[feature] Support running single node.

[runkecheng]

Is your feature request related to a problem? Please describe.

When the FOLLOWER node of the cluster cannot quickly recover quickly, cluster automatically switches to single-node to ensure the fastest recovery service.

Describe the solution you'd like

• Support read and write single node
• Support automatically switches into single nodes when necessary

Describe alternatives you've considered

Additional context