I followed the guides to launch a k8s cluster with a p4de instance and I think the default security groups it created had two groups with the kubernetes.io/cluster/k8s-gpu-cluster tag included. When trying to later create an ingress load balancer using the aws load balancer controller, the loadbalancer failed to configure everything.
This matches this error report and deleting the tag in one security group resolved the issue.
It'd be very helpful to figure out where this double tagging is happening and add a fix so that eks clusters are well configured for load balancers.
I followed the guides to launch a k8s cluster with a p4de instance and I think the default security groups it created had two groups with the
kubernetes.io/cluster/k8s-gpu-clustertag included. When trying to later create an ingress load balancer using the aws load balancer controller, the loadbalancer failed to configure everything.This matches this error report and deleting the tag in one security group resolved the issue.
It'd be very helpful to figure out where this double tagging is happening and add a fix so that eks clusters are well configured for load balancers.