### Terraform Version ```shell Terraform v1.13.1 on linux_amd64 + provider registry.terraform.io/hashicorp/aws v6.11.0 + provider registry.terraform.io/hashicorp/http v3.5.0 + provider registry.terraform.io/hashicorp/local v2.5.3 + provider registry.terraform.io/hashicorp/random v3.7.2 + provider registry.terraform.io/hashicorp/tls v4.1.0 ``` ### Terraform Configuration Files ``` variable "eks_k8s_version" { description = "If a Kubernetes version other than the current default for the default EKS version is desired, set this variable to the desired K8s version. Note that if you set this to a version in extended support status, a warning will be generated but the cluster will still be provisioned and may incur extra costs." type = string default = "" } resource "aws_eks_cluster" "simple_eks" { name = "[cluster name]" [...] version = var.eks_k8s_version == "" ? null : var.eks_k8s_version [...] } resource "aws_eks_node_group" "ec2" { cluster_name = aws_eks_cluster.simple_eks.name node_group_name = "[nodegroup name]" version = aws_eks_cluster.simple_eks.version [...] } ``` ### Debug Output Running with trace on generated so much trace that it ran off my scrollback, but if needed I can generate a minimal test case config that I can share and that hopefully will not generate quite so much trace output. ### Expected Behavior Terraform should have either upgraded both the cluster and nodegroup in-place, or destroyed and then recreated both. ### Actual Behavior Terraform attempts to destroy the cluster but only change the nodegroup, which results in the following failure: ``` Error: deleting EKS Cluster ([cluster-name]): operation error EKS: DeleteCluster, https response error StatusCode: 409, RequestID: 64ba437d-f278-4dec-9ed1-af7ffb0828fb, ResourceInUseException: Cluster has nodegroups attached ``` ### Steps to Reproduce 1. Create an `aws_eks_cluster` and `aws_eks_node_group` with an explicit `version`. 2. Update the resources with a null `version` (e.g. by deleting the `version` arguments). 3. Apply again. No changes are planned. 4. Taint the `aws_eks_cluster` resource. 5. Apply again. Terraform generates a plan that will delete the existing `aws_eks_cluster`, create a new `aws_eks_cluster`, and update the `aws_eks_node_group` in-place, however the cluster delete fails with the error noted above. ### Additional Context If I taint the nodegroup resource as well, then both resources are deleted and recreated in the correct order. ### References This appears to be the same error that was reported in #29732. ### Generative AI / LLM assisted development? None