fix: Disable creation of cluster security group rules that map to nod…

…e security group when `create_node_security_group` = `false` (#2274) * fix: Disable creation of cluster security group rules that map to node security group when `create_node_security_group` = `false` * feat: Update Fargate example to run only Fargate and show disabling of both cluster and node security groups * fix: Ensure CoreDNS changes are made ahead of install
terraform-aws-modules · Oct 14, 2022 · 28ccece · 28ccece
1 parent 8dc5ad4
commit 28ccece
Show file tree

Hide file tree

Showing 5 changed files with 171 additions and 53 deletions.
diff --git a/docs/faq.md b/docs/faq.md
@@ -26,7 +26,7 @@ By default, EKS creates a cluster primary security group that is created outside
   attach_cluster_primary_security_group = true # default is false
 
   node_security_group_tags = {
-    "kubernetes.io/cluster/<CLUSTER_NAME>" = "" # or any other value other than "owned"
+    "kubernetes.io/cluster/<CLUSTER_NAME>" = null # or any other value other than "owned"
   }
 ```
 
@@ -36,7 +36,7 @@ By default, EKS creates a cluster primary security group that is created outside
   attach_cluster_primary_security_group = true # default is false
 
   cluster_tags = {
-    "kubernetes.io/cluster/<CLUSTER_NAME>" = "" # or any other value other than "owned"
+    "kubernetes.io/cluster/<CLUSTER_NAME>" = null # or any other value other than "owned"
   }
 ```
 

diff --git a/examples/fargate_profile/README.md b/examples/fargate_profile/README.md
@@ -21,13 +21,16 @@ Note that this example may create resources which cost money. Run `terraform des
 |------|---------|
 | <a name="requirement_terraform"></a> [terraform](#requirement\_terraform) | >= 0.13.1 |
 | <a name="requirement_aws"></a> [aws](#requirement\_aws) | >= 3.72 |
-| <a name="requirement_kubernetes"></a> [kubernetes](#requirement\_kubernetes) | >= 2.10 |
+| <a name="requirement_helm"></a> [helm](#requirement\_helm) | >= 2.7 |
+| <a name="requirement_null"></a> [null](#requirement\_null) | >= 3.0 |
 
 ## Providers
 
 | Name | Version |
 |------|---------|
 | <a name="provider_aws"></a> [aws](#provider\_aws) | >= 3.72 |
+| <a name="provider_helm"></a> [helm](#provider\_helm) | >= 2.7 |
+| <a name="provider_null"></a> [null](#provider\_null) | >= 3.0 |
 
 ## Modules
 
@@ -41,6 +44,11 @@ Note that this example may create resources which cost money. Run `terraform des
 | Name | Type |
 |------|------|
 | [aws_kms_key.eks](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/kms_key) | resource |
+| [helm_release.coredns](https://registry.terraform.io/providers/hashicorp/helm/latest/docs/resources/release) | resource |
+| [null_resource.modify_kube_dns](https://registry.terraform.io/providers/hashicorp/null/latest/docs/resources/resource) | resource |
+| [null_resource.remove_default_coredns_deployment](https://registry.terraform.io/providers/hashicorp/null/latest/docs/resources/resource) | resource |
+| [aws_eks_addon_version.this](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/data-sources/eks_addon_version) | data source |
+| [aws_eks_cluster_auth.this](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/data-sources/eks_cluster_auth) | data source |
 
 ## Inputs
 

diff --git a/examples/fargate_profile/main.tf b/examples/fargate_profile/main.tf
@@ -2,6 +2,20 @@ provider "aws" {
   region = local.region
 }
 
