kubernetes上启动tensorflow:latest-gpu-jupyter

阳光穿透心脏的1/2处 2022-12-01 01:25 71阅读 0赞

首先，按[https://blog.csdn.net/vah101/article/details/108098827][https_blog.csdn.net_vah101_article_details_108098827]这里提到的方式为k8s集群配置好GPU结点，

之后编写如下的tensflow.yaml文件

apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: tensorflow-gpu-jupyter
      labels:
        app: tensorflow-gpu-jupyter
    spec:
      replicas: 1
      selector: # define how the deployment finds the pods it mangages
        matchLabels:
          app: tensorflow-gpu-jupyter
      template: # define the pods specifications
        metadata:
          labels:
            app: tensorflow-gpu-jupyter
        spec:
          containers:
          - name: tensorflow-gpu-jupyter
            image: tensorflow/tensorflow:latest-gpu-jupyter
            resources:
              limits:
                nvidia.com/gpu: 1
    ---
    apiVersion: v1
    kind: Service
    metadata:
     name: tensorflow-gpu-jupyter
     labels:
      app: tensorflow-gpu-jupyter
    spec:
     type: NodePort
     ports:
     - port: 8888
       targetPort: 8888
       nodePort: 30888
     selector:
      app: tensorflow-gpu-jupyter
    ---
    apiVersion: extensions/v1beta1
    kind: Ingress
    metadata:
      name: tensorflow-gpu-jupyter
    spec:
      rules:
      - host: gpu.k8s
        http:
          paths:
          - backend:
              serviceName: tensorflow-gpu-jupyter
              servicePort: 8888
            path: /
    status:
      loadBalancer: {}

执行：

kubectl create -f tensflow.yaml

之后，通过kubectl get pod 检查tensflow是否启动，如果启动会显示如下内容：

NAME                                      READY   STATUS    RESTARTS   AGE
    tensorflow-gpu-jupyter-66874fd7db-rvjt2   1/1     Running   0          4m29s

之后执行kubectl logs -f tensorflow-gpu-jupyter-66874fd7db-rvjt2

会看到日志显示：

[I 07:56:42.097 NotebookApp] Writing notebook server cookie secret to /root/.local/share/jupyter/runtime/notebook_cookie_secret
    [I 07:56:42.318 NotebookApp] Serving notebooks from local directory: /tf
    [I 07:56:42.318 NotebookApp] The Jupyter Notebook is running at:
    [I 07:56:42.318 NotebookApp] http://tensorflow-gpu-jupyter-66874fd7db-rvjt2:8888/?token=1ba80e3978b1c7da673aae668c9750105b37cb3efd0b223a
    [I 07:56:42.318 NotebookApp]  or http://127.0.0.1:8888/?token=1ba80e3978b1c7da673aae668c9750105b37cb3efd0b223a
    [I 07:56:42.318 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
    [C 07:56:42.321 NotebookApp] 
        
        To access the notebook, open this file in a browser:
            file:///root/.local/share/jupyter/runtime/nbserver-1-open.html
        Or copy and paste one of these URLs:
            http://tensorflow-gpu-jupyter-66874fd7db-rvjt2:8888/?token=1ba80e3978b1c7da673aae668c9750105b37cb3efd0b223a
         or http://127.0.0.1:8888/?token=1ba80e3978b1c7da673aae668c9750105b37cb3efd0b223a

说明启动成功，在浏览器中输入

master结点ip:30888/?token=1ba80e3978b1c7da673aae668c9750105b37cb3efd0b223a

即可进入jupyter界面

[https_blog.csdn.net_vah101_article_details_108098827]: https://blog.csdn.net/vah101/article/details/108098827