Terraform-aws-eks: Node taints for Node Groups

[mattlawnz]

I have issues

I'm submitting a...

• [ ] bug report
• [ ] feature request
• [X] support request - read the FAQ first!
• [ ] kudos, thank you, warm fuzzy

What is the current behavior?

I need to be able to set taints on newly created node groups. These node groups will be highly varialbe in terms of sizing, so the taints are required on creation.

If this is a bug, how to reproduce? Please include a code sample if relevant.

I thought of adding kubelt_extra_args to the node group defination, but it didnt work, or perhaps thats not the right way to do it?

    az3-c = {
      desired_capacity = 1
      max_capacity     = 2
      min_capacity     = 1
      subnets          = ["subnet-004514be54bcd7eb7"]
      kubelet_extra_args = "--node-labels=function=data_only --register-with-taints=function=data_only:NoSchedule"
      instance_type = "m5a.4xlarge"
      additional_tags = {
        Name = module.eks_dev_label.id
      }
    }

What's the expected behavior?

Was hoping for a node taint to be applied.

Are you able to fix this problem and submit a PR? Link here if you have already.

Environment details

• Affected module version: "12.1.0"
• OS: Mac
• Terraform version: Terraform v0.12.26

Any other relevant info

[max-rocket-internet]

Maybe it helps you to see what I use?

  worker_groups_launch_template = [
    {
      name                    = "prometheus-spot-1"
      override_instance_types = ["m5.2xlarge", "r5.xlarge", "r5ad.xlarge"]
      spot_instance_pools     = 3
      asg_max_size            = 5
      additional_userdata     = module.bastion_vpc1.user_data_users
      kubelet_extra_args = join(" ", [
        "--node-labels=cluster_name=xxxxxxx,kubernetes.io/lifecycle=spot,worker_group=prometheus-spot-1,app=prometheus",
        "--register-with-taints app=prometheus:NoSchedule"
      ])
      enabled_metrics = local.enabled_metrics
      subnets         = aws_subnet.private_extra.*.id
      tags = [
        {
          "key"                 = "k8s.io/cluster-autoscaler/enabled"
          "propagate_at_launch" = "false"
          "value"               = "true"
        },
        {
          "key"                 = "k8s.io/cluster-autoscaler/xxxxx"
          "propagate_at_launch" = "false"
          "value"               = "true"
        }
      ]
    }
  ]

[dpiddockcmp]

Hi @mattlawnz, it looks like you're using the Managed Node Groups. These do not currently support node tainting. There is a request over on the AWS team's roadmap for this feature: https://github.com/aws/containers-roadmap/issues/864

If you really need to work with node taints then I suggest you use the traditional worker groups similar to Max's example above.

[mattlawnz]

Thanks all for the examples & explanation. I have this working as a worker group.