What happened:

If kubelet plugin registration fails during the registration process, the kubelet never attempts to register the socket again (assuming the error was transient), unless the socket for the driver is deleted & recreated to trigger a new registration.

What you expected to happen:
Maybe retry with exponential backoff when kubelet plugin registration fails instead of just logging and waiting for another filesystem event like it does today.

Of course, if a socket returns a terminal error (does not implement the kubelet plugin registration service, for example), we should not retry.

Follow up on #71440

/milestone v1.14
/sig storage

How to reproduce it (as minimally and precisely as possible):

Anything else we need to know?:

Environment:

• Kubernetes version (use kubectl version):
• Cloud provider or hardware configuration:
• OS (e.g. from /etc/os-release):
• Kernel (e.g. uname -a):
• Install tools:
• Others:

/kind bug