Changes to desired node count do not result in changes to autoscaling group
thecooldrop opened this issue · 2 comments
Summary
When desired_node_count input variable is changed this does not lead to change in the backing autoscaling group, thus making it impossible to increase the desired number of nodes deployed into cluster via GitOps.
Steps to reproduce the behavior
- Generate a clone of this repository for yourself
- In clone repository change the main.tf looks like:
module "eks-jx" {
source = "github.com/jenkins-x/terraform-aws-eks-jx?ref=v1.15.12"
cluster_name = var.cluster_name
cluster_version = var.cluster_version
region = var.region
vault_user = var.vault_user
is_jx2 = false
jx_git_url = var.jx_git_url
jx_bot_username = var.jx_bot_username
jx_bot_token = var.jx_bot_token
force_destroy = var.force_destroy
desired_node_count = 2
min_node_count = 1
max_node_count = 3
node_machine_type = "t3.medium"
}
-
Clone the repository localy
-
Execute
terraform init
andterraform apply
and approve -
Change the main.tf so that
desired_node_count=3
-
Execute
terraform apply
-
Observe that no changes are made to autoscaling group.
Expected behavior
Changes to desired_node_count
are propagated the autoscaling group used by the EKS cluster
Actual behavior
Changes to desired_node_count
do not result in changes to the autoscaling group.
Terraform version
Terraform v0.13.2
- provider registry.terraform.io/hashicorp/aws v3.37.0
- provider registry.terraform.io/hashicorp/helm v2.1.1
- provider registry.terraform.io/hashicorp/kubernetes v2.1.0
- provider registry.terraform.io/hashicorp/local v2.1.0
- provider registry.terraform.io/hashicorp/null v3.1.0
- provider registry.terraform.io/hashicorp/random v3.1.0
- provider registry.terraform.io/hashicorp/template v2.2.0
Module version
1.15.2
Operating system
Linux Ubuntu
@thecooldrop this is the expected behavior, that setting is ignored by terraform. See: https://github.com/terraform-aws-modules/terraform-aws-eks/blob/master/docs/autoscaling.md#notes. Internally the eks jx module uses that module. I wonder if we should just remove the desired count. Personally, I let cluster autoscaler to set the desired count. Also this: terraform-aws-modules/terraform-aws-eks#510 (comment)
closing this, as this is a desired behavior.