The bug report for this issue is here
The underlying cause is that the AWS cli shipped a breaking change in a minor version release. You can see this here
I'm assuming here you're using the pulumi-eks
package in order to provision an EKS cluster greater than v1.22
. The EKS package uses a resource provider to configure some EKS resources like the aws-auth
config map, and this isn't the same transient kubeconfig you're referring to in ~/.kube/config
In order to fix this, you need to do the following:
- Ensure your
aws-cli
version is greater than 1.24.0
or 2.7.0
- Ensure you've updated your
pulumi-eks
package in your language SDK package manager to greater than 0.40.0
. This will mean also updated the provider in your existing stack.
- Ensure you have the version of
kubectl
installed locally that matches your cluster version that has been provisioned