As of RKE
v1.3.0, dual-stack networking support has been added, which allows allocating both IPv4 and IPv6 addresses to pods and services.
In order to use the dual-stack feature, RKE and the infrastructure it's deploy to must be configured as follows:
- Kubernetes 1.21 or newer is used.
- RKE is configured to use Calico as the Container Network Interface (CNI) provider. Other providers are not supported.
- RKE is deployed on Amazon EC2 instances with the following prerequisites:
- Enable IPv6 support: set the network range at VPC and its subnetworks.
- Add a IPv6 default gateway to VPC routes.
- Add inbound/outbound rules for IPv6 traffic to your cluster's security group(s).
- Ensure instances have
Auto-assign IPv6 IPenabled. See the AWS documentation for instructions.
- Disable source/destination checks on all instances in the cluster. See the AWS documentation for instructions.
For more information on configuring your AWS enivronment for IPv6, refer to the AWS Getting started with IPv6 documentation.
Example RKE Configuration
The following is an example RKE configuration that can be used to deploy RKE with dual-stack support configured: