Multi-Node Kubernetes Basics for Infrastructure Engineers

In this lab, you will expand a pre-initialized single-node Kubernetes cluster into a full three-node setup (one control-plane, two workers) using kubeadm. Modules guide you through node labeling and scheduling, installing MetalLB for LoadBalancer services on bare-metal, deploying Rook-Ceph for dynamic persistent storage, and running real-world applications: a persistent web app with MongoDB and a full TIG (Telegraf/InfluxDB/Grafana) monitoring stack—all with external access and data durability demonstrations.

Upon completion of this lab, you will be able to:

• Bootstrap a multi-node Kubernetes cluster (1 control-plane, 2 workers) using kubeadm and containerd.
• Join additional nodes with kubeadm tokens and verify cluster health.
• Use node labels and nodeSelector to control pod placement.
• Install and configure MetalLB and expose apps via Service type: LoadBalancer.
• Install Rook-Ceph, create a cluster, and provision storage with StorageClasses.
• Build self-contained modules that create their own Namespace and StorageClass.
• Deploy two applications:
  • Asteroid Mining (web + MongoDB) pinned to a labeled node and exposed via MetalLB.
  • TIG stack (Telegraf, InfluxDB, Grafana) with Ceph-backed PVCs and a Grafana LoadBalancer.
• Access services from a Windows jumpbox and perform safe cleanup.

Software: Kubernetes (kubeadm, kubelet, kubectl), containerd, Calico CNI, MetalLB, Rook-Ceph, Red Hat Enterprise Linux 9 (RHEL), PuTTY, NGINX, MongoDB, Telegraf, InfluxDB, Grafana. Virtual Hardware: Three RHEL VMs (node1 control-plane, node2/node3 workers) and a Windows Jumpbox for browser access and SSH, interconnected on an internal lab network.