CloudBeaver Full JDBC Docker

Custom CloudBeaver Community image builder that keeps the standard CloudBeaver CE drivers and automatically pre-bundles additional JDBC drivers already defined by DBeaver Community when their runtime JARs can be resolved from Maven.

Why this repository exists

CloudBeaver Community does not dynamically download JDBC drivers like DBeaver Desktop. A driver must:

1. already have a driver definition in the DBeaver platform;
2. have its runtime JARs available in CloudBeaver driversLocation;
3. be registered in CloudBeaver as a resource, bundle, and enabled driver.

This repository automates steps 2 and 3 for Maven-backed JDBC drivers and then builds CloudBeaver from source into a Docker image.

What "full" means

The generated image contains:

• every driver already pre-bundled by CloudBeaver Community;
• additional DBeaver Community drivers whose connection URL is JDBC, whose driver definition exposes a CloudBeaver-compatible drivers/... bundle path, and whose required JARs are available from Maven;
• a build report at /opt/cloudbeaver/full-drivers-report.json showing what was added and what was skipped.

It intentionally does not pretend that every DBeaver driver can be redistributed. Drivers that require vendor-local/proprietary files, have no Maven artifact, or are not JDBC are skipped. Review vendor licenses before publishing the resulting image to a public registry.

Build

The Dockerfile is self-contained: it installs the documented CloudBeaver build toolchain in a builder stage, clones CloudBeaver/DBeaver sources, generates the driver pack, builds CloudBeaver, and packages the final runtime image.

docker build \
  --build-arg CLOUDBEAVER_REF=devel \
  --build-arg DBEAVER_REF=devel \
  -t cloudbeaver-full:local \
  .

For production, pin CLOUDBEAVER_REF, DBEAVER_REF, and DBEAVER_COMMON_REF to compatible commits/tags instead of tracking devel.

The build is large because it compiles CloudBeaver and DBeaver. BuildKit cache is strongly recommended.

Run

docker compose up -d

Open:

http://localhost:8978

The Compose file enables anonymous access and private/custom connections for the discovery-service POC. Review these settings before exposing CloudBeaver outside a trusted network.

Inspect installed drivers

docker exec cloudbeaver-full \
  find /opt/cloudbeaver/drivers -maxdepth 2 -type f | sort

View the generated report:

docker exec cloudbeaver-full \
  cat /opt/cloudbeaver/full-drivers-report.json

Driver generation policy

Edit config/driver-policy.json.

The default policy:

• scans DBeaver plugins/**/plugin.xml;
• selects network JDBC drivers (jdbc: URL);
• skips embedded/file drivers unless they are already part of stock CloudBeaver;
• requires a drivers/<name> prebundle mapping in the DBeaver definition;
• requires Maven-backed runtime JARs;
• skips drivers that require unresolved ${...} Maven versions;
• keeps stock CloudBeaver registrations untouched;
• aggregates dependencies by CloudBeaver driver bundle to avoid duplicate modules.

You can explicitly exclude a driver with its full DBeaver ID, e.g. provider:driver-id.

Project layout

.
├── Dockerfile
├── docker-compose.yml
├── config/
│   └── driver-policy.json
├── scripts/
│   └── generate_full_driver_pack.py
├── .github/workflows/
│   └── build.yml
└── Makefile

Useful commands

make build
make up
make logs
make report
make down

Important limitations

• This is a JDBC-focused full pack. Non-JDBC/NoSQL drivers are not automatically added.
• A Maven artifact being downloadable does not automatically mean you are allowed to redistribute it publicly. Check the database vendor's license.
• A driver may be present but still require database-side configuration, TLS certificates, native libraries, or vendor authentication components.
• CloudBeaver and DBeaver branches must be compatible. Pin compatible commits for repeatable production builds.
• The goal here is broad connection/discovery support. Database-specific UI features may differ from DBeaver Desktop for drivers that run through generic JDBC metadata.

Security

The supplied Compose configuration is optimized for an internal POC. Do not expose anonymous private connections directly to the Internet. Put CloudBeaver behind an authenticated reverse proxy/gateway and restrict network egress to approved database networks for production.