Reputation: 346
using for each with terraform version upgrade from 0.12 to 0.13
I am working on version upgrade of terraform from 0.12 to 0.13. So far I am able to create the necessary pieces by changing providers as well as my modules.
Current Status of service.tf
module "services" {
source = "../../modules/aws-ecs-servicev2"
for_each = var.applications
container_name = each.value.container_name
container_cpu = each.value.container_cpu
container_image = each.value.container_image
container_memory_total = each.value.container_memory_total
container_memory_reservation = each.value.container_memory_reservation
deployment_max_percent = each.value.deployment_max_percent
deployment_min_healthy = each.value.deployment_min_healthy
container_host_port = each.value.container_host_port
container_port = each.value.container_port
command = each.value.command
replica_count = each.value.replica_count
stage = each.value.stage
alb_enabled = each.value.alb_enabled
alb_subnets = module.vpc.public_subnets
alb_https_enabled = each.value.alb_https_enabled
alb_ssl_arn = each.value.alb_ssl_arn
health_check_path = each.value.health_check_path
internal_lb = each.value.health_check_path
vault_key = each.value.vault_key
vault_backend = each.value.vault_backend
cpu_based_autoscaling = each.value.cpu_based_autoscaling
target_cpu_value = each.value.target_cpu_value
max_capacity = each.value.max_capacity
min_capacity = each.value.min_capacity
scale_out_adjustment = each.value.scale_out_adjustment
scale_out_cooldown = each.value.scale_out_cooldown
scale_in_adjustment = each.value.scale_in_adjustment
scale_in_cooldown = each.value.scale_in_cooldown
disable_scale_in = each.value.disable_scale_in
ecs_arn = module.ecs.arn
namespace = var.namespace
sd_namespace_id = aws_service_discovery_private_dns_namespace.default.id
vpc_id = module.vpc.vpc_id
health_check_response = "200"
alb_sgs = [aws_security_group.ecs_sg.id]
}
and application.tf
variable "applications" {
description = "Map for the services to be deployed"
type = any
default = {
service-a = {
container_cpu = 512
container_image = service-a-image:latest
container_memory_total = 2048
container_memory_reservation = 1024
container_name = "service-a"
command = null
deployment_max_percent = 200
deployment_min_healthy = 50
container_host_port = [0]
container_port = [80]
replica_count = 1
stage = "dev"
health_check_path = null
internal_lb = false
vault_key = null
vault_backend = null
alb_enabled = false
alb_ssl_arn = null
alb_https_enabled = false
cpu_based_autoscaling = false
target_cpu_value = null
max_capacity = null
min_capacity = null
scale_out_adjustment = null
scale_out_cooldown = null
scale_in_adjustment = null
scale_in_cooldown = null
disable_scale_in = false
}
}
}
What I want to achieve?
Is it possible to pass somehow default values to few of the variables such as health_check_path= null so that I do not have to worry about the not including this in services which do not require these values. There are also some services which does require such values which I will then provide from the same variables. Example
service-b = {
container_name = "service-b-container"
container_cpu = 1024
container_image = "service-b-image:latest"
container_memory_total = 1024
container_memory_reservation = 1024
deployment_max_percent = 200
deployment_min_healthy = 50
container_host_port = [0]
container_port = [80]
command = null
replica_count = 1
stage = "dev"
alb_enabled = true
alb_https_enabled = true
alb_ssl_arn = "ssl-arn"
health_check_response = 200
health_check_path = "/status"
internal_lb = false
vault_key = "servia-b"
vault_backend = "service"
cpu_based_autoscaling = true
target_cpu_value = 40
max_capacity = 8
min_capacity = 1
scale_out_adjustment = 2
scale_out_cooldown = 60
scale_in_adjustment = 1
scale_in_cooldown = 300
disable_scale_in = false
},
I tried using local or any other such variables inside the variable section like follwing
variable "cpu" {
type = number
default = 1024
description = "cpu to associate"
}
variable "applications" {
description = "Map for the services to be deployed"
type = any
default = {
service-b = {
container_name = "service-b-container"
container_cpu = var.cpu
container_image = "service-b-image:latest"
container_memory_total = 1024
container_memory_reservation = 1024
},
service-a = {
container_name = "service-a-container"
container_cpu = var.cpu
container_image = "service-b-image:latest"
container_memory_total = 1024
}
}
}
but turns out terraform does not allow that. To elaborate on this, let me add it's log here
Error: Variables not allowed
on application.tf line 56, in variable "applications":
56: container_cpu = var.cpu
Variables may not be used here.
If I try by not including the variables which are present in my services.tf then I get the followiing error
Acquiring state lock. This may take a few moments...
Refreshing Terraform state in-memory prior to plan...
The refreshed state will be used to calculate this plan, but will not be
persisted to local or remote state storage.
module.services["service-a"].module.ecs_alb_service_task.data.aws_iam_policy_document.ecs_exec[0]: Refreshing state...
module.services["service-a"].module.ecs_alb_service_task.data.aws_iam_policy_document.ecs_service_policy[0]: Refreshing state...
module.services["service-a"].module.ecs_alb_service_task.data.aws_iam_policy_document.ecs_task_exec[0]: Refreshing state...
module.services["service-a"].module.ecs_alb_service_task.data.aws_iam_policy_document.ecs_task[0]: Refreshing state...
module.services["service-a"].module.ecs_alb_service_task.data.aws_iam_policy_document.ecs_service[0]: Refreshing state...
module.vpc.aws_vpc.this[0]: Refreshing state... [id=vpc-0641c1d3435df195a]
aws_service_discovery_private_dns_namespace.default: Refreshing state... [id=ns-vyggm2ltoxjufhfa]
Error: Unsupported attribute
on Service.tf line 22, in module "services":
22: alb_enabled = each.value.alb_enabled
|----------------
| each.value is object with 12 attributes
This object does not have an attribute named "alb_enabled".
Error: Unsupported attribute
on Service.tf line 25, in module "services":
25: alb_https_enabled = each.value.alb_https_enabled
|----------------
| each.value is object with 12 attributes
This object does not have an attribute named "alb_https_enabled".
Is there anyway I can achieve this, since there are more than 10 services which needs to be deployed in this way, also some services lack the variable while the other requires.
Upvotes: 0
Views: 547
Answers (1)
Reputation: 151
Option 1. Based on the documentation this could help - declaring some special object type and there you could set default values: https://www.terraform.io/docs/configuration/variables.html#declaring-an-input-variable
Option 2Also may be this could help (have not tried it) - using lookup with default value https://www.terraform.io/docs/configuration/functions/lookup.html
Simple example for option 2:
child module:
main.tf
resource "local_file" "simple_file" {
filename = "file/${var.file_name}"
content = "foo!"
}
variables.tf
variable "file_name" {
description = "file name"
}
root module: main.tf
module "file" {
source = "./child_module"
for_each = var.applications
file_name = lookup(each.value, "Not_existing_key", "default_value" )
}
default_value is used in this case in the child module
terraform version 0.14.3
Upvotes: 3
Related Questions
- Looping with for_each
- Terraform : for_each one by one
- Terraform for_each loop
- Terraform for_each argument is unsuitable error
- terraform Error: Invalid for_each argument error
- Terraform : Invalid dynamic for_each value => Cannot use a set of object value in for_each. An iterable collection is required
- terraform for_each index issue, invalid index
- Missing resource instance key when using for_each in terraform
- Terraform 0.13 - Modules, for_each and providers
- Terraform : The "each" object does not support this operation