MongoDB Replica Set Deployment in Kubernetes

This guide explains the deployment of a MongoDB Replica Set in Kubernetes using a StatefulSet, headless service, and Persistent Volume Claims (PVCs).

Table of Contents

• Overview
• Components
• YAML Specification
• How It Works
• Key Benefits
• Accessing MongoDB

---

Overview

This setup deploys a 3-node MongoDB Replica Set in Kubernetes using the following components:

• StatefulSet for managing MongoDB pods.
• Headless Service for stable DNS names for inter-pod communication.
• Persistent Volumes for data durability.

Replica Set Members

• Primary Node: Handles all reads and writes.
• Secondary Nodes: Replicate data from the primary.
• DNS Names:
  • mongodb-0.mongodb-headless-service
  • mongodb-1.mongodb-headless-service
  • mongodb-2.mongodb-headless-service

---

Components

1. ServiceAccount

A ServiceAccount for secure inter-pod communication.

2. Headless Service

Used for stable DNS records for MongoDB pods.

3. StatefulSet

Manages the MongoDB pods, ensuring:

• Stable pod names.
• Persistent storage for each pod.
• Ordered pod creation and deletion.

4. Persistent Volumes

Ensures that each pod has dedicated storage for MongoDB data, which persists across pod restarts.

---

YAML Specification

ServiceAccount

apiVersion: v1
kind: ServiceAccount
metadata:
  name: mongodb-sa

Accessing MongoDB

1. Get the LoadBalancer IP:

kubectl get service mongodb-loadbalancer

2. Connect to MongoDB using localhost:

mongosh "mongodb://localhost:27017"

Note: The LoadBalancer service is exposing MongoDB on localhost:27017. Make sure no other service is using this port locally.

For connecting from applications, use the connection string:

mongodb://localhost:27017/?directConnection=true