StatefulSets

Introduction to StatefulSets

While Deployments are ideal for stateless applications, StatefulSets are designed for applications that require stable, unique network identifiers, stable persistent storage, and ordered deployment and scaling. They are ideal for stateful applications like databases, distributed caches, and messaging systems.

Key Features of StatefulSets

• Stable, Unique Network Identifiers: Each pod in a StatefulSet gets a predictable hostname based on the StatefulSet name and an ordinal index (0, 1, 2, etc.).
• Ordered Deployment and Scaling: Pods are created and terminated in a defined order, starting from index 0.
• Stable Storage: Each pod can be associated with its own persistent volume that follows it even if it’s rescheduled to another node.
• Stable Network Identity: Services can route traffic to specific pods based on their stable hostnames.

When to Use StatefulSets

• Database clusters (MySQL, PostgreSQL, MongoDB)
• Messaging systems with persistent queues (RabbitMQ, Kafka)
• Distributed caches (Redis)
• Any application requiring stable identity and ordered operations

Example StatefulSet

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: postgres
spec:
  serviceName: "postgres"
  replicas: 3
  selector:
    matchLabels:
      app: postgres
  template:
    metadata:
      labels:
        app: postgres
    spec:
      containers:
      - name: postgres
        image: postgres:13
        ports:
        - containerPort: 5432
          name: postgres
        env:
        - name: POSTGRES_PASSWORD
          valueFrom:
            secretKeyRef:
              name: postgres-secret
              key: password
        volumeMounts:
        - name: postgres-data
          mountPath: /var/lib/postgresql/data
  volumeClaimTemplates:
  - metadata:
      name: postgres-data
    spec:
      accessModes: ["ReadWriteOnce"]
      storageClassName: "managed-premium"
      resources:
        requests:
          storage: 10Gi

In this example:

• The StatefulSet creates 3 PostgreSQL pods with predictable names: postgres-0, postgres-1, and postgres-2
• Each pod gets its own 10GB persistent volume for data storage
• The pods are created and terminated in order
• We’re using Azure’s managed-premium storage class for better performance

StatefulSets vs Deployments

AKS Considerations for StatefulSets

When using StatefulSets in AKS:

• Use Azure managed disks or Azure Files for persistent storage
• Consider zone redundancy for high availability
• For databases, evaluate managed services (Azure Database) vs. self-managed StatefulSets
• Be aware of the performance characteristics of different storage classes (see later section on storage)

StatefulSet Best Practices

• Consider using a headless service (clusterIP: None) to enable stable DNS names for pods
• Create appropriate PodDisruptionBudgets to ensure high availability during cluster operations
• Implement proper backup and restore procedures for stateful data
• Consider using Helm charts for complex StatefulSet deployments
• Carefully test scaling operations, as they can impact distributed stateful applications

Further Reading and Resources

• Kubernetes StatefulSets Documentation