Facing Issue while Planning Eks Cluster, How to By Pass the error ?

[bangupoc]

Currently we are doing the validate and Plan for eks cluster. Getting the below error.

Error
│ Error: Failed to configure client: exec plugin: invalid apiVersion "client.authentication.k8s.io/v1alpha1"
│ with module.eks_k8s_role_mapping.kubernetes_config_map.eks_to_k8s_role_mapping,
on .terraform/modules/eks_k8s_role_mapping/modules/eks-k8s-role-mapping/main.tf line 33, in resource "kubernetes_config_map" "eks_to_k8s_role_mapping":
33: resource "kubernetes_config_map" "eks_to_k8s_role_mapping" {

in _envcommon seeing below segment

Code : _envcommon/services/eks-cluster.hcl line 151
iam_role_to_rbac_group_mapping = {
"arn:aws:iam::${local.common_vars.locals.account_ids[local.account_name]}:role/GruntworkAccountAccessRole" = ["system:masters"]
(dependency.ops_admin_role.outputs.iam_role_arn) = ["system:masters"]
(dependency.baseline.outputs.allow_dev_access_from_other_accounts_iam_role_arn) = ["system:masters"]
(dependency.ecs_deploy_runner.outputs.ecs_task_iam_roles["terraform-planner"]["arn"]) = ["system:masters"]
(dependency.ecs_deploy_runner.outputs.ecs_task_iam_roles["terraform-applier"]["arn"]) = ["system:masters"]
}

As we are using Plan only currently Not sure why this error is coming up ? Is there any other way to by pass and do the plan ?

---

Tracked in ticket #109834

[ryehowell]

Hi @bangupoc. Could you provide what version of the service catalog/EKS module(s) you are using as well as what version of Kuberntes/EKS you are specifying?

[ryehowell]

Hi @bangupoc,

I just wanted to follow up to see if the provided workaround resolved the issue you were facing. Were you able to resolve the issue?

[bangupoc]

Hi @ryehowell still the error is not resolved tried using the ?ref=v0.100.0 and pined the kubernetes version to 2.17 as well still v1aplha1 error is coming up and its mentioned https://www-app.gruntwork.io/repos/v0.14.0/terraform-aws-eks/modules/eks-k8s-role-mapping/README.md#aws-auth-configmap-generator-binary the iam roles will not be created and they have provide kubectl ref any idea about it what needs to be done for it ?

[gorkemgoknar]

I am facing a similar problem (trying to get from v0.82.0 to v0.100.0 before AWS updates 1.21 to 1.22)
and did override so uses helm/kubernetes <2.17

Override in reference architecture does not seem to override this one : https://github.com/gruntwork-io/terraform-aws-eks/blob/94aef4221985bfd314ef33aaf0f956f035cadfa3/modules/eks-k8s-role-mapping/main.tf#L21

getting credentials: exec plugin is configured to use API version client.authentication.k8s.io/v1, plugin returned version client.authentication.k8s.io/v1alpha1
[ecs-deploy-runner][2023-02-18T17:52:05+0000] │ 
[ecs-deploy-runner][2023-02-18T17:52:05+0000] │   with module.eks_k8s_role_mapping.kubernetes_config_map.eks_to_k8s_role_mapping,
[ecs-deploy-runner][2023-02-18T17:52:05+0000] │   on .terraform/modules/eks_k8s_role_mapping/modules/eks-k8s-role-mapping/main.tf line 36, in resource "kubernetes_config_map" "eks_to_k8s_role_mapping":
[ecs-deploy-runner][2023-02-18T17:52:05+0000] │   36: resource "kubernetes_config_map" "eks_to_k8s_role_mapping" {
[ecs-deploy-runner][2023-02-18T17:52:05+0000] │ 
[ecs-deploy-runner][2023-02-18T17:52:05+0000] ╵
[ecs-deploy-runner][2023-02-18T17:52:05+0000] time=2023-02-18T17:52:05Z level=error msg=1 error occurred:
[ecs-deploy-runner][2023-02-18T17:52:05+0000] 	* exit status 1

[gorkemgoknar]

At last found the problem after a day
for a v1alpha1 supported api EKS (1.21,1.22,1.23) The final version that does not cause trouble is terraform-aws-service-catalog -> v0.99.2 . on v0.100.0 eks-cluster uses its own provider and uses v1 api (terraform override has no use). on v0.99.2 it uses v1alpha1 api. You need a provider kubernetes override withing eks cluster you have to use your own version, or use v0.99.2 (for supported kubernetes releases <=1.23 I think)

Note that for v0.100.0 you will have problems with other dependend modules that use rest api for > v0.100.0 service catalog or terraform-aws-eks.
You should still use on root provider override hashicorp/kubernetes v2.17.0 as 2.18 and onwards use v1. (it will give trouble with namespace etc and other modules if you do not).

[ryehowell]

Hi everyone, just for future reference, a KB article was created to address this issue. Please see the article for details. Hope this helps!