Open zioproto opened 2 years ago
@lonegunmanb this is a trivial patch, but it is also a change of default behaviour that triggers the cluster to be destroyed and recreated. It is very dangerous. I believe we should target this for 7.0.0 with a very visible release note. What is your opinion on this ?
Thanks to @paolosalvatori that noticed this discrepancy in default behaviour between the AKS product and the Terraform Module.
Thanks @zioproto for opening this issue. According to azurerm_kubernetes_cluster
's document, the os_disk_type
's default value is Managed
in Terraform AzureRM Provider, so I think it's ok to keep it in this module. As the downstream of Terraform Provider I think it's better to keep consist with Terraform Provider, because the Provider has low tolerance of breaking change, whilst API and portal are tend to change more frequently.
Is there an existing issue for this?
Greenfield/Brownfield provisioning
greenfield
Terraform Version
1.2.8
Module Version
6.0.0
AzureRM Provider Version
3.0.0
Affected Resource(s)/Data Source(s)
azurerm_kubernetes_cluster
Terraform Configuration Files
tfvars variables values
Debug Output/Panic Output
Expected Behaviour
The default value for
os_disk_type
should beEphemeral
.https://learn.microsoft.com/en-us/azure/aks/cluster-configuration#ephemeral-os
Actual Behaviour
The current default value is
Managed
https://github.com/Azure/terraform-azurerm-aks/blob/master/variables.tf#L363Steps to Reproduce
No response
Important Factoids
n/a
References
It is all about this variable https://github.com/Azure/terraform-azurerm-aks/blob/master/variables.tf#L360