Skip to main content
RadishDB ships as a multi-stage Docker image. The builder stage compiles the binary with GCC; the runtime stage runs it on a minimal debian:bookworm-slim base, keeping the final image small.

​
Initial deploy

1

Pull the image

Fetch the latest RadishDB image from Docker Hub:
docker pull piee314/radishdb:latest
2

Create a named volume

RadishDB writes all data to /app/aof/radish.aof. Create a named volume so the file persists independently of the container:
docker volume create radish-data
Data lives at /app/aof/radish.aof inside the container. Mounting a named volume at /app/aof is the recommended way to make writes durable across container lifecycle events.
3

Run the container

Start RadishDB with port mapping and volume mount:
docker run -d \
  --name radishdb \
  -p 6379:6379 \
  -v radish-data:/app/aof \
  piee314/radishdb:latest
Flag breakdown:
FlagPurpose
-dRun in detached (background) mode
--name radishdbAssign a stable name for subsequent commands
-p 6379:6379Map host port 6379 to container port 6379
-v radish-data:/app/aofMount the named volume at the AOF directory
4

Verify the container is running

docker ps --filter name=radishdb
Connect and send a test command:
echo "SET ping pong" | nc localhost 6379
Expected output:
OK

​
Volume management

List all volumes and confirm radish-data exists: 
docker volume ls
Inspect volume details, including the host mount path: 
docker volume inspect radish-data
The Mountpoint field shows where Docker stores the AOF file on the host. You can copy the file out for backup: 
docker run --rm \
  -v radish-data:/data \
  debian:bookworm-slim \
  cat /data/radish.aof > radish-backup.aof

​
Container lifecycle

Stop the container gracefully: 
docker stop radishdb
Start it again (data is recovered from the AOF on startup): 
docker start radishdb
View live logs: 
docker logs -f radishdb
Remove the container (the volume and its data are preserved): 
docker rm radishdb
To also remove the volume and all stored data: 
docker rm radishdb
docker volume rm radish-data

​
How the image is built

The Dockerfile uses a two-stage build: 
# Stage 1 — compile
FROM gcc:13 AS builder
WORKDIR /build
COPY src ./src
COPY Makefile .
RUN make

# Stage 2 — runtime
FROM debian:bookworm-slim
WORKDIR /app
COPY --from=builder /build/radishdb .
RUN mkdir -p /app/aof
EXPOSE 6379
VOLUME ["/app/aof"]
ENTRYPOINT ["./radishdb"]
CMD ["--server"]
Stage 1 (builder): Pulls the full GCC 13 toolchain, copies source files, and runs make. The resulting radishdb binary is the only artifact carried forward. Stage 2 (runtime): Starts from debian:bookworm-slim — no compiler, no build tools. The binary is copied in, the AOF directory is created, port 6379 is declared, and the default command is --server. This keeps the final image lean: the GCC layer is discarded and never shipped.

​
Overriding the startup command

The default CMD ["--server"] runs RadishDB in TCP server mode. You can override it at runtime: 
# Print the version and exit
docker run --rm piee314/radishdb:latest --version

# Print help and exit
docker run --rm piee314/radishdb:latest --help
Interactive REPL mode is not useful inside a non-interactive container, but you can attach a TTY for debugging: 
docker run -it --rm \
  -v radish-data:/app/aof \
  piee314/radishdb:latest

Build docs developers (and LLMs) love

Get started for freeTalk to us