Differences

This shows you the differences between two versions of the page.

--- k8s:core:cluster_model [2026/05/31 02:30] – created phong2018
+++ k8s:core:cluster_model [2026/05/31 02:48] (current) – phong2018
@@ Line 1: / Line 1: @@
-====== Kubernetes Cluster Model ======
+====== Cluster Model ======
 A Kubernetes cluster is a group of machines that work together to run applications.
-It has two main parts:
+It is made of:
+  * Control Plane (brain)
+  * Worker Nodes (workers)
+  * Pods (applications)
+---
 ===== 1. Control Plane =====
-This is the "brain" of Kubernetes.
-It manages:
+The Control Plane is the **brain of Kubernetes**.
-  * Scheduling
-  * Scaling
-  * Health checks
-  * Desired state
-Key components:
+It decides:
-  * API Server → entry point for all commands
+  * Where applications run
-  * Scheduler → decides where pods run
+  * How many copies run
-  * Controller Manager → ensures desired state
+  * What happens when something fails
-  * etcd → key-value database storing cluster state
+Example:
+  * You say: "Run 3 replicas of my app"
+  * Control Plane ensures it happens
+---
 ===== 2. Worker Nodes =====
-These are the machines that run your applications.
-Each node contains:
+Worker Nodes are machines that **run your applications**.
-  * kubelet → agent that communicates with control plane
-  * container runtime → runs containers (Docker/containerd)
-  * kube-proxy → handles networking
-Nodes run:
+Example:
-  * Pods (your application containers)
+Node A → runs Pod 1, Pod 2
+Node B → runs Pod 3
+If a node fails:
+  * Kubernetes moves Pods to another node
+---
 ===== 3. Pods =====
-The smallest unit in Kubernetes.
-A pod:
+A Pod is the **smallest unit in Kubernetes**.
-  * Contains one or more containers
-  * Shares network and storage
-  * Represents a single running instance
-===== Cluster flow =====
+Example:
+A Pod contains:
+  * Your application container (e.g. nginx)
+  * Optional helper containers
 Example:
-User → Service → Pod → Node
-Control Plane decides:
+Pod = Web App Container
-  * Where the pod should run
-  * How many replicas exist
-  * When to restart failed pods
+---
+===== Full flow example =====
+User request flow:
+User
+↓
+Service
+↓
+Pod
+↓
+Node
+Control Plane ensures:
+  * Pods are created
+  * Pods are healthy
+  * Pods are distributed across nodes
+---
+===== Simple analogy =====
-===== Key idea =====
+Think of a factory:
-Kubernetes cluster =
-  Control Plane (brain)
+  * Control Plane → factory manager
-  +
+  * Worker Nodes → machines
-  Worker Nodes (workers)
+  * Pods → products being produced
-  +
-  Pods (applications)