Chapter 9

Here are the sample answers to the questions of this chapter:

1. Here are some reasons why we need an orchestration engine:
   • Containers are ephemeral and only an automated system (the orchestrator) can handle this efficiently.
   • For high availability reasons, we want to run multiple instances of each container. The number of containers to manage quickly becomes huge.
   • To meet the demand of today’s internet, we need to quickly scale up and down.
   • Containers, contrary to VMs, are not treated as pets and fixed or healed when they misbehave, but are treated as cattle. If one misbehaves, we kill it and replace it with a new instance. The orchestrator quickly terminates an unhealthy container and schedules a new instance.
2. Here are some responsibilities of a container orchestration engine:
   • Manages a set of nodes in a cluster
   • Schedules workload to the nodes with sufficient free resources
   • Monitors the health of nodes and workload
   • Reconciles current state with desired state of applications and components
   • Provides service discovery and routing
   • Load balances requests
   • Secures confident data by providing support for secrets

3. Here is an (incomplete) list of orchestrators, sorted by their popularity:
   • Kubernetes by Google, donated to the CNCF
   • SwarmKit by Docker—that is, Operations Support System (OSS)
   • AWS ECS by Amazon
   • Azure AKS by Microsoft
   • Mesos by Apache—that is, OSS
   • Cattle by Rancher
   • Nomad by HashiCorp