Henry Xu (Aug 09 2024 at 18:55): Henry Xu (Aug 09 2024 at 18:55):I just build a working docker solution for dataverse from scratch based on the official installation guideline.
Unlike the official docker project which depends on release images and new docker specified scripts,
this solution only depends on artifacts mention in the official installation guideline and popular base images.
which means you don't have to maintain two different code set for manual deployment and docker.
Here is the project: https://github.com/link89/dataverse-docker
Philip Durbin 🚀 (Aug 12 2024 at 14:49):Looks interesting. Thanks for also mentioning it at https://groups.google.com/g/dataverse-community/c/6Q9C0KMOndo/m/ZPAaxXzeBAAJ
Philip Durbin 🚀 (Aug 12 2024 at 14:50):I added it to the list here: Document competing containerization efforts and how to choose #10522
Philip Durbin 🚀 (Aug 20 2024 at 13:46):@Henry Xu hi! Your new Docker environment was discussed during last week's meeting. You can watch it at https://harvard.zoom.us/rec/share/MtAOhUuG-abIF72gXCNQBOHshawfXG5WLdZNNiD5zbfC3FrezReOoyCFk2pvYxv9.cZCHOtECVqC0DRUt
Henry Xu (Aug 20 2024 at 16:55):Great to know that you have discussed it. I started it with the goal of building a production ready container solution for my organization . When I tried the community's container solutions, I found that the latest versions didn't run as expected, and the documentation indicated that it isn't yet production-ready. As I delved into fixing issues on top of the community's work, I noticed that containerization efforts were scattered and lacked documentation, there are multiple Dockerfiles and compose.yml files spread across the Dataverse project. So I decided to take a less risky path: automating the installation and configuration based directly on Dataverse’s production deployment documentation within a container environment from scratch.
Currently, this project can run Dataverse version 6.3 (and my own patched version) correctly via Docker Compose. Since all installation resources come from official packages (like Solr schemas), I believe it will have good compatibility with past and future versions. Kubernetes is not my priority though. Kubernetes excels at scaling stateless components . Although recent Kubernetes versions offer better support for stateful applications, most ops engineers would still hesitate to let you deploy databases in their managed Kubernetes clusters, as it increases complexity without getting any benefits .A lightweight Docker Swarm solution can offer higher performance and availability options if needed.
Last updated: Oct 30 2025 at 05:14 UTC