What is the purpose of Kubernetes Deployment? How does it implement rolling updates and rollbacks?

Kubernetes Deployment is a declarative update controller for managing Pods and ReplicaSets. It provides declarative application deployment and update capabilities.

Core Functions of Deployment

1. Declarative Management: Define the desired state through YAML files, and Kubernetes automatically implements the transition from the current state to the desired state.

2. Rolling Updates: Supports zero-downtime application updates by gradually replacing old version Pods.

3. Rollback Capability: Can easily roll back to previous versions, supports viewing update history.

4. Auto-scaling: Supports manual or automatic adjustment of the number of Pod replicas.

5. Self-healing: Automatically creates new Pods to maintain the desired replica count when Pods fail or are deleted.

How Deployment Works

Deployment manages Pods through ReplicaSet:

1. ReplicaSet: Ensures that a specified number of Pod replicas are always running.

2. Pod Template: Defines the Pod specification, including container images, resource limits, etc.

3. Update Strategy: Controls the behavior of rolling updates, such as the maximum number of unavailable Pods and the maximum number of surge Pods.

Deployment YAML Example

yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx:1.14.2
        ports:
        - containerPort: 80

Update Strategy

Deployment supports two update strategies:

1. RollingUpdate (default):

   • Gradually replaces Pods, ensuring that there are always available Pods
   • Can configure maxUnavailable and maxSurge
   • Suitable for most applications

2. Recreate:

   • First deletes all old Pods, then creates new Pods
   • Causes a brief service interruption
   • Suitable for applications that cannot run new and old versions simultaneously

Rolling Update Parameters

• maxUnavailable: The maximum number of Pods that can be unavailable during the update process (default 25%)

• maxSurge: The maximum number of additional Pods that can be created during the update process (default 25%)

Rollback Operations

View update history:

bash
kubectl rollout history deployment/nginx-deployment

Rollback to previous version:

bash
kubectl rollout undo deployment/nginx-deployment

Rollback to specific version:

bash
kubectl rollout undo deployment/nginx-deployment --to-revision=2

Scaling

Manual scaling:

bash
kubectl scale deployment/nginx-deployment --replicas=5

Auto-scaling (HPA):

yaml
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: nginx-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: nginx-deployment
  minReplicas: 2
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 50

Deployment vs Other Controllers

1. vs ReplicaSet:

   • Deployment is a higher-level controller of ReplicaSet
   • Deployment provides update and rollback capabilities, while ReplicaSet can only maintain replica counts

2. vs StatefulSet:

   • Deployment is suitable for stateless applications
   • StatefulSet is suitable for stateful applications, providing stable network identities and persistent storage

3. vs DaemonSet:

   • Deployment can run a specified number of Pods on any node
   • DaemonSet ensures that one Pod replica runs on each node

Best Practices

1. Use declarative configuration: Always use YAML files to define Deployment, not imperative commands.

2. Set reasonable resource limits: Set CPU and memory requests and limits for containers.

3. Configure health checks: Use livenessProbe and readinessProbe to ensure application health.

4. Use multi-stage builds: Optimize image size to improve deployment speed.

5. Set appropriate update strategies: Choose RollingUpdate or Recreate based on application characteristics.

6. Use labels and annotations: Add meaningful labels and annotations to Deployment for easier management and tracking.

7. Monitor the update process: Use kubectl rollout status to monitor update progress.