+provider "helm" {
+  kubernetes {
+    host                   = module.eks.cluster_endpoint
+    cluster_ca_certificate = base64decode(module.eks.cluster_certificate_authority_data)
+
+    exec {
+      api_version = "client.authentication.k8s.io/v1beta1"
+      command     = "aws"
+      # This requires the awscli to be installed locally where Terraform is executed
+      args = ["eks", "get-token", "--cluster-name", module.eks.cluster_id]
+    }
+  }
+}
+
 locals {
   name            = "ex-${replace(basename(path.cwd), "_", "-")}"
   cluster_version = "1.22"
@@ -27,14 +41,8 @@ module "eks" {
   cluster_endpoint_public_access  = true
 
   cluster_addons = {
-    # Note: https://docs.aws.amazon.com/eks/latest/userguide/fargate-getting-started.html#fargate-gs-coredns
-    coredns = {
-      resolve_conflicts = "OVERWRITE"
-    }
     kube-proxy = {}
-    vpc-cni = {
-      resolve_conflicts = "OVERWRITE"
-    }
+    vpc-cni    = {}
   }
 
   cluster_encryption_config = [{
@@ -45,28 +53,13 @@ module "eks" {
   vpc_id     = module.vpc.vpc_id
   subnet_ids = module.vpc.private_subnets
 
-  # You require a node group to schedule coredns which is critical for running correctly internal DNS.
-  # If you want to use only fargate you must follow docs `(Optional) Update CoreDNS`
-  # available under https://docs.aws.amazon.com/eks/latest/userguide/fargate-getting-started.html
-  eks_managed_node_groups = {
-    example = {
-      desired_size = 1
-
-      instance_types = ["t3.large"]
-      labels = {
-        Example    = "managed_node_groups"
-        GithubRepo = "terraform-aws-eks"
-        GithubOrg  = "terraform-aws-modules"
-      }
-      tags = {
-        ExtraTag = "example"
-      }
-    }
-  }
+  # Fargate profiles use the cluster primary security group so these are not utilized
+  create_cluster_security_group = false
+  create_node_security_group    = false
 
   fargate_profiles = {
-    default = {
-      name = "default"
+    example = {
+      name = "example"
       selectors = [
         {
           namespace = "backend"
@@ -75,15 +68,18 @@ module "eks" {
           }
         },
         {
-          namespace = "default"
+          namespace = "app-*"
           labels = {
-            WorkerType = "fargate"
+            Application = "app-wildcard"
           }
         }
       ]
 
+      # Using specific subnets instead of the subnets supplied for the cluster itself
+      subnet_ids = [module.vpc.private_subnets[1]]
+
       tags = {
-        Owner = "default"
+        Owner = "secondary"
       }
 
       timeouts = {
@@ -92,29 +88,138 @@ module "eks" {
       }
     }
 
-    secondary = {
-      name = "secondary"
+    kube_system = {
+      name = "kube-system"
       selectors = [
-        {
-          namespace = "default"
-          labels = {
-            Environment = "test"
-            GithubRepo  = "terraform-aws-eks"
-            GithubOrg   = "terraform-aws-modules"
-          }
-        }
+        { namespace = "kube-system" }
       ]
+    }
+  }
 
-      # Using specific subnets instead of the subnets supplied for the cluster itself
-      subnet_ids = [module.vpc.private_subnets[1]]
+  tags = local.tags
+}
 
-      tags = {
-        Owner = "secondary"
+################################################################################
+# Modify EKS CoreDNS Deployment
+################################################################################
+
+data "aws_eks_cluster_auth" "this" {
+  name = module.eks.cluster_id
+}
+
+locals {
+  kubeconfig = yamlencode({
+    apiVersion      = "v1"
+    kind            = "Config"
+    current-context = "terraform"
+    clusters = [{
+      name = module.eks.cluster_id
+      cluster = {
+        certificate-authority-data = module.eks.cluster_certificate_authority_data
+        server                     = module.eks.cluster_endpoint
+      }
+    }]
+    contexts = [{
+      name = "terraform"
+      context = {
+        cluster = module.eks.cluster_id
+        user    = "terraform"
       }
+    }]
+    users = [{
+      name = "terraform"
+      user = {
+        token = data.aws_eks_cluster_auth.this.token
+      }
+    }]
+  })
+}
+
+# Separate resource so that this is only ever executed once
+resource "null_resource" "remove_default_coredns_deployment" {
+  triggers = {}
+
+  provisioner "local-exec" {
+    interpreter = ["/bin/bash", "-c"]
+    environment = {
+      KUBECONFIG = base64encode(local.kubeconfig)
     }
+
+    # We are removing the deployment provided by the EKS service and replacing it through the self-managed CoreDNS Helm addon
+    # However, we are maintaing the existing kube-dns service and annotating it for Helm to assume control
+    command = <<-EOT
+      kubectl --namespace kube-system delete deployment coredns --kubeconfig <(echo $KUBECONFIG | base64 --decode)
+    EOT
   }
+}
 
