GKE networking (VPC-native) in GCP - Step-by-Step Execution

Process Flow - GKE networking (VPC-native)

Create VPC Network

↓

Enable Alias IP Ranges

↓

Create Subnet with Secondary Ranges

↓

Create GKE Cluster with VPC-native Enabled

↓

Pods get IPs from Secondary Range

↓

Pods communicate using VPC routing

↓

Cluster nodes and Pods share VPC network

This flow shows how a VPC-native GKE cluster is set up to use VPC networking with alias IP ranges, allowing pods to get IPs from a subnet's secondary range and communicate within the VPC.

Execution Sample

GCP

gcloud container clusters create my-cluster \
  --enable-ip-alias \
  --create-subnetwork name=my-subnet,range=10.0.0.0/24,secondary-range-pods=10.0.1.0/24,secondary-range-services=10.0.2.0/24 \
  --network my-vpc

This command creates a GKE cluster with VPC-native networking enabled, using a subnet with secondary IP ranges for pods and services.

Process Table

Step	Action	Resource Created/Modified	IP Range Assigned	Result
1	Create VPC network	VPC network 'my-vpc'	10.0.0.0/16	VPC network ready
2	Create subnet with secondary ranges	Subnet 'my-subnet'	Primary: 10.0.0.0/24 Pods: 10.0.1.0/24 Services: 10.0.2.0/24	Subnet with alias IP ranges ready
3	Create GKE cluster with --enable-ip-alias	GKE cluster 'my-cluster'	Pods use 10.0.1.0/24 Services use 10.0.2.0/24	Cluster nodes and pods get IPs from VPC ranges
4	Pods scheduled	Pods in cluster	IPs from 10.0.1.0/24	Pods communicate over VPC network
5	Services created	Kubernetes services	IPs from 10.0.2.0/24	Services reachable inside VPC
6	Cluster running	All resources	N/A	VPC-native networking active

💡 All resources created and pods/services assigned IPs from secondary ranges, enabling VPC-native networking.

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	Final
VPC Network	None	my-vpc (10.0.0.0/16)	my-vpc (10.0.0.0/16)	my-vpc (10.0.0.0/16)	my-vpc (10.0.0.0/16)	my-vpc (10.0.0.0/16)
Subnet	None	None	my-subnet (10.0.0.0/24) with pods 10.0.1.0/24 and services 10.0.2.0/24	my-subnet (10.0.0.0/24) with alias ranges	my-subnet (10.0.0.0/24) with alias ranges	my-subnet (10.0.0.0/24) with alias ranges
Cluster	None	None	None	my-cluster with IP alias enabled	my-cluster running	my-cluster running
Pods IP Range	None	None	None	10.0.1.0/24	Pods assigned IPs in 10.0.1.0/24	Pods assigned IPs in 10.0.1.0/24
Services IP Range	None	None	None	10.0.2.0/24	Services assigned IPs in 10.0.2.0/24	Services assigned IPs in 10.0.2.0/24

Key Moments - 3 Insights

Why do pods get IPs from the secondary range and not the primary subnet range?

Pods get IPs from the secondary range because the cluster is created with --enable-ip-alias and the subnet has secondary IP ranges assigned for pods and services, as shown in execution_table step 3 and 4.

What happens if the subnet does not have secondary IP ranges configured?

Without secondary IP ranges, the cluster cannot assign IPs to pods from the VPC, so VPC-native networking cannot be enabled. This is clear from step 2 where secondary ranges are created before cluster creation.

How do pods communicate with other resources in the VPC?

Pods use IPs from the VPC subnet's secondary range, so they communicate over the VPC network like any other resource, as shown in step 5 and 6.

Visual Quiz - 3 Questions

Look at the execution_table, at which step are pods assigned their IPs from the secondary range?

A Step 2

B Step 3

C Step 4

D Step 5

💡 Hint

Check the 'IP Range Assigned' and 'Result' columns in execution_table rows for pod IP assignment.

According to variable_tracker, what is the state of the subnet after step 2?

A Subnet with secondary ranges for pods and services

B Subnet with primary range only

C Subnet with no secondary ranges

D Subnet deleted

💡 Hint

Look at the 'Subnet' row after 'After Step 2' column in variable_tracker.

If the cluster was created without --enable-ip-alias, how would the execution_table change?

A Pods would get IPs from primary subnet range

B Pods would not get IPs from VPC ranges, breaking VPC-native networking

C Pods would still get IPs from secondary range

D Services would get IPs from secondary range but pods would not

💡 Hint

Refer to the importance of --enable-ip-alias in step 3 of execution_table.

Concept Snapshot

GKE VPC-native networking uses alias IP ranges.
Create a VPC and subnet with secondary IP ranges.
Enable --enable-ip-alias when creating the cluster.
Pods get IPs from secondary ranges, communicating inside VPC.
This improves IP management and network performance.

Full Transcript

This visual execution trace shows how to set up Google Kubernetes Engine (GKE) networking using VPC-native mode. First, a VPC network is created. Then, a subnet is created with secondary IP ranges for pods and services. Next, a GKE cluster is created with the --enable-ip-alias flag, which enables pods to receive IP addresses from the subnet's secondary ranges. Pods and services get IPs from these ranges and communicate over the VPC network. The variable tracker shows how resources and IP ranges evolve step-by-step. Key moments clarify why secondary ranges are needed and how pods communicate. The quiz tests understanding of when pods get IPs, subnet state, and the role of --enable-ip-alias.