Make session and jdbc interface more adaptable to the cluster

While the cluster is running, the coordinator node may go down, which may cause the session/jdbc to be unable to read or write properly. However, it is possible that most of the nodes in the cluster are running, and a different coordinator node can work just fine. Therefore, it is necessary to make the Session/jdbc interface more adaptable to the cluster, including but not limited to transparent fault tolerance for the business layer and query load balancing.