-  tags = local.tags
+resource "null_resource" "modify_kube_dns" {
+  triggers = {}
+
+  provisioner "local-exec" {
+    interpreter = ["/bin/bash", "-c"]
+    environment = {
+      KUBECONFIG = base64encode(local.kubeconfig)
+    }
+
+    # We are maintaing the existing kube-dns service and annotating it for Helm to assume control
+    command = <<-EOT
+      echo "Setting implicit dependency on ${module.eks.fargate_profiles["kube_system"].fargate_profile_pod_execution_role_arn}"
+      kubectl --namespace kube-system annotate --overwrite service kube-dns meta.helm.sh/release-name=coredns --kubeconfig <(echo $KUBECONFIG | base64 --decode)
+      kubectl --namespace kube-system annotate --overwrite service kube-dns meta.helm.sh/release-namespace=kube-system --kubeconfig <(echo $KUBECONFIG | base64 --decode)
+      kubectl --namespace kube-system label --overwrite service kube-dns app.kubernetes.io/managed-by=Helm --kubeconfig <(echo $KUBECONFIG | base64 --decode)
+    EOT
+  }
+
+  depends_on = [
+    null_resource.remove_default_coredns_deployment
+  ]
+}
+
+################################################################################
+# CoreDNS Helm Chart (self-managed)
+################################################################################
+
+data "aws_eks_addon_version" "this" {
+  for_each = toset(["coredns"])
+
+  addon_name         = each.value
+  kubernetes_version = module.eks.cluster_version
+  most_recent        = true
+}
+
+resource "helm_release" "coredns" {
+  name             = "coredns"
+  namespace        = "kube-system"
+  create_namespace = false
+  description      = "CoreDNS is a DNS server that chains plugins and provides Kubernetes DNS Services"
+  chart            = "coredns"
+  version          = "1.19.4"
+  repository       = "https://coredns.github.io/helm"
+
+  # For EKS image repositories https://docs.aws.amazon.com/eks/latest/userguide/add-ons-images.html
+  values = [
+    <<-EOT
+      image:
+        repository: 602401143452.dkr.ecr.eu-west-1.amazonaws.com/eks/coredns
+        tag: ${data.aws_eks_addon_version.this["coredns"].version}
+      deployment:
+        name: coredns
+        annotations:
+          eks.amazonaws.com/compute-type: fargate
+      service:
+        name: kube-dns
+        annotations:
+          eks.amazonaws.com/compute-type: fargate
+      podAnnotations:
+        eks.amazonaws.com/compute-type: fargate
+      EOT
+  ]
+
+  depends_on = [
+    # Need to ensure the CoreDNS updates are peformed before provisioning
+    null_resource.modify_kube_dns
+  ]
 }
 
 ################################################################################

diff --git a/examples/fargate_profile/versions.tf b/examples/fargate_profile/versions.tf
@@ -6,9 +6,13 @@ terraform {
       source  = "hashicorp/aws"
       version = ">= 3.72"
     }
-    kubernetes = {
-      source  = "hashicorp/kubernetes"
-      version = ">= 2.10"
+    helm = {
+      source  = "hashicorp/helm"
+      version = ">= 2.7"
+    }
+    null = {
+      source  = "hashicorp/null"
+      version = ">= 3.0"
     }
   }
 }
diff --git a/main.tf b/main.tf
@@ -128,7 +128,8 @@ locals {
 
   cluster_security_group_id = local.create_cluster_sg ? aws_security_group.cluster[0].id : var.cluster_security_group_id
 
-  cluster_security_group_rules = {
+  # Do not add rules to node security group if the module is not creating it
+  cluster_security_group_rules = local.create_node_sg ? {
     ingress_nodes_443 = {
       description                = "Node groups to cluster API"
       protocol                   = "tcp"
@@ -153,7 +154,7 @@ locals {
       type                       = "egress"
       source_node_security_group = true
     }
-  }
+  } : {}
 }
 
 resource "aws_security_group" "cluster" {