mixed_instances_policy is not working

[shaibs3]

Description

i am trying to create a self managed node group with the eks module latest version 18.2.2
i am using mixed_instances_policy to create
mixed_instances_policy = { instances_distribution = { on_demand_base_capacity = "1" on_demand_percentage_above_base_capacity = "100" on_demand_allocation_strategy = "prioritized" } override = [ { instance_type = "m5.2xlarge" }, { instance_type = "c5.2xlarge" }, ] }

along with
use_mixed_instances_policy = true
unfortunatly the plan does not show that the correct policy was created.

looking at the code in node_groups.tf in module "self_managed_node_group"
i see that the variable mixed_instances_policy is not passed to the module so that why it cant work.
if we added a line
mixed_instances_policy = try(each.value.mixed_instances_policy, var.self_managed_node_group_defaults.mixed_instances_policy, null)

to the node_groups.tf this should fix the issue

Versions

• Terraform: 1.1.4

• Provider(s):

• Module: eks 18.2.2

Reproduction

Expected behavior

Actual behavior

Terminal Output Screenshot(s)

Additional context

[daroga0002]

please past a terraform module configuration which you are using

[nickvanwegen]

I seem to have the same issue.

Following terraform config

module "eks" {
  source          = "terraform-aws-modules/eks/aws"
  version         = "18.2.1"
  cluster_version = local.cluster_version
  cluster_name    = local.name
  vpc_id          = var.vpc_id
  subnet_ids      = var.subnets

  iam_role_name = "test"

  enable_irsa = true

  self_managed_node_group_defaults = local.self_managed_node_group_defaults
  self_managed_node_groups         = local.self_managed_node_group

this is a copy from the given example in github repo

  self_managed_node_group = {
    one = {
      name = "spot-1"

      public_ip    = true
      max_size     = 5
      desired_size = 2

      use_mixed_instances_policy = true
      mixed_instances_policy = {
        instances_distribution = {
          on_demand_base_capacity                  = 0
          on_demand_percentage_above_base_capacity = 10
          spot_allocation_strategy                 = "capacity-optimized"
        }

        override = [
          {
            instance_type     = "m5.large"
            weighted_capacity = "1"
          },
          {
            instance_type     = "m6i.large"
            weighted_capacity = "2"
          },
        ]
      }

      pre_bootstrap_user_data = <<-EOT
      echo "foo"
      export FOO=bar
      EOT

      bootstrap_extra_args = "--kubelet-extra-args '--node-labels=node.kubernetes.io/lifecycle=spot'"

      post_bootstrap_user_data = <<-EOT
      cd /tmp
      sudo yum install -y https://s3.amazonaws.com/ec2-downloads-windows/SSMAgent/latest/linux_amd64/amazon-ssm-agent.rpm
      sudo systemctl enable amazon-ssm-agent
      sudo systemctl start amazon-ssm-agent
      EOT
    }
  }

output seems to show that input should be fine according to examples.

self_managed_node_groups = {
  "one" = {
    "bootstrap_extra_args" = "--kubelet-extra-args '--node-labels=node.kubernetes.io/lifecycle=spot'"
    "desired_size" = 2
    "max_size" = 5
    "mixed_instances_policy" = {
      "instances_distribution" = {
        "on_demand_base_capacity" = 0
        "on_demand_percentage_above_base_capacity" = 10
        "spot_allocation_strategy" = "capacity-optimized"
      }
      "override" = [
        {
          "instance_type" = "m5.large"
          "weighted_capacity" = "1"
        },
        {
          "instance_type" = "m6i.large"
          "weighted_capacity" = "2"
        },
      ]
    }
    "name" = "spot-1"
    "post_bootstrap_user_data" = <<-EOT
    cd /tmp
    sudo yum install -y https://s3.amazonaws.com/ec2-downloads-windows/SSMAgent/latest/linux_amd64/amazon-ssm-agent.rpm
    sudo systemctl enable amazon-ssm-agent
    sudo systemctl start amazon-ssm-agent
    
    EOT
    "pre_bootstrap_user_data" = <<-EOT
    echo "foo"
    export FOO=bar
    
    EOT
    "public_ip" = true
    "use_mixed_instances_policy" = true
  }
}
self_managed_node_groups_defaults = {
  "ami_id" = "ami-02b3f04ab50ffd9f1"
  "block_device_mappings" = {
    "xvda" = {
      "device_name" = "/dev/xvda"
      "ebs" = {
        "encrypted" = true
        "iops" = "3000"
        "volume_size" = 25
        "volume_type" = "gp3"
      }
    }
  }
  "bootstrap_extra_args" = "--kubelet-extra-args '--node-labels=node.kubernetes.io/lifecycle=spot'"
  "cluster_name" = "poc-kajuuc0e"
  "create_launch_template" = true
  "desired_size" = "1"
  "force_delete" = true
  "iam_role_additional_policies" = [
    "arn:aws:iam::aws:policy/AmazonSSMManagedInstanceCore",
  ]
  "launch_template_name" = "poc-kajuuc0e"
  "launch_template_use_name_prefix" = true
  "max_size" = "5"
  "min_size" = "1"
  "name" = "poc"
  "pre_bootstrap_user_data" = <<-EOT
  export CONTAINER_RUNTIME="containerd"
  export USE_MAX_PODS=false
  
  EOT
  "propagate_tags" = [
    {
      "key" = "k8s.io/cluster-autoscaler/enabled"
      "propagate_at_launch" = "true"
      "value" = "true"
    },
    {
      "key" = "k8s.io/cluster-autoscaler/poc-kajuuc0e"
      "propagate_at_launch" = "true"
      "value" = "poc-kajuuc0e"
    },
  ]
  "protect_from_scale_in" = true
  "update_default_version" = true
  "use_mixed_instances_policy" = true
  "use_name_prefix" = true
}

Tho deployed in AWS i see the following ASG.
As you can see there is 0% spot
And also the spot_allocation_strategy is not set.

I could be that i am missing something? not sure but it sure seems like @shaibs3 has a point

[bryantbiggs]

yes @shaibs3 is correct - I've opened a PR to fix and am just going to double check the rest of the variables between the root module and the sub-module. thanks for reporting @shaibs3 !

[bryantbiggs]

ok looks good with the fix in

[antonbabenko]

This issue has been resolved in version 18.2.3 🎉

[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.