liveBook · Manning

Virtual ip address

As explained in Google Anthos in Action

A virtual IP address is a logical network address that can be assigned to a service or endpoint independently of a particular physical interface or host.

A virtual IP address (VIP) gives clients a stable address for a service while traffic is directed to its backends. Inside a Kubernetes cluster, a Service’s ClusterIP acts as a VIP for multiple Pods. It remains static even when Pods move, but is reachable only within the cluster. section 10.2.1

On Anthos bare metal, reserve VIPs before installation and choose bundled or manual load balancing. The bundled option creates or assigns VIPs from installation configuration; the manual option requires you to configure the load balancer and associate each VIP with its backends. section 17.3.4 section 10.2.1

Plan VIPs by endpoint:

  • Control plane: reserve one per cluster. With manual load balancing, map it to all control-plane node IPs on backend port 6444. section 17.3.4
  • Ingress: reserve one per user cluster. With manual load balancing, match the istio-ingress Service IP and use cluster nodes as backends on that Service’s NodePort. HTTP and HTTPS may require separate VIPs and backend pools. section 17.3.4
  • LoadBalancer Services: reserve one VIP for each Service; bundled load balancing assigns addresses from the configured pool. section 17.3.4

Specify a Service VIP

A LoadBalancer Service manifest can set `spec.loadBalancerIP: [SERVICE VIP]`. Choose that address before creating the cluster; in manual mode, configure the load balancer to match the Service address. section 10.2.1 section 17.3.4

Table 10.2 How to prepare for load balancing

 

Integrated/bundled mode

Manual mode

Choose VIPs before you create your clusters.

Yes

Yes

Choose node IP addresses before you create your clusters.

No, if using DHCPYes, if using static IP addresses

Yes

Choose nodePort values before you create your clusters.

No

Yes

Manually configure your load balancer

No

Yes

Compare bundled and manual preparation: both require VIP selection, while manual mode also requires node IPs, NodePort values, and load-balancer configuration. section 10.2.1

How Google Anthos in Action covers it

How deep it goes
Details VIP reservation, backend mapping, ports, and bundled-versus-manual preparation; explains ClusterIP behavior without tracing packet internals.
Examples
Uses control-plane and Ingress backend mappings, a LoadBalancer Service manifest, and a bundled-versus-manual preparation table.
What it assumes
Assumes familiarity with Kubernetes Services, cluster nodes, NodePorts, and load balancers while spelling out their VIP configuration requirements.
What you can do after
Operate/configure VIPs by reserving addresses and matching manual load-balancer backends to cluster endpoints.
How it connects
Connects VIPs to ClusterIP, Pods, Kubernetes Services, Ingress, NodePort, and load balancers.

Also explained in

Compare how 2 books explain virtual ip address →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Google Anthos in Action. The section links point to the parts of the book it draws on.