add boostrap action var

aws_datalake/modules/emr/main.tf

@@ -17,6 +17,16 @@ resource "aws_emr_cluster" "segment_data_lake_emr_cluster" {
   service_role           = var.iam_emr_service_role
   autoscaling_role       = var.iam_emr_autoscaling_role
   security_configuration = var.security_configuration
+
+  dynamic "bootstrap_action" {
+    for_each = var.bootstrap_action
+
+    content {
+      args = try(bootstrap_action.value.args, null)
+      name = bootstrap_action.value.name
+      path = bootstrap_action.value.path
+    }
+  }
 
   master_instance_group {
     instance_type = var.master_instance_type

aws_datalake/modules/emr/variables.tf

@@ -101,6 +101,12 @@ variable "task_instance_max_count" {
   default     = "4"
 }
 
+variable "bootstrap_action" {
+  description = "Ordered list of bootstrap actions that will be run before Hadoop is started on the cluster nodes"
+  type        = any
+  default     = {}
+}
+
 locals {
   tags = merge(tomap({"vendor" = "segment"}), var.tags)
 }

aws_datalake/modules/iam/main.tf

@@ -289,9 +289,8 @@ resource "aws_iam_instance_profile" "segment_emr_instance_profile" {
   role = aws_iam_role.segment_emr_instance_profile_role.name
 }
 
-resource "aws_iam_role_policy" "segment_emr_instance_profile_policy" {
+resource "aws_iam_policy" "segment_emr_instance_profile_policy" {
   name = "SegmentEMRInstanceProfilePolicy${var.suffix}"
-  role = aws_iam_role.segment_emr_instance_profile_role.id
 
   policy = <<EOF
 {
@@ -364,6 +363,17 @@ resource "aws_iam_role_policy" "segment_emr_instance_profile_policy" {
 EOF
 }
 
+resource "aws_iam_role_policy_attachment" "segment_emr_instance_profile_policy_attachment" {
+  role = aws_iam_role.segment_emr_instance_profile_role.name
+  policy_arn = aws_iam_policy.segment_emr_instance_profile_policy.arn
+}
+
+resource "aws_iam_role_policy_attachment" "segment_emr_instance_profile_policy_custom_attachment" {
+  count = var.attach_custom_emr_isntance_profile_policy ? 1 : 0
+  role = aws_iam_role.segment_emr_instance_profile_role.name
+  policy_arn = var.custom_emr_instance_profile_policy_arn
+}
+
 # IAM Role for EMR Autoscaling role
 resource "aws_iam_role" "segment_emr_autoscaling_role" {
   name = "SegmentEMRAutoscalingRole${var.suffix}"
@@ -389,6 +399,7 @@ EOF
   tags = local.tags
 }
 
+
 resource "aws_iam_role_policy" "segmnet_emr_autoscaling_policy" {
   name = "SegmentEMRAutoscalingPolicy${var.suffix}"
   role = aws_iam_role.segment_emr_autoscaling_role.id

aws_datalake/modules/iam/variables.tf

@@ -46,6 +46,18 @@ variable "tags" {
   default     = {}
 }
 
+variable "custom_emr_instance_profile_policy_arn" {
+  description = "ARN of a IAM policy. This policy will be attached to the instance profile role"
+  type        = string
+  default     = ""
+}
+
+variable "attach_custom_emr_isntance_profile_policy" {
+  description = "boolean flag to use the above policy arn"
+  type       = bool
+  default    = false
+}
+
 locals {
   tags = merge(tomap({"vendor" = "segment"}), var.tags)
 }