Deployment 滚动更新深度解析

Deployment 核心功能

Deployment 是最常用的工作负载资源，提供：

• 声明式更新：描述期望状态，自动滚动更新
• 版本回滚：保留历史 ReplicaSet，支持快速回滚
• 弹性伸缩：随时调整副本数
• 暂停/恢复：批量修改后一次性发布

完整配置示例

apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app
  namespace: production
  labels:
    app: my-app
  annotations:
    kubernetes.io/change-cause: "升级到 v2.0，新增用户认证功能"
spec:
  replicas: 3

  selector:
    matchLabels:
      app: my-app

  # 滚动更新策略
  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxSurge: 1        # 最多额外创建 1 个 Pod（可以是百分比：25%）
      maxUnavailable: 0  # 最多 0 个不可用（保证服务不中断）

  # 历史版本保留数量
  revisionHistoryLimit: 10

  # Pod 就绪后等待时间（防止假就绪）
  minReadySeconds: 10

  # 更新超时时间
  progressDeadlineSeconds: 600

  template:
    metadata:
      labels:
        app: my-app
        version: v2.0
    spec:
      containers:
      - name: app
        image: my-app:v2.0
        ports:
        - containerPort: 8080
        resources:
          requests:
            cpu: "100m"
            memory: "128Mi"
          limits:
            cpu: "500m"
            memory: "512Mi"
        readinessProbe:
          httpGet:
            path: /ready
            port: 8080
          initialDelaySeconds: 5
          periodSeconds: 5
        lifecycle:
          preStop:
            exec:
              command: ["/bin/sh", "-c", "sleep 10"]  # 优雅退出
      terminationGracePeriodSeconds: 30

滚动更新流程

初始状态：RS-v1 (3 pods)

更新镜像后：
Step 1: RS-v2 创建 1 个 Pod（maxSurge=1，总数=4）
Step 2: RS-v2 Pod Ready 后，RS-v1 缩减 1 个（总数=3）
Step 3: 重复直到 RS-v2 = 3，RS-v1 = 0

最终：RS-v2 (3 pods)，RS-v1 (0 pods，保留用于回滚)

常用操作

# 查看 Deployment 状态
kubectl get deployment my-app
kubectl rollout status deployment/my-app

# 更新镜像（触发滚动更新）
kubectl set image deployment/my-app app=my-app:v2.1

# 查看更新历史
kubectl rollout history deployment/my-app
kubectl rollout history deployment/my-app --revision=2

# 回滚到上一版本
kubectl rollout undo deployment/my-app

# 回滚到指定版本
kubectl rollout undo deployment/my-app --to-revision=2

# 暂停更新（批量修改时使用）
kubectl rollout pause deployment/my-app
kubectl set image deployment/my-app app=my-app:v2.2
kubectl set resources deployment/my-app -c app --limits=cpu=1
kubectl rollout resume deployment/my-app  # 一次性发布所有修改

# 扩缩容
kubectl scale deployment/my-app --replicas=5

# 强制重启所有 Pod（不改变镜像）
kubectl rollout restart deployment/my-app

蓝绿部署

# 蓝色版本（当前生产）
apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app-blue
spec:
  replicas: 3
  template:
    metadata:
      labels:
        app: my-app
        slot: blue
    spec:
      containers:
      - name: app
        image: my-app:v1.0
---
# 绿色版本（新版本）
apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app-green
spec:
  replicas: 3
  template:
    metadata:
      labels:
        app: my-app
        slot: green
    spec:
      containers:
      - name: app
        image: my-app:v2.0
---
# Service 切换流量（修改 selector）
apiVersion: v1
kind: Service
metadata:
  name: my-app
spec:
  selector:
    app: my-app
    slot: green  # 从 blue 改为 green 即完成切换

金丝雀发布

# 稳定版本（90% 流量）
apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app-stable
spec:
  replicas: 9
  template:
    metadata:
      labels:
        app: my-app
        track: stable
---
# 金丝雀版本（10% 流量）
apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app-canary
spec:
  replicas: 1
  template:
    metadata:
      labels:
        app: my-app
        track: canary
---
# Service 同时选中两个 Deployment
apiVersion: v1
kind: Service
spec:
  selector:
    app: my-app  # 不区分 track，按副本数比例分流

HPA 自动伸缩

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: my-app-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: my-app
  minReplicas: 2
  maxReplicas: 20
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70
  - type: Resource
    resource:
      name: memory
      target:
        type: AverageValue
        averageValue: 400Mi
  behavior:
    scaleUp:
      stabilizationWindowSeconds: 60
      policies:
      - type: Pods
        value: 4
        periodSeconds: 60
    scaleDown:
      stabilizationWindowSeconds: 300