AzureRM's minimum version is >= 3.51, < 4.0
now.
var.sku_tier
cannot be set to Paid
anymore, now possible values are Free
and Standard
.
Related issue: #335
Two new resources would be created when upgrading from v6.x to v7.x:
null_resource.kubernetes_version_keeper
azapi_update_resource.aks_cluster_post_create
azurerm_kubernetes_cluster.main
resource would ignore change on kubernetes_version
from outside of Terraform in case AKS cluster's patch version has been upgraded automatically.
When you change var.kubernetes_version
's value, it would trigger a re-creation of null_resource.kubernetes_version_keeper
and re-creation of azapi_update_resource.aks_cluster_post_create
, which would upgrade the AKS cluster's kubernetes_version
.
azapi
provider is required to be configured in your Terraform configuration.
In v6.0, default_node_pool.linux_os_config
block won't be added to azurerm_kubernetes_cluster.main
resource when var.enable_auto_scaling
is true
. This bug has been fixed in v7.0.0 so you might see a diff on azurerm_kubernetes_cluster.main
resource.
var.log_analytics_solution_id
is now an object with id
attribute. This change is to fix #263.
var.net_profile_docker_bridge_cidr
has been deprecated and is not used in the module anymore and has been removed.
Now azurerm_kubernetes_cluster_node_pool.node_pool
resource has create_before_destroy=true
to avoid downtime when upgrading node pools. Users must be aware that there would be a "random" suffix added into pool's name, this suffix's length is 4
, so your previous node pool's name nodepool1
would be nodepool1xxxx
. This suffix is calculated from node pool's config, the same configuration would lead to the same suffix. You might need to shorten your node pool's name because of this new added suffix.
To enable this feature, we've also added new null_resource.pool_name_keeper
to track node pool's name in case you've changed the name.
As the document described:
When
public_network_access_enabled
is set to true,0.0.0.0/32
must be added toauthorized_ip_ranges
in theapi_server_access_profile block
.
We'll add api_server_access_profile
nested block after AzureRM provider's v4.0, but starting from v7.0 we'll enforce such pre-condition check.
If you have azurerm_kubernetes_cluster_node_pool
resources not managed with this module (var.nodepools
) you
must have an explicit depends_on
on those resources to avoid conflicting nodepools operations.
See issue #418 for more details.