Docker Compose Deployment Strategy

How to package and deploy Nexus 3.0 to new servers.

Docker vs Docker Compose

Nexus uses Docker Compose because it has 20+ environments, each with vault and operational containers.

Image Sharing

Docker is smart about layers: - FalkorDB image: ~500MB - 40 containers (20 envs x 2) share the SAME base image - Total storage: ~2GB for infrastructure + data - Data grows independently per environment

Compose File Structure

version: '3.8'

networks:
  nexus-storage-network:
    driver: bridge

services:
  # SECURE ENVIRONMENTS (vault only)

  user-vault:
    image: falkordb/falkordb:latest
    container_name: nexus3-user-vault
    restart: unless-stopped
    networks:
      - nexus-storage-network
    ports:
      - "6610:6379"
    volumes:
      - /data/nexus3/user/vault:/var/lib/falkordb/data
    command: redis-server --loadmodule /var/lib/falkordb/bin/falkordb.so

  locker-vault:
    image: falkordb/falkordb:latest
    container_name: nexus3-locker-vault
    restart: unless-stopped
    networks:
      - nexus-storage-network
    ports:
      - "6720:6379"
    volumes:
      - /data/nexus3/locker/vault:/var/lib/falkordb/data
    command: redis-server --loadmodule /var/lib/falkordb/bin/falkordb.so

  # STANDARD ENVIRONMENTS (vault + operational)

  track-vault:
    image: falkordb/falkordb:latest
    container_name: nexus3-track-vault
    restart: unless-stopped
    networks:
      - nexus-storage-network
    ports:
      - "6640:6379"
    volumes:
      - /data/nexus3/track/vault:/var/lib/falkordb/data
    command: redis-server --loadmodule /var/lib/falkordb/bin/falkordb.so
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 5s
      timeout: 3s
      retries: 3

  track-operational:
    image: falkordb/falkordb:latest
    container_name: nexus3-track-operational
    restart: unless-stopped
    networks:
      - nexus-storage-network
    ports:
      - "6641:6379"
    volumes:
      - /data/nexus3/track/operational:/var/lib/falkordb/data
    depends_on:
      track-vault:
        condition: service_healthy
    command: redis-server --loadmodule /var/lib/falkordb/bin/falkordb.so

  # Repeat pattern for each environment...

Replication Init Service

Add a one-shot service to configure replication after containers start:

  replication-init:
    image: redis:alpine
    container_name: nexus3-replication-init
    networks:
      - nexus-storage-network
    depends_on:
      - track-vault
      - track-operational
      - context-vault
      - context-operational
      # ... all standard environments
    entrypoint: /bin/sh
    command: |
      -c '
      sleep 5
      # Track replication
      redis-cli -h nexus3-track-vault CONFIG SET requirepass $TRACK_VAULT_PASS
      redis-cli -h nexus3-track-operational CONFIG SET masterauth $TRACK_VAULT_PASS
      redis-cli -h nexus3-track-operational REPLICAOF nexus3-track-vault 6379

      # Context replication
      redis-cli -h nexus3-context-vault CONFIG SET requirepass $CONTEXT_VAULT_PASS
      redis-cli -h nexus3-context-operational CONFIG SET masterauth $CONTEXT_VAULT_PASS
      redis-cli -h nexus3-context-operational REPLICAOF nexus3-context-vault 6379

      # Repeat for each standard environment...
      echo "Replication configured"
      '
    environment:
      - TRACK_VAULT_PASS=${TRACK_VAULT_PASS}
      - CONTEXT_VAULT_PASS=${CONTEXT_VAULT_PASS}
    restart: "no"

Environment Variables (.env file)

# Vault passwords - generated on first deploy
TRACK_VAULT_PASS=yeaeoW
CONTEXT_VAULT_PASS=abc123
SESSION_VAULT_PASS=def456
# ... etc

Deployment Workflow

New Server Setup

1. Install Docker & Docker Compose

curl -fsSL https://get.docker.com | sh
sudo apt install docker-compose-plugin

1. Clone Nexus deployment repo

git clone https://github.com/corlera/nexus-deploy.git /opt/nexus

1. Generate passwords

cd /opt/nexus
./scripts/generate-passwords.sh > .env

1. Create data directories

./scripts/init-directories.sh

1. Start all services

docker-compose up -d

1. Verify replication

./scripts/verify-replication.sh

Modular Deployment

Not every customer needs all environments. Use compose profiles:

services:
  track-vault:
    profiles: ["core", "full"]
    # ...

  transcript-vault:
    profiles: ["media", "full"]
    # ...

  kb-vault:
    profiles: ["knowledge", "full"]
    # ...

Deploy with specific profile:

docker-compose --profile core up -d  # Just essentials
docker-compose --profile full up -d  # Everything

Core vs Optional Environments

Core (Always Deployed)

• User (authentication)
• Locker (credentials)
• Session (work tracking)
• Workflow (AI protocols)
• Track (projects/tasks)

Optional Modules

• Knowledge: KB, Context
• Media: Transcript, Corpus, Voice
• CRM: Contact, Links
• Storage: Documents, CDN, Archive
• Collaboration: Assignment (multi-tenant)

Backup & Restore

Data lives in /data/nexus3/{env}/. Backup strategy:

# Backup all data
tar -czf nexus-backup-$(date +%Y%m%d).tar.gz /data/nexus3/

# Restore
tar -xzf nexus-backup-20260103.tar.gz -C /
docker-compose restart

Updating Nexus

cd /opt/nexus
git pull
docker-compose pull  # Get latest FalkorDB
docker-compose up -d --force-recreate