Kubesphere deploys Kubernetes using external IP

Our company has always had the need to connect all the cloud servers we currently manage. We have not been able to make an overall cluster and manage it with Kubernetes. Today, after modifying the original network model and performing relevant configuration on the server, we have finally connected all servers that can only be accessed through external network IP.

We have deployed Kubernetes using the Calico network model. After deployment, we found that all servers need to be connected through a virtual network before they can be connected together. As a unified distributed server, this has caused a relatively large limitation on the expansion of our servers. So we changed the default network model of kubesphere to flannel. Each deployment takes a long time, and basically each node cannot communicate directly with each other. After looking for information, we modified the network configuration in the server:

First, add a virtual network card to all servers and bind the external IP to this virtual network card. Note that it is best to add :1 to the current active network card to indicate that it is the configuration information on the current network card.

 cat > / etc / sysconfig / network - scripts / ifcfg - eth0 : 1 << EOF
 BOOTPROTO = static
 DEVICE = eth0 : 1
 IPADDR = your public IP
 PREFIX = 32
 TYPE = Ethernet
 USERCTL = no
 ONBOOT = yes
 EOF

After restarting, use ip addr to check whether it has taken effect. If it has taken effect, you can use kubesphere to install multiple nodes. Also note that the node address and master address in config.yaml both use the external network IP.