Docker Deployment¶

Barrel VectorDB provides official Docker images for easy deployment. Images are built for both linux/amd64 and linux/arm64 architectures.

Quick Start¶

Standalone Mode¶

docker run -d \
  --name barrel-vectordb \
  -p 8080:8080 \
  -v barrel-data:/app/data \
  registry.gitlab.enki.io/barrel-db/barrel_vectordb:latest

Cluster Mode¶

docker run -d \
  --name barrel-node1 \
  -e BARREL_ENABLE_CLUSTER=true \
  -e BARREL_NODE_NAME=barrel@node1 \
  -e RELEASE_COOKIE=my_secret_cookie \
  -p 8080:8080 \
  -p 4369:4369 \
  -p 9100-9200:9100-9200 \
  -v barrel-data:/app/data \
  registry.gitlab.enki.io/barrel-db/barrel_vectordb:latest

Image Tags¶

Tag	Description
`latest`	Latest stable release
`x.y.z`	Specific version (e.g., `1.0.0`)
`main`	Latest main branch build
`main-<sha>`	Specific main branch commit

Configuration¶

All configuration is done via environment variables.

Node Configuration¶

Variable	Default	Description
`BARREL_NODE_NAME`	`barrel_vectordb@<hostname>`	Erlang node name
`RELEASE_COOKIE`	`barrel_vectordb_secret`	Erlang distribution cookie

Security

Always set a strong RELEASE_COOKIE in production. All nodes in a cluster must use the same cookie.

Storage¶

Variable	Default	Description
`BARREL_DATA_PATH`	`/app/data`	Data directory path

HTTP API¶

Variable	Default	Description
`BARREL_HTTP_PORT`	`8080`	HTTP API port
`BARREL_HTTP_IP`	`0.0.0.0`	HTTP bind address

Clustering¶

Variable	Default	Description
`BARREL_ENABLE_CLUSTER`	`false`	Enable cluster mode
`BARREL_SEED_NODES`	-	Comma-separated seed nodes
`BARREL_CLUSTER_NAME`	`barrel_vectordb`	Cluster name

Seed nodes format:

BARREL_SEED_NODES="'barrel@node1.example.com','barrel@node2.example.com'"

Sharding¶

Variable	Default	Description
`BARREL_DEFAULT_SHARDS`	`1`	Default shard count for new collections
`BARREL_REPLICATION_FACTOR`	`1`	Default replication factor

Embedding¶

Variable	Default	Description
`BARREL_EMBEDDER_PROVIDER`	-	Provider: `ollama`, `openai`, `fastembed`, `local`
`BARREL_EMBEDDER_MODEL`	(provider default)	Model name
`BARREL_EMBEDDER_URL`	-	URL for ollama/openai
`OPENAI_API_KEY`	-	OpenAI API key

Performance¶

Variable	Default	Description
`BARREL_SCHEDULER_COUNT`	(auto)	Erlang scheduler count
`BARREL_ASYNC_THREADS`	`64`	Async thread pool size

Logging¶

Variable	Default	Description
`BARREL_LOG_LEVEL`	`info`	Log level: `debug`, `info`, `warning`, `error`

Docker Compose¶

Single Node¶

version: '3.8'

services:
  vectordb:
    image: registry.gitlab.enki.io/barrel-db/barrel_vectordb:latest
    ports:
      - "8080:8080"
    environment:
      BARREL_LOG_LEVEL: info
    volumes:
      - vectordb-data:/app/data
    restart: unless-stopped

volumes:
  vectordb-data:

3-Node Cluster¶

version: '3.8'

x-vectordb-common: &vectordb-common
  image: registry.gitlab.enki.io/barrel-db/barrel_vectordb:latest
  restart: unless-stopped
  networks:
    - vectordb-net

services:
  node1:
    <<: *vectordb-common
    hostname: node1
    environment:
      RELEASE_COOKIE: my_production_cookie
      BARREL_NODE_NAME: barrel@node1
      BARREL_ENABLE_CLUSTER: "true"
      BARREL_SEED_NODES: ""  # First node, no seeds
      BARREL_REPLICATION_FACTOR: "2"
    ports:
      - "8081:8080"
    volumes:
      - node1-data:/app/data

  node2:
    <<: *vectordb-common
    hostname: node2
    environment:
      RELEASE_COOKIE: my_production_cookie
      BARREL_NODE_NAME: barrel@node2
      BARREL_ENABLE_CLUSTER: "true"
      BARREL_SEED_NODES: "'barrel@node1'"
      BARREL_REPLICATION_FACTOR: "2"
    ports:
      - "8082:8080"
    volumes:
      - node2-data:/app/data
    depends_on:
      - node1

  node3:
    <<: *vectordb-common
    hostname: node3
    environment:
      RELEASE_COOKIE: my_production_cookie
      BARREL_NODE_NAME: barrel@node3
      BARREL_ENABLE_CLUSTER: "true"
      BARREL_SEED_NODES: "'barrel@node1'"
      BARREL_REPLICATION_FACTOR: "2"
    ports:
      - "8083:8080"
    volumes:
      - node3-data:/app/data
    depends_on:
      - node1

networks:
  vectordb-net:
    driver: bridge

volumes:
  node1-data:
  node2-data:
  node3-data:

