CNI Common Questions

1. Why did the CNI plugin upgrade fail?

1. Check if the node has set taints and prohibited scheduling (the default taint of the master node does not count). Nodes with taints need to choose to forcefully upgrade (which will not have special effects on the node).
2. If the forced upgrade fails, please click on 'cni force upgrade' again.
3. Execute the kubectl get pods -n kube-system |grep plugin-operation to find the associated plugin upgrade pod, and use describe pod to see the reason for the pod's failure.

2. Why can't my cluster connect to the external network?

UK8S uses VPC network for internal communication, and uses SurferCloud's DNS by default. To get information using wget, you need to configure a gateway for the subnets of the VPC. You need to go to the VPC product in the area where UK8S is located, and configure a NAT gateway for the specific subnet, so that cluster nodes can pull data from the external network through the NAT gateway. For specific operations, see VPC Creation of NAT Gateway.