Service Management¶

Deploy¶

task deploy                        # Every stack
task deploy -- homepage            # One stack
task deploy -- cicd forgejo        # A subset
task deploy:plan -- homepage       # What a deploy would change, without doing it

A named stack always deploys. Anything it needs, and anything swept in by a bare task deploy, is ensured: deployed only if it has not already converged. That is also the resume — a run killed partway is finished by running the same command again.

task ansible:cluster:up            # bring the cluster up (swarm, labels, storage)
task deploy:full                   # bootstrap + cluster:up + deploy

Status¶

docker stack ls                                    # List deployed stacks
docker stack services homepage                     # Services in a stack
docker service logs homepage_homepage --tail 50 -f # Logs
docker service ps homepage_homepage                # Task placement/errors

Remove¶

task ansible:teardown:service -- -e "stack_name=homepage"                       # Keep volumes
task ansible:teardown:service -- -e "stack_name=homepage remove_volumes=true"   # Delete data

Add a New Service¶

mkdir stacks/apps/myservice

# stacks/apps/myservice/docker-compose.yml
services:
  myservice:
    image: myapp:latest
    volumes:
      - myservice_data:/data
    networks:
      - traefik-public
    deploy:
      labels:
        - "traefik.enable=true"
        - "traefik.http.routers.myservice.rule=Host(`myapp.${BASE_DOMAIN}`)"
        - "traefik.http.routers.myservice.entrypoints=websecure"
        - "traefik.http.routers.myservice.tls.certresolver=dns"
        - "traefik.http.services.myservice.loadbalancer.server.port=8080"
        - "traefik.swarm.network=traefik-public"

networks:
  traefik-public:
    external: true

volumes:
  myservice_data:
    driver: local

task deploy -- myservice

Storage Provisioning¶

iSCSI (databases, configs)¶

sudo mkdir -p /mnt/iscsi/app-data/myservice
sudo chown -R 1000:1000 /mnt/iscsi/app-data/myservice

# PostgreSQL needs UID 999
sudo chown -R 999:999 /mnt/iscsi/app-data/myservice/postgresql

CIFS (media, bulk files)¶

Create the share on your NAS. The deployment creates the Docker volume automatically using .env credentials.

volumes:
  myservice_media:
    driver: local
    driver_opts:
      type: "cifs"
      o: "username=${SMB_USERNAME},password=${SMB_PASSWORD},vers=3.0"
      device: "//${NAS_SERVER}/myservice"

Cluster Operations¶

task ansible:cluster:sync           # Re-join nodes with changed IPs
task ansible:cluster:update-labels  # Apply inventory label changes to live swarm
task ansible:volume:ls              # List all Docker volumes
task ansible:volume:cleanup -- -e "stack_name=sonarr"  # Remove stale CIFS volumes