• Need to ask the following questions:
    • Learning
    • Testing
    • Production-grade application?
    • Cloud or On-prem
    • What workloads to be ran?
      • What applications?
        • Web
        • Big data / Analytics?
      • Application resource requirements?
        • Heavy traffic or burst?
  • Education purposes:
    • Minikube
    • Single node cluster with kubeadm/GCP/AWS
  • Development and Testing:
    • Multi-node cluster.
    • Use kubeadm or quickly provisioning with (Google Container Engine) GCP, AWS or AKS.
  • Hosting Production-grade Applications
    • High availability multi node cluster with ultiple master nodes.
    • kubeadm or GCP or Kops on AWS
    • Cluster restrictions:
      • Up to 5000 nodes.
      • Up to 150,000 PODs in the cluster.
      • Up to 300,000 total containers
      • Up to 100 Pods per node

  • The size of the node really depends. A good chart is the following: 1-5 nodes –> GCP - N1-standard-1 1 vCPU 3.75GB RAM / AWS - M3.medium - 1 vCPU 3.75GB 101-250 –> GCP - N1-standard-8 8 vCPU 30GB RAM / AWS - M3.2xlarge - 8 vCPU 30GB f> 500 –> GCP - N1-standard-32 32 vCPU 120GB RAM / AWS - C4.8xlarge - 36 vCPU 60GB

  • Deploying OnPrem clusters, use the above guides.
  • Use kubeadm for on-prem
  • GKE for GCP
    • Has one-click cluster upgrade feature for simple upgrades.
  • Kops for AWS

  • Azure Kubernetes Service for Azure

  • Storage concerns:
    • High performance - SSD backed storage.
    • Multiple concurrent connections - network-based storage
    • Persistent shared volumes for shared access across multiple PODs.
    • Label nodes with specific disk types.
    • Use Node Selectors to assign applications to nodes with specific disk types.
  • Nodes:
    • Physical or virtual is fine.
    • Minimum of 4 Node Cluster (Sise-based on workload)
    • Master vs. Worker Nodes
      • Not a strict requirements that Worker Nodes should use
      • Master Nodes should be ran with Control Plane items only.
      • Deployment tools like kubeadm, prevent deployments being placed on Master Nodes.
        • It adds a taint to a Master Node to stop this.
    • Linux x86_64 Architiectures.
  • Master Nodes:
    • On large clusters, can separate the ETCD cluster away from the Master Node.
      • It will be placed on its own cluster node.
  • Won’t be tested on anything in this section in the exam.

Updated: