Docker
We provide three Docker images for various deployments: "all-in-one" for quick testing and experiments, a development image for debugging problems, and a production image for all other cases.
All-in-one image is documented in the README.md file in the repository. The rest are covered below.
Production image
Our production image ghcr.io/ferretdb/ferretdb
is recommended for most deployments.
It does not include PostgreSQL or other backends, so you must run them separately.
You can do that with Docker Compose, Kubernetes, or other means.
Setup with Docker Compose
The following steps describe a quick local setup:
Store the following in the
docker-compose.yml
file:services:
postgres:
image: postgres
environment:
- POSTGRES_USER=username
- POSTGRES_PASSWORD=password
- POSTGRES_DB=ferretdb
volumes:
- ./data:/var/lib/postgresql/data
ferretdb:
image: ghcr.io/ferretdb/ferretdb
restart: on-failure
ports:
- 27017:27017
environment:
- FERRETDB_POSTGRESQL_URL=postgres://postgres:5432/ferretdb
networks:
default:
name: ferretdbpostgres
container runs PostgreSQL that would store data in the./data
directory on the host.ferretdb
runs FerretDB.Start services with
docker compose up -d
.If you have
mongosh
installed, just run it to connect to FerretDB. It will use credentials passed inmongosh
flags or MongoDB URI to authenticate to the PostgreSQL database. You'll also need to setauthMechanism
toPLAIN
. The example URI would look like:mongodb://username:password@127.0.0.1/ferretdb?authMechanism=PLAIN
See Authentication for more details.
If you don't have
mongosh
, run the following command to run it inside the temporary MongoDB container, attaching to the same Docker network:docker run --rm -it --network=ferretdb --entrypoint=mongosh mongo \
"mongodb://username:password@ferretdb/ferretdb?authMechanism=PLAIN"
You can improve that setup by:
- securing connections with TLS;
- adding backups.
Find out more about:
Development image
The development image ghcr.io/ferretdb/ferretdb-dev
contains the debug build
of FerretDB with test coverage instrumentation, race detector,
and other changes that make it more suitable for debugging problems.
It can be used exactly the same way as the production image, as described above.