Because docker lacked a docker bridge,So i created a new bridge and add it to docker's daemon.json file. But In my k8s cluster, the master is a single node, not HA. Now I need to restart all my work nodes and master node in my cluster. Is there anything I should pay attention to?
If you are using replicaset, daemonset, deployment, statefulset etc for which there is a controller that is always trying to converge the state of your workload pods to the desired state there will be new pods created automatically. You will experience a brief downtime until new pods are spawned. But if you are running pods directly then pods will not be recreated. That's why it's not recommended to deploy pods directly.
Since you have single master node which is also hosting etcd you need to backup ETCD data and restore it after you restarted master.