Kubernetes 网络：无法从在不同节点上运行的另一个 pod 访问 pod

Question

我正在尝试在自己的裸机 Kubernetes 基础架构上安装 nextcloud（我有 3 个工作节点），但当 nextcloud 容器尝试连接到 mariadb 容器时遇到了问题。由于我使用 nodePort 服务和 hostNetwork，我可以使用工作节点上的所有 IP 地址访问 mariadb。问题出在两个容器之间的通信上。

我的 nextcloud 配置：

apiVersion: v1
kind: Service
metadata:
  name: nextcloud
  namespace: nextcloud
spec:
  type: NodePort
  ports:
    - name: http
      port: 80
      targetPort: 80
      nodePort: 30001
      protocol: TCP
  selector:
    app: nextcloud
    tier: frontend
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: nextcloud-pv-claim
  namespace: nextcloud
  labels:
    app: nextcloud
spec:
  storageClassName: rook-ceph-block
  accessModes:
  - ReadWriteOnce
  resources:
    requests:
      storage: 20Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: nextcloud
  name: nextcloud
  namespace: nextcloud
spec:
  replicas: 1
  strategy:
    type: Recreate
  selector:
    matchLabels:
      app: nextcloud
      tier: frontend
  template:
    metadata:
      labels:
        app: nextcloud
        tier: frontend
    spec:
      hostNetwork: true
      containers:
      - env:
        - name: NEXTCLOUD_ADMIN_USER
          value: admin
        - name: NEXTCLOUD_ADMIN_PASSWORD
          value: ********
        - name: MYSQL_DATABASE
          value: nextcloud
        - name: MYSQL_HOST
          value: database-nextcloud
        - name: MYSQL_PASSWORD
          value: **********
        - name: MYSQL_USER
          value: nextcloud
        name: nextcloud
        image: nextcloud
        ports:
        - containerPort: 80
          protocol: TCP
        terminationMessagePath: /dev/termination-log
        terminationMessagePolicy: File
        volumeMounts:
        - mountPath: /var/www/html
          name: html
      restartPolicy: Always
      volumes:
        - name: html
          persistentVolumeClaim:
              claimName:  nextcloud-pv-claim

我的 mariadb 配置：

apiVersion: v1
kind: Service
metadata:
  name: database-nextcloud
  namespace: nextcloud
spec:
  type: NodePort
  ports:
    - name: mariadb
      port: 3306
      targetPort: 3306
      nodePort: 30002
      protocol: TCP
  selector:
    app: database-nextcloud
    tier: backend
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: database-nextcloud-pv-claim
  namespace: nextcloud
  labels:
    app: database-nextcloud
spec:
  storageClassName: rook-ceph-block
  accessModes:
  - ReadWriteOnce
  resources:
    requests:
      storage: 8Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: database-nextcloud
    tier: backend
  name: database-nextcloud
  namespace: nextcloud
spec:
  replicas: 1
  strategy:
    type: Recreate
  selector:
    matchLabels:
      app:  database-nextcloud
      tier: backend
  template:
    metadata:
      labels:
        app: database-nextcloud
        tier: backend
    spec:
      containers:
      - env:
        - name: MYSQL_DATABASE
          value: nextcloud
        - name: MYSQL_PASSWORD
          value: **********
        - name: MYSQL_USER
          value: nextcloud
        - name: MYSQL_ROOT_PASSWORD
          value: ***********
        image: mariadb
        name: database-nextcloud
        volumeMounts:
        - mountPath: /var/lib/mysql
          name: database-nextcloud
        ports:
        - containerPort: 3306
          protocol: TCP
      restartPolicy: Always
      volumes:
      - name: database-nextcloud
        persistentVolumeClaim:
            claimName: database-nextcloud-pv-claim

豆荚：

kubectl get pods -o wide -n nextcloud
NAME                                  READY   STATUS    RESTARTS   AGE     IP              NODE   NOMINATED NODE   READINESS GATES
database-nextcloud-7f488f5894-5bmqz   1/1     Running   0          6m22s   192.168.0.101   pwe1   <none>           <none>
nextcloud-67bf849f44-m42nc            1/1     Running   1          13d     192.168.0.101   pwe1   <none>           <none>

服务：

kubectl get service -n nextcloud
NAME                 TYPE       CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
database-nextcloud   NodePort   10.106.57.87     <none>        3306:30002/TCP   13d
nextcloud            NodePort   10.106.140.209   <none>        80:30001/TCP     13d

nextcloud 容器内的 /etc/resolv.conf：

search nextcloud.svc.cluster.local svc.cluster.local cluster.local v.cablecom.net
nameserver 10.96.0.10
options ndots:5

我删除并重新创建了部署（这次没有hostNetwork在容器配置中设置，现在出现此错误：

Error while trying to create admin user: Failed to connect to the database: An exception occurred in driver: SQLSTATE[HY000] [2002] php_network_getaddresses: getaddrinfo failed: Temporary failure in name resolution