SDBM Jena Fuseki

A customized Apache Jena Fuseki Docker image for the Schoenberg Database of Manuscripts (SDBM), providing a SPARQL 1.1 endpoint over the SDBM linked data.

Based on stain/jena-fuseki.

What It Is

The SDBM tracks the history of pre-modern manuscripts through auction records, dealer catalogs, and other sources, exposing that provenance data as RDF. This image runs Fuseki pre-configured with an sdbm TDB dataset, a custom SPARQL query interface, and a security policy that makes queries publicly accessible while restricting administrative operations.

Customizations

What	How
Base image	ubuntu:24.04 (replaces Alpine)
Fuseki version	3.14.0
Dataset	Pre-configured sdbm TDB1 dataset (sdbm.ttl) at /fuseki/databases/sdbm
Authentication	ADMIN_PASSWORD is required at startup; the container exits if it is missing
shiro.ini	Always overwritten at startup from the image copy, with ${ADMIN_PASSWORD} substituted — redeployment reliably picks up a new password
Access control	SPARQL query endpoints and the UI are public (anon); all other operations require admin credentials
Query UI	Custom qonsole-config.js and dataset.html configure the SPARQL query interface for SDBM

Requirements

• Docker
• ADMIN_PASSWORD environment variable (required — container will not start without it)

Running

Quick start (foreground)

docker run -p 3030:3030 -e ADMIN_PASSWORD=changeme sdbm/jena-fuseki

Fuseki will be available at http://localhost:3030/. The sdbm dataset is at http://localhost:3030/sdbm.

Local development

Copy jena.env.example to jena.env and edit as needed:

cp jena.env.example jena.env
# edit jena.env — do not commit this file

Then run with --env-file:

docker run -p 3030:3030 --env-file jena.env sdbm/jena-fuseki

Environment variables

Variable	Required	Default	Description
ADMIN_PASSWORD	Yes	—	Password for the admin user
JVM_ARGS	No	-Xmx1200m	JVM heap settings, e.g. -Xmx3g for large uploads
VIRTUAL_HOST	No	—	Hostname for a reverse proxy (e.g. jena.example.com)
TDB	No	1	Set to 2 to use TDB v2

Data Persistence

Fuseki data lives in the Docker volume /fuseki. Without a named volume or bind mount, data is lost when the container is removed.

Named volume (recommended)

docker volume create fuseki-data
docker run -d --name fuseki -p 3030:3030 -e ADMIN_PASSWORD=changeme -v fuseki-data:/fuseki sdbm/jena-fuseki

Bind mount

docker run -d --name fuseki -p 3030:3030 -e ADMIN_PASSWORD=changeme -v /path/to/data:/fuseki sdbm/jena-fuseki

Loading Data

For large datasets, load directly from the command line using the bundled load.sh script rather than the web UI. Stop Fuseki first, then load into the volume:

docker stop fuseki
docker run --rm -v fuseki-data:/fuseki -v /path/to/rdf/files:/staging sdbm/jena-fuseki ./load.sh sdbm /staging/sdbm.ttl

If no filenames are given, load.sh loads all matching files under /staging (*.rdf, *.ttl, *.owl, *.nt, *.nquads, and their .gz variants).

Restart Fuseki after loading:

docker restart fuseki

Building

docker build -t sdbm/jena-fuseki .

Common Operations

# View logs
docker logs fuseki

# Stop
docker stop fuseki

# Restart (retains volume and port config)
docker restart fuseki

# Upgrade: pull new image, recreate container against existing volume
docker pull sdbm/jena-fuseki
docker stop fuseki && docker rm fuseki
docker run -d --name fuseki -p 3030:3030 -e ADMIN_PASSWORD=changeme -v fuseki-data:/fuseki sdbm/jena-fuseki

Access Control Summary

The shiro.ini configuration grants anonymous access to:

• The Fuseki web UI (/, /index.html, /dataset.html, static assets)
• SPARQL query endpoints (/**/query, /$/sdbm)
• Status and ping endpoints

All other endpoints — including dataset management, data upload, and graph store writes — require HTTP Basic authentication as admin.

Upstream Resources

• Apache Jena Fuseki documentation
• stain/jena-docker source
• Apache Jena TDB
• Jena users mailing list

---

🤖 AI Usage Disclosure: README substantially rewritten by Claude (Anthropic) to document SDBM-specific customizations, required configuration, and operational usage.