In k8s, start pod failed, debugging method.

         In k8s, it is not possible to start the container directly. The container must be started in the form of a pod. When the pod fails to start, you can debug like this.

        1. Find the image started by the pod, re-encapsulate the image, and modify the entry of the image

FROM xxxxx            //pod启动的镜像
ENTRYPOINT ["python", "-m", "SimpleHTTPServer", "8080"] //修改镜像的入口

        2. Upload the image to the image library

        3. Modify the yaml file of the pod, modify the mirror address, modify the probe

containers:
    - name: containers_name
    image: image_path

//这里要注意要把levenesspoint给屏蔽了，要不会一直kill掉containers，然后重启

        4. Use kubectl exec -it -n <namespaces> <podname> /bin/sh to enter the containers to run your own code and debug.