With Ollama Embeddings¶

version: '3.8'

services:
  ollama:
    image: ollama/ollama:latest
    volumes:
      - ollama-data:/root/.ollama
    ports:
      - "11434:11434"

  vectordb:
    image: registry.gitlab.enki.io/barrel-db/barrel_vectordb:latest
    depends_on:
      - ollama
    environment:
      BARREL_EMBEDDER_PROVIDER: ollama
      BARREL_EMBEDDER_MODEL: nomic-embed-text
      BARREL_EMBEDDER_URL: http://ollama:11434
    ports:
      - "8080:8080"
    volumes:
      - vectordb-data:/app/data

volumes:
  ollama-data:
  vectordb-data:

Kubernetes¶

Basic Deployment¶

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: barrel-vectordb
spec:
  serviceName: barrel-vectordb
  replicas: 3
  selector:
    matchLabels:
      app: barrel-vectordb
  template:
    metadata:
      labels:
        app: barrel-vectordb
    spec:
      containers:
        - name: vectordb
          image: registry.gitlab.enki.io/barrel-db/barrel_vectordb:latest
          ports:
            - containerPort: 8080
              name: http
            - containerPort: 4369
              name: epmd
            - containerPort: 9100
              name: dist
          env:
            - name: POD_NAME
              valueFrom:
                fieldRef:
                  fieldPath: metadata.name
            - name: BARREL_NODE_NAME
              value: "barrel@$(POD_NAME).barrel-vectordb"
            - name: BARREL_ENABLE_CLUSTER
              value: "true"
            - name: BARREL_SEED_NODES
              value: "'barrel@barrel-vectordb-0.barrel-vectordb'"
            - name: RELEASE_COOKIE
              valueFrom:
                secretKeyRef:
                  name: barrel-secrets
                  key: erlang-cookie
          volumeMounts:
            - name: data
              mountPath: /app/data
          livenessProbe:
            httpGet:
              path: /vectordb/cluster/status
              port: 8080
            initialDelaySeconds: 60
            periodSeconds: 15
          readinessProbe:
            httpGet:
              path: /vectordb/cluster/status
              port: 8080
            initialDelaySeconds: 30
            periodSeconds: 10
  volumeClaimTemplates:
    - metadata:
        name: data
      spec:
        accessModes: ["ReadWriteOnce"]
        resources:
          requests:
            storage: 10Gi
---
apiVersion: v1
kind: Service
metadata:
  name: barrel-vectordb
spec:
  clusterIP: None  # Headless service for StatefulSet
  selector:
    app: barrel-vectordb
  ports:
    - port: 8080
      name: http
    - port: 4369
      name: epmd
    - port: 9100
      name: dist
---
apiVersion: v1
kind: Service
metadata:
  name: barrel-vectordb-lb
spec:
  type: LoadBalancer
  selector:
    app: barrel-vectordb
  ports:
    - port: 8080
      targetPort: 8080

Health Checks¶

The container includes a built-in health check:

curl http://localhost:8080/vectordb/cluster/status

Response when healthy:

{
  "state": "member",
  "is_leader": true,
  "nodes": ["barrel@node1", "barrel@node2", "barrel@node3"]
}

Ports¶

Port	Protocol	Description
8080	TCP	HTTP API
4369	TCP	EPMD (Erlang Port Mapper Daemon)
9100-9200	TCP	Erlang distribution
9090	TCP	Prometheus metrics (optional)

Volumes¶

Path	Description
`/app/data`	Persistent data (RocksDB, Ra logs)

Data Persistence

Always mount /app/data to a persistent volume in production to avoid data loss.

Building Custom Images¶

# Clone repository
git clone https://gitlab.enki.io/barrel-db/barrel_vectordb.git
cd barrel_vectordb

# Build for current architecture
docker build -t my-barrel-vectordb .

# Build for multiple architectures
docker buildx build \
  --platform linux/amd64,linux/arm64 \
  -t my-registry/barrel-vectordb:latest \
  --push \
  .