azurerm_kubernetes_cluster: provider wants to destroy/recreate cluster due to case mismatch in network_plugin_mode

[jeffhuenemann]

Is there an existing issue for this?

• I have searched the existing issues

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment and review the contribution guide to help.

Terraform Version

1.5.3

AzureRM Provider Version

3.65.0

Affected Resource(s)/Data Source(s)

azurerm_kubernetes_cluster

Terraform Configuration Files

resource "azurerm_kubernetes_cluster" "aks1" {
# ...
  network_profile {
    network_plugin      = "azure"
    network_plugin_mode = "Overlay"
  }
}

Debug Output/Panic Output

# module.kubernetes_cluster_standard.azurerm_kubernetes_cluster.aks1 must be replaced
-/+ resource "azurerm_kubernetes_cluster" "aks1" {
#[snip]
      ~ network_profile {
#[snip]
          + network_mode        = (known after apply)
          ~ network_plugin_mode = "overlay" -> "Overlay" # forces replacement
        }
    }

Expected Behaviour

Per the azurerm_kubernetes_cluster docs, the acceptable value of network_plugin_mode is Overlay (uppercase O). With that value in play, subsequent plans/applies without any code change should not represent a mismatch of value, and should not suggest recreating the cluster.

Actual Behaviour

In the state file, the state is represented with a lowercase o, creating a mismatch with uppercase O required in the configuration. terraform plan wants to replace the entire cluster due to network_plugin_mode changing (it thinks) from overlay in state to Overlay in configuration.

Steps to Reproduce

1. Create an azurerm_kubernetes_cluster using network_plugin = azure and network_plugin_mode = Overlay
2. Subsequent terraform plan suggests to recreate the cluster on account of drift innetwork_plugin_mode

Important Factoids

Did not observe this behavior until upgrading from 3.63 to 3.65

References

No response

[jeffhuenemann]

Improved search on my part reveals #22151 -> #22153 for an update to the API version in use, so maybe this is really just a documentation change and overlay is the only appropriate input value? If the provider accepts both overlay and Overlay as inputs, then they both need to be non-impactful on subsequent plans/applies, to my thinking.

[stephybun]

Thanks for raising this issue @jeffhuenemann.

The resource will allow the value to be lowercased to overlay without replacing the cluster - this is unfortunately an update that needs to be done in the user's config since this change was shipped by the upstream API.

The latest release v3.66.0 shipped a documentation fix to the casing of the value in #22554.

Since this is working as expected and the documentation has been fixed I'm going to close this issue.

[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.