Container Orchestration

  What is Docker Swarm?

• Native cluster management and orchestration for Docker
• A pool of Docker hosts to a single, virtual Docker host
• Docker Swarm serves the standard Docker API
• Standalone Swarm before Docker version 1.12
• In Docker 1.12 and higher, Swarm Mode integrated with Docker Engine

Key Features

• Integrated cluster management with Docker engine
• Declarative service model with YAML based files
• Scaling with declaring the number of tasks
• Make sure apps are up and running with Desired state reconciliation
• Managed Multi-host networking
• Service discovery with unique DNS names

 

• Load balancing

• Secure by default with TLS mutual authentication and encryption for communications

 

• Rolling updates for delaying deployments and rollback

Docker Swarm Concepts

• Nodes: An instance of the Docker engine participating in the swarm
• Manager: Performs the orchestration and cluster management. Dispatches unit of works to worker nodes.

• Worker: Receives and executes tasks dispatched from manager nodes

• Task: Atomic scheduling unit of swarm. A running container which is part of a swarm service and managed by a swarm manager, as opposed to a standalone container

• Service: The definition of the tasks to execute on swarm cluster. Replicated Services and Global Services

• Stack: The definition of the services to execute on swarm cluster

• Load Balancing: Swarm manager uses ingress load balancing to expose the services via PublishedPort

Docker Compose Vs. Docker Stack

• docker-compose -f docker-compose.yml up -d

• docker stack deploy -c docker-compose.yml stack_name

• Docker stack is ignoring “build” instructions.
• Docker compose is a Python project. Origin is “fig”. You should install docker-compose separately to use it.
• Docker Stack is included with the Docker Engine from version 1.12
• To use Docker Stack, Swarm mode should be enabled
• Docker Stack not supporting docker-compose files which are written according to Version 2 specification. Docker Compose supports Version 2 and 3

How Nodes Work   Manager nodes handle cluster management tasks:

• maintaining cluster state
• scheduling services
• serving swarm mode HTTP API endpoints

For fault-tolerance features, use an odd number of manager nodes An N manager cluster tolerates the loss of at most (N-1)/2 managers Docker recommends maximum 7 manager node Worker nodes not handle cluster management tasks. By default, all Manager nodes are also workers. How Services Work

• To deploy an application image in swarm mode, we must create a service.
• When we create a service, we specify some options;
  • Image name
  • Commands
  • Port
  • Networking
  • Resources
  • Rolling update policy
  • Replica count