Kubernetes部署PostgreSQL遇Pending状态：PVC未找到问题排查

问题现象 #

部署NextJS应用时，PostgreSQL Pod始终处于Pending状态，执行kubectl describe pod postgres-deployment-69894b4656-sqwvh输出如下：

Name:             postgres-deployment-69894b4656-sqwvh
Namespace:        k8s-datamaker
Priority:         0
Service Account:  default
Node:             <none>
Labels:           component=postgres-deployment
                  pod-template-hash=69894b4656
Annotations:      <none>
Status:           Pending
IP:               
IPs:              <none>
Controlled By:    ReplicaSet/postgres-deployment-69894b4656
Containers:
  postgres:
    Image:      postgres
    Port:       5432/TCP
    Host Port:  0/TCP
    Limits:
      cpu:     125m
      memory:  32Mi
    Requests:
      cpu:     63m
      memory:  16Mi
    Environment Variables from:
      postgres-config  ConfigMap  Optional: false
    Environment:
      POSTGRES_PASSWORD:  <set to the key 'DB_PASSWORD' in secret 'dbpassword'>  Optional: false
    Mounts:
      /var/lib/postgresql/data from postgredb (rw,path="postgres")
      /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-zv7hj (ro)
Conditions:
  Type           Status
  PodScheduled   False 
Volumes:
  postgredb:
    Type:       PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
    ClaimName:  postgres-pv-claim
    ReadOnly:   false
  kube-api-access-zv7hj:
    Type:                    Projected (a volume that contains injected data from multiple sources)
    TokenExpirationSeconds:  3607
    ConfigMapName:           kube-root-ca.crt
    ConfigMapOptional:       <nil>
    DownwardAPI:             true
QoS Class:                   Burstable
Node-Selectors:              <none>
Tolerations:                 node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
                             node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
  Type     Reason            Age                   From               Message
  ----     ------            ----                  ----               -------
  Warning  FailedScheduling  4m19s (x111 over 9h)  default-scheduler  0/1 nodes are available: persistentvolumeclaim "postgres-pv-claim" not found. preemption: 0/1 nodes are available: 1 Preemption is not helpful for scheduling..

用户提供的PostgreSQL Deployment文件：

apiVersion: apps/v1
kind: Deployment
metadata:
  name: postgres-deployment
  namespace: k8s-app
spec:
  replicas: 1
  selector:
    matchLabels:
      component: postgres-deployment
  template:
    metadata:
      labels:
        component: postgres-deployment
    spec:
      containers:
        - name: postgres
          image: postgres
          imagePullPolicy: "IfNotPresent"
          ports:
            - containerPort: 5432
          resources:
            requests:
              memory: "16Mi"
              cpu: "63m"
            limits:
              memory: "32Mi"
              cpu: "125m"
          envFrom:
            - configMapRef:
                name: postgres-config
          env:
            - name: POSTGRES_PASSWORD
              valueFrom:
                secretKeyRef:
                  name: dbpassword
                  key: DB_PASSWORD
          volumeMounts:
            - mountPath: /var/lib/postgresql/data
              subPath: postgres
              name: postgredb
      volumes:
        - name: postgredb
          persistentVolumeClaim:
            claimName: postgres-pv-claim

PostgreSQL PersistentVolumeClaim文件：

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: postgres-pv-claim
  labels:
    component: postgres-pv-claim
spec:
  resources:
    requests:
      storage: 1Gi
  accessModes:
    - ReadWriteMany

更新信息：执行kubectl describe pv输出：

Name:            pvc-d5e43434-972f-4385-99a6-1af11117c8b3
Labels:          <none>
Annotations:     hostPathProvisionerIdentity: a87b44e1-28cc-4f23-bc79-28c3fccd1cbb
                 pv.kubernetes.io/provisioned-by: k8s.io/minikube-hostpath
Finalizers:      [kubernetes.io/pv-protection]
StorageClass:    standard
Status:          Bound
Claim:           default/postgres-pv-claim
Reclaim Policy:  Delete
Access Modes:    RWX
VolumeMode:      Filesystem
Capacity:        1Gi
Node Affinity:   <none>
Message:         
Source:
    Type:          HostPath (bare host directory volume)
    Path:          /tmp/hostpath-provisioner/default/postgres-pv-claim
    HostPathType:  
Events:            <none>

问题根源 #

1. Namespace不匹配：

   • Pod实际运行在k8s-datamaker namespace，但需要的PVCpostgres-pv-claim被创建在default namespace（从PV绑定信息可确认）。Kubernetes要求PVC必须与Pod处于同一个namespace才能被引用。
   • Deployment文件中指定的namespace是k8s-app，但实际Pod却在k8s-datamaker，说明部署Deployment时可能使用了错误的namespace参数，导致资源实际所在namespace与定义不符。

2. PVC部署位置错误：
   创建PVC时未指定namespace，默认部署到default namespace，而Pod所在的k8s-datamaker namespace中不存在同名PVC，因此调度失败。

解决步骤 #

方法一：将PVC部署到Pod所在的k8s-datamaker namespace #

1. 修改PVC文件，添加namespace字段：

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: postgres-pv-claim
  namespace: k8s-datamaker
  labels:
    component: postgres-pv-claim
spec:
  resources:
    requests:
      storage: 1Gi
  accessModes:
    - ReadWriteMany

2. 删除原PVC：

kubectl delete pvc postgres-pv-claim -n default

3. 重新创建PVC：

kubectl apply -f pvc.yaml -n k8s-datamaker

方法二：统一所有资源到同一个namespace（推荐） #

选择一个目标namespace（如k8s-app），将所有相关资源部署到该namespace下：

1. 确保目标namespace存在：

kubectl create namespace k8s-app

2. 删除当前错误的资源：

kubectl delete deployment postgres-deployment -n k8s-datamaker
kubectl delete pvc postgres-pv-claim -n default

3. 重新部署Deployment和PVC到目标namespace：

kubectl apply -f deployment.yaml -n k8s-app
kubectl apply -f pvc.yaml -n k8s-app

4. 验证关联资源（ConfigMap、Secret）是否在同一namespace：

kubectl get configmap postgres-config -n k8s-app
kubectl get secret dbpassword -n k8s-app

若不在，需重新创建或移动到目标namespace。

验证 #

执行以下命令确认PVC状态：

kubectl get pvc -n <目标namespace>

当PVC状态变为Bound后，Pod会自动开始调度。

内容的提问来源于stack exchange，提问作者Stefan Radonjic