ensures the cache is global to the image - every repository opened in RStudio will reuse this shared cache.

Effectively:

  • Each repository still has its own renv.lock file (to ensure reproducibility).
  • But the package binaries are stored once in the shared cache, instead of reinstalling them for every project.

3) The truly simple path (what to consider)

If you want to deploy the same setup as VLab5 in your own VLab, here are the essentials to look at and adapt per VLab:

  1. D4Science ticket (reference implementation & notes):
    • See: https://support.d4science.org/issues/23536#note-15
      (Contains key decisions & parameters used for the Fisheries Atlas VLab RStudio.)
  2. RStudio Server image for D4Science (to adapt):
    • Repo: https://code-repo.d4science.org/gCubeSystem/rstudio-d4science/src/branch/fisheriesatlas-vlab5
    • Files to review:
      • Dockerfile
        https://code-repo.d4science.org/gCubeSystem/rstudio-d4science/src/branch/fisheriesatlas-vlab5/Dockerfile
      • 07_vlab5_github_repositories.sh (preloads the course repositories)
        https://code-repo.d4science.org/gCubeSystem/rstudio-d4science/src/branch/fisheriesatlas-vlab5/07_vlab5_github_repositories.sh

If this is interesting for your team, we can do a short technical walkthrough per VLab, since each environment may require minor adjustments (base image, permissions, storage mounts, etc.).

4) Summary

  • Basic users: You don’t need to worry about this - VLab5 is already configured.
  • Advanced users: You can adapt the Dockerfile and scripts for your own VLab deployment, preload specific repos, and configure a global renv cache.
  • Best practice: Use a cache in blue-cloud-dataspace for shared accessibility across all VLab instances.

✅ If you are interested in deploying or customizing your own RStudio environment with these settings, reach out - we can provide a step-by-step session on adapting the Dockerfile and repository scripts for your use case.