Instance Refresh not triggering on launch_template changes

[Patricol]

Description

With the new instance refresh functionality, changes to the launch template will not trigger a refresh, because the autoscaling_group's launch_template block does not actually change if its version = $Latest, as is the case by default.

This is unexpected based on the comment here: A refresh will always be triggered by a change in any of launch_configuration, launch_template, or mixed_instances_policy.

I tested by changing ami_id multiple times; which never triggered an instance refresh.

All of the terraform resources appear to be created correctly; and I'm on the newest AWS provider (3.40.0).

I think this was just a misunderstanding of the intended behavior of aws_autoscaling_group's instance_refresh block.

I suppose this module could handle bumping an explicitly set launch_template version.

Versions

• Terraform: v0.14.11

• Provider(s):

+ provider registry.terraform.io/hashicorp/aws v3.40.0
+ provider registry.terraform.io/hashicorp/cloudinit v2.2.0
+ provider registry.terraform.io/hashicorp/external v2.1.0
+ provider registry.terraform.io/hashicorp/kubernetes v2.2.0
+ 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
+ provider registry.terraform.io/terraform-aws-modules/http v2.3.0

• Module: 16.0.0

Reproduction

Skipping most repro descriptions because the module does correctly setup all the resources; they just don't behave as documented.

[barryib]

@Patricol Thanks for opening this. @bashims can you please have a look ? We probably missed something in your PR.

[barryib]

@Patricol did you tried the instace refresh's example ? Is that example works for you ? Can you please share your configuration ?

[bashims]

@barryib absolutely.

@Patricol can you try changing/adding a tag to the example ASG config. That should trigger the refresh.

In the meantime I will take a closer look to see what's going on.

[bashims]

@Patricol as a work around please add the following to your worker_groups_launch_template:

      instance_refresh_triggers            = ["tag"]
      tags = [
        {
          key                 = "test_trigger_refresh"
          value               = "true"
          propagate_at_launch = true
        },
     ]

[bashims]

@barryib I'm working on a fix now. We need to bump/change the launch_template version to trigger the instance_refresh. Triggering from tags works fine however.

[Patricol]

@bashims That does work for me.

[stevehipwell]

@bashims is there a reason why the node termination handler policy is added to the worker node role in the example?

[barryib]

@bashims is there a reason why the node termination handler policy is added to the worker node role in the example?

Nope. It should be attached to the node termination handler through IRSA.

@stevehipwell can you please open a PR to fix that ?

[stevehipwell]

@barryib that's what I thought. Would you like me to also add the the spot instance termination and rebalance suggestion events to the example queue from the aws-node-termination-handler repo docs ?

[barryib]

@barryib that's what I thought. Would you like me to also add the the spot instance termination and rebalance suggestion events to the example queue from the aws-node-termination-handler repo docs ?

Yes you can if it's simple. Please add a link to the NTH docs also. People must follow that docs. The example is there for help only and not to address every NTH use cases.

[stevehipwell]

@barryib #1373.

[barryib]

@barryib #1373.

Thanks. Merged.

[bashims]

@bashims is there a reason why the node termination handler policy is added to the worker node role in the example?

@stevehipwell I believe it is uses IRSA so there is no need for attaching it to the worker node role. This was probably a leftover i.e. before I switched to using IRSA. Thanks for removing the attachment.

[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.