Docker Swarm Cheat Sheet.md

Docker Swarm Cheat Sheet

Starting/Stopping a Docker Swarm

# Turns your docker server into a manager of the Docker Swarm
$ docker swarm init

# Turns off docker swarm
$ docker swarm leave -f

Deploying application

connect cli to manager run commands on control plane.

# Creates a docker stack. Deploys compose file to your cluster
$ docker stack deploy -c docker-compose.yml csv-merger

# Removes a created docker stack
$ docker stack rm csv-merger

Networking

# Create network
$ docker network create --driver overlay portfolio-network

# View all networks
$ docker network ls

Storing Configurations

Creating Configuration

# Create a config file within the cluster database
$ docker config create <config-name> <path-to-config-file> 

$ docker config inspect --pretty <config-name>

$ docker config ls

Creating Secrets (Configuration)

# Create a config file within the cluster database
$ docker secret create <secret-name> <path-to-secret-file>

$ docker secret inspect --pretty <secret-name>

$ docker secret ls

Using Secrets & Configurations

version: "3.7"

services:
  products-db:
    image: ...
    ports:
      - ...
    environment:
      - ...
    config:
      - source: <config-name>
        target: <path-to-config-file>
    secrets:
      - source: <secret-name>
        target: <path-to-secret-file>
        mode: 0400

configs:
    <config-name>:
        external: true
    
secrets:    
    <secret-name>:
        external: true

Helpful Commands

# List nodes in the swarm
$ docker node ls

# Add worker to cluster
$ docker swarm join-token worker

# Add manager to control plane
$ docker swarm join-token manager

# list services
$ docker service ls

# list stacks
$ docker stack ls

# list networks
$ docker network ls

# list containers
$ docker ps

Terminology

• Node - a single physical or virtual machine that is part of the swarm cluster. A node can either me a manager or a worker node.

• Manger Node - These nodes are responsible for managing the state of the swarm, scheduling tasks, and maintaining the desired state of the cluster. There is usually one or more manager nodes in a swarm for high availability and fault tolerance. Manager nodes are also responsible for coordinating the worker nodes.

• Worker Node - These nodes are responsible for executing the tasks that are assigned to them by the manager nodes. Worker nodes do not participate in the decision-making process; instead, they run the services and tasks as instructed by the manager nodes

• Docker Swarm Cluster - a group of interconnected nodes (physical or virtual machines) that work together to manage and run containerized applications

• Control Plane - primary consisting of manager nodes, this is responsible for managing the overal state and behavior of the cluser. Scheduling, monitoring and orchestrating task and services across the cluster.

Abstractions

• Services - a higher-level abstraction that represents a group of containers running the same application or performing the same function. First class objects

• Stacks - a higher-level abstraction that allows you to define and manage a group of interrelated services, networks, and volumes as a single unit. A stack is essentially a blueprint for deploying a multi-service application in a Docker Swarm cluster.

When in swarm connect CLI to manager and run commands on the control plane

References

• Deploying Containerized Applications by Elton Stoneman