Imported EKS cluster stuck in Waiting for API to be available state

[peishuli]

Hello there! I am evaluating using Rancher to manage our k8s clusters. Rancher is installed with the stable helm chart (v2.6.9) running on a local k3 cluster (k8s version 1.24.4). I was trying to import an existing EKS cluster (k8s version 1.24.7) but the state of the imported cluster stuck in "Waiting for API to be available" state. I researched the issue over the Internet and tried a few things as people suggested but nothing worked so far. What I've tried include:

• Terminate eks nodes (ec2 instances) and let the Autoscaling Group automatically relaunch new instances
• Patch the rancher eks opertor: kubectl -n cattle-system patch deploy/eks-config-operator -p '{"spec":{"template":{"spec":{"dnsPolicy":"Default"}}}}'

Your help in this matter will be greatly appreciated!

Thanks, Peishu

[peishuli]

Hi there! I figured out what's going on - I used k8s port-forward to access rancher console. e.g., https://localhost:9443. When I imported the EKS cluster from the UI console, the generated cluster agent pod in EKS will try to communicate with rancher server via localhost:9443 and thus failed. I moved Rancher to an EKS cluster with DNS name configured and now it works fine.

Please close the ticket!

Thanks, Peishu