如何让DaemonSet Pod访问StatefulSet创建的Azure Disk PVC？

开发者阿泽

2026-8-10

问题描述

通过StatefulSet部署了3个app-1 Pod，配置如下：

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: app-1
spec:
  replicas: 3
  selector:
    matchLabels:
      app: app-1
  serviceName: "app-1"
  template:
    metadata:
      labels:
        app: app-1
    spec:
      containers:
      - name: app-1
        image: registry.k8s.io/nginx-slim:0.8
        resources:
          requests:
            cpu: 100m
            memory: 128Mi
          limits:
            cpu: 250m
            memory: 256Mi
        ports:
        - containerPort: 4567
        volumeMounts:
        - name: app-1-state-volume-claim
          mountPath: /app1Data
        - name: app-2-data-volume-claim
          mountPath: /app2Data
  volumeClaimTemplates:
  - metadata:
      name: app-1-state-volume-claim
    spec:
      accessModes: [ "ReadWriteOnce" ]
      storageClassName: "managed-csi-premium"
      resources:
        requests:
          storage: 1Gi
  - metadata:
      name: app-2-data-volume-claim
    spec:
      accessModes: [ "ReadWriteOnce" ]
      storageClassName: "managed-csi-premium"
      resources:
        requests:
          storage: 1Gi

app-1的状态存储在app-1-state-volume-claim PVC中，同时会在app-2-data-volume-claim PVC模板生成的存储中写入供app-2使用的数据。

尝试通过DaemonSet部署app-2 Pod访问该存储，DaemonSet配置如下：

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: app-2
spec:
  selector:
    matchLabels:
      name: app-2
  template:
    metadata:
      labels:
        name: app-2
    spec:
      containers:
      - name: app-2
        image: quay.io/fluentd_elasticsearch/fluentd:v2.5.2
        resources:
          limits:
            memory: 200Mi
          requests:
            cpu: 100m
            memory: 200Mi
        volumeMounts:
        - name: app2Data
          mountPath: /app2Data
      volumes:
      - name: app2Data
        persistentVolumeClaim:
          claimName: app-2-data-volume-claim

部署时出现错误：

persistentvolumeclaim "app-2-data-volume-claim" not found

由于app-1的限制，无法使用Azure文件共享，请问该如何解决此问题？

解决方案

1. 先理清PVC命名规则

StatefulSet的volumeClaimTemplates会生成带Pod序号后缀的PVC，格式为<模板名称>-<StatefulSet名称>-<Pod序号>。因此3个app-1 Pod对应的app-2-data-volume-claim系列PVC实际名称为：

app-2-data-volume-claim-app-1-0
app-2-data-volume-claim-app-1-1
app-2-data-volume-claim-app-1-2
直接引用app-2-data-volume-claim肯定找不到，这是错误的核心原因。

2. 方案一：将app-2作为sidecar嵌入app-1的StatefulSet

这是最简单的解决方式，无需额外处理PVC调度问题：把app-2容器直接添加到app-1的Pod模板中，和app-1共享同一份PVC挂载，直接访问/app2Data路径。修改后的app-1 StatefulSet配置如下：

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: app-1
spec:
  replicas: 3
  selector:
    matchLabels:
      app: app-1
  serviceName: "app-1"
  template:
    metadata:
      labels:
        app: app-1
    spec:
      containers:
      - name: app-1
        image: registry.k8s.io/nginx-slim:0.8
        resources:
          requests:
            cpu: 100m
            memory: 128Mi
          limits:
            cpu: 250m
            memory: 256Mi
        ports:
        - containerPort: 4567
        volumeMounts:
        - name: app-1-state-volume-claim
          mountPath: /app1Data
        - name: app-2-data-volume-claim
          mountPath: /app2Data
      # 添加app-2作为sidecar容器
      - name: app-2
        image: quay.io/fluentd_elasticsearch/fluentd:v2.5.2
        resources:
          limits:
            memory: 200Mi
          requests:
            cpu: 100m
            memory: 200Mi
        volumeMounts:
        - name: app-2-data-volume-claim
          mountPath: /app2Data
  volumeClaimTemplates:
  - metadata:
      name: app-1-state-volume-claim
    spec:
      accessModes: [ "ReadWriteOnce" ]
      storageClassName: "managed-csi-premium"
      resources:
        requests:
          storage: 1Gi
  - metadata:
      name: app-2-data-volume-claim
    spec:
      accessModes: [ "ReadWriteOnce" ]
      storageClassName: "managed-csi-premium"
      resources:
        requests:
          storage: 1Gi

