self managed node group fails with default cluster_version = null

[aidanmelen]

Description

The self managed node group uses the cluster_version input used to lookup default AMI ID if one is not provided. But the default for cluster_version is null.

The data.aws_ami.eks_default data resource will fail to filter with the null value. e.g. amazon-eks-node-null-v*

Versions

• Terraform: v1.1.7

• Provider(s): terraform-aws-modules/eks/aws (0.18.3)

• Sub-Module: self-managed-node-group

Reproduction

Steps to reproduce the behavior:

1. run terraform-aws-modules/eks/aws with self_managed_node_groups sub-module and don't set cluster_version

Code Snippet to Reproduce

module "eks" {
  source  = "terraform-aws-modules/eks/aws"
  version = ">= 18.0.0"

  cluster_name    = local.name
  # cluster_version = "1.21"
  vpc_id          = module.vpc.vpc_id
  subnet_ids      = module.vpc.private_subnets

  self_managed_node_groups = {
    boo = {
      instance_type = "t3.medium"
      instance_market_options = {
        market_type = "spot"
      }
    }
  }
}

Expected behavior

defaults to the cluster version.

Actual behavior

defaults to null

Terminal Output Screenshot(s)

Additional context

one possible solution is to retrieve the cluster_version from a data resource and set that as the default

data "aws_eks_cluster" "example" {
  name = "example"
}

output "dynamic_cluster_version" {
  value = data.aws_eks_cluster.example.version
}

[bryantbiggs]

I think its better if we just rely on users specifying the version - it aligns with the other sub-modules as well

[aidanmelen]

it aligns with the other sub-modules as well

The behavior of the self managed node group is inconsistent with the other sub-modules. for example

module "eks" {
  source  = "terraform-aws-modules/eks/aws"
  version = ">= 18.0.0"

  cluster_name    = local.name
  # cluster_version = "1.21"

  vpc_id     = module.vpc.vpc_id
  subnet_ids = module.vpc.private_subnets

  self_managed_node_groups = {
    boo = {
      instance_type = "t3.medium"
      instance_market_options = {
        market_type = "spot"
      }
    }
  }
}

FAIL: The plan fails as the self managed node group fails to look up the EKS AMI with latest cluster version.

╷
│ Error: Invalid template interpolation value
│ 
│   on ../../modules/self-managed-node-group/main.tf line 10, in data "aws_ami" "eks_default":
│   10:     values = ["amazon-eks-node-${var.cluster_version}-v*"]
│     ├────────────────
│     │ var.cluster_version is null
│ 
│ The expression result is null. Cannot include a null value in a string template.
╵

module "eks" {
  source  = "terraform-aws-modules/eks/aws"
  version = ">= 18.0.0"

  cluster_name    = local.name
  # cluster_version = "1.21"

  vpc_id     = module.vpc.vpc_id
  subnet_ids = module.vpc.private_subnets

  eks_managed_node_groups = {
    foo = {}
  }
}

PASS: creates eks managed node group on latest cluster version

module "eks" {
  source  = "terraform-aws-modules/eks/aws"
  version = ">= 18.0.0"

  cluster_name    = local.name
  # cluster_version = "1.21"

  vpc_id     = module.vpc.vpc_id
  subnet_ids = module.vpc.private_subnets

  fargate_profiles = {
    bar = {
      selectors = [
        {
          namespace = "bar"
        }
      ]
    }
  }
}

PASS: creates fargate profile on latest cluster version

[bryantbiggs]

The behavior of the self managed node group is inconsistent with the other sub-modules. for example

Self managed node groups by their very nature are different from EKS managed or Fargate profiles because all management is up to users

[github-actions]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues. If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.