Kubernetes部署报错：无法找到指定Python文件，需PV加载可更新文件

创建Kubernetes Deployment时遇到报错：

python: can't open file '/home/validator_scheduler/validator_rq_scheduler.py': [Errno 2] No such file or directory

需求是让Python文件从Persistent Volume加载，且该卷支持版本控制或手动更新。以下是相关配置文件：

Dockerfile #

FROM python

WORKDIR /home/validator_scheduler

COPY requirements.txt .

RUN apt-get update && apt-get install -y dnsutils

RUN pip install --no-cache-dir -r requirements.txt

COPY . /home/validator_scheduler

CMD ["python" , "validator_rq_scheduler.py"]

文件夹结构 #

monitoring-system
  validator_scheduler
    validator_rq_scheduler.py

Deployment清单 #

apiVersion: apps/v1
kind: Deployment
metadata:
  name: validator-rq-scheduler-deployment
  labels:
    app: validator-rq-scheduler
spec:
  replicas: 1
  selector:
    matchLabels:
      app: validator-rq-scheduler
  template:
    metadata:
      labels:
        app: validator-rq-scheduler
    spec:
      containers:
        - name: validator-rq-scheduler
          image: BLABLACOMPANY/monitoring-system:validator-rq-scheduler
          workingDir: /home/validator_scheduler
          volumeMounts:
            - name: validator-scheduler
              mountPath: /home/validator_scheduler
          env:
            - name: REDIS_URL
              valueFrom:
                secretKeyRef:
                  name: app-secrets
                  key: REDIS_URL
      volumes:
        - name: validator-scheduler
          persistentVolumeClaim:
            claimName: validator-scheduler-pvc
      imagePullSecrets:
        - name: docker-registry-secret

PVC清单 #

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: validator-scheduler-pvc
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 5Gi

核心问题 #

你把Persistent Volume直接挂载到了容器的/home/validator_scheduler目录，这会完全覆盖镜像中原有的该目录内容。而你的PV里并没有预先放入validator_rq_scheduler.py文件，容器启动自然找不到该文件。另外，Dockerfile把代码打包进镜像的逻辑，和你想从PV加载代码的需求冲突。

分步解决 #

1. 调整PV挂载路径 #

不要挂载整个工作目录，改成挂载到子目录，避免覆盖镜像里的依赖和配置。修改Deployment的volumeMounts：

volumeMounts:
  - name: validator-scheduler
    mountPath: /home/validator_scheduler/app

2. 修改Dockerfile适配PV加载逻辑 #

去掉复制本地代码到镜像的步骤，让CMD从PV挂载的目录启动脚本：

FROM python

WORKDIR /home/validator_scheduler

COPY requirements.txt .

RUN apt-get update && apt-get install -y dnsutils

RUN pip install --no-cache-dir -r requirements.txt

# 不再复制本地代码
CMD ["python" , "/home/validator_scheduler/app/validator_rq_scheduler.py"]

3. 将代码上传到Persistent Volume #

把validator_rq_scheduler.py放到PV对应的存储中：

• 如果是HostPath类型PV，直接在节点的对应目录创建文件；
• 云存储或PVC可以用临时Pod上传：

  apiVersion: v1
  kind: Pod
  metadata:
    name: pv-uploader
  spec:
    containers:
      - name: uploader
        image: busybox
        command: ["sleep", "3600"]
        volumeMounts:
          - name: validator-scheduler
            mountPath: /data
    volumes:
      - name: validator-scheduler
        persistentVolumeClaim:
          claimName: validator-scheduler-pvc
  

  创建Pod后执行kubectl cp validator_rq_scheduler.py pv-uploader:/data/，完成后删除临时Pod。

4. 代码更新方案 #

• 手动更新：重复上传文件步骤，然后执行kubectl rollout restart deployment/validator-rq-scheduler-deployment重启Pod；
• 版本控制自动更新：用GitSync工具或自定义Sidecar容器，定期从Git仓库拉取代码到PV目录。

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