3. 方案二：节点绑定式DaemonSet部署

如果必须保留DaemonSet独立部署的模式，可通过节点标签绑定实现每个节点的DaemonSet Pod挂载对应节点上的app-1 PVC：

给每个PVC打上对应节点的标签：

for pod in app-1-0 app-1-1 app-1-2; do
  node=$(kubectl get pod $pod -o jsonpath='{.spec.nodeName}')
  pvc=$(kubectl get pod $pod -o jsonpath='{.spec.volumes[?(@.name=="app-2-data-volume-claim")].persistentVolumeClaim.claimName}')
  kubectl label pvc $pvc node=$node
done

为每个节点创建专属DaemonSet，只调度到目标节点并挂载对应PVC（以节点node-xx为例）：

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: app-2-node-xx
spec:
  selector:
    matchLabels:
      name: app-2-node-xx
  template:
    metadata:
      labels:
        name: app-2-node-xx
    spec:
      nodeSelector:
        kubernetes.io/hostname: node-xx
      containers:
      - name: app-2
        image: quay.io/fluentd_elasticsearch/fluentd:v2.5.2
        resources:
          limits:
            memory: 200Mi
          requests:
            cpu: 100m
            memory: 200Mi
        volumeMounts:
        - name: app2Data
          mountPath: /app2Data
      volumes:
      - name: app2Data
        persistentVolumeClaim:
          claimName: app-2-data-volume-claim-app-1-0

重复此步骤为每个运行app-1的节点创建对应DaemonSet即可。

4. 方案三：将app-2改为StatefulSet与app-1一一对应

如果app-2不需要每个节点一个Pod，而是和app-1 Pod一一对应，可将app-2改为StatefulSet，直接引用对应序号的PVC：

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: app-2
spec:
  replicas: 3
  selector:
    matchLabels:
      name: app-2
  serviceName: "app-2"
  template:
    metadata:
      labels:
        name: app-2
    spec:
      # 亲和性配置，确保app-2 Pod和对应app-1 Pod调度到同一节点
      affinity:
        podAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
          - labelSelector:
              matchExpressions:
              - key: app
                operator: In
                values:
                - app-1
            topologyKey: "kubernetes.io/hostname"
      containers:
      - name: app-2
        image: quay.io/fluentd_elasticsearch/fluentd:v2.5.2
        resources:
          limits:
            memory: 200Mi
          requests:
            cpu: 100m
            memory: 200Mi
        volumeMounts:
        - name: app2Data
          mountPath: /app2Data
      volumes:
      - name: app2Data
        persistentVolumeClaim:
          claimName: app-2-data-volume-claim-app-1-0 # 对应app-1-0的PVC

注意：此方案需要手动为每个app-2 Pod指定对应序号的PVC，或通过自定义控制器实现自动绑定。

内容的提问来源于stack exchange，提问作者Atul Takekar

如何让DaemonSet Pod访问StatefulSet创建的Azure Disk PVC？

1. 先理清PVC命名规则 #

2. 方案一：将app-2作为sidecar嵌入app-1的StatefulSet #

3. 方案二：节点绑定式DaemonSet部署 #

4. 方案三：将app-2改为StatefulSet与app-1一一对应 #

1. 先理清PVC命名规则

2. 方案一：将app-2作为sidecar嵌入app-1的StatefulSet

3. 方案二：节点绑定式DaemonSet部署

4. 方案三：将app-2改为StatefulSet与app-1一一对应