Lesson 4: Deployment of Kubernetes

General overview

• What is Kubernetes and what is the main purpose behind it (ease of deployment, management and scaling)

• Networking (the basics of it, and why it’s hard to do networking of Docker containers without Kubernetes)

Deploying Kubernetes

• Quick overview of the infrastructure (API, nodes, network provider, ie. flannel)

• Using Kubeadm to initialize and join nodes to a cluster

• Explain why it’s a bad idea to have the Kubernetes master participate as a node in the cluster (for security reasons)

Using Kubernetes

• Basic overview of the different workload types (Daemonsets, Replication Controllers, etc.)

• Basic overview of how services work (and how they integrate with the rest of Kubernetes)

• The structure of the Kubernetes API YAML and how to create resources with it

• Deploy Zeppelin, Spark Master, Spark Slave workloads

• Verify that the work that we’ve done in previous meetups is working correctly on the cluster

• Scale up Spark Slaves to see performance improvements

• Show that we can deploy an arbitrary application and integrate it with our existing framework very easily (such as a distributed database)