managed_resource_group_name not working for databricks

[mevenks]

Hi ,

I am using terraform to create a databricks . If I do NOT include the [managed_resource_group_name] parameter , databricks is created successfully in my resource group and a new managed resource groupe is created automatically , until now everything is normal .

But if I include the [managed_resource_group_name] parameter adding a personalised freshly new created resource group , databricks get created but not the resources that are supposed to be created under the managed resource group (storage,vnet and workers-sg) .

Terraform v1.1.5
on windows_amd64

provider registry.terraform.io/hashicorp/azurerm v2.99.0
My naming convention :
databricks resource gp naming convension : xx-xx-rg-project
databricks managed resource gp naming convension : xx-xx-rg-project-databricks-workspace

The same main.tf is used to created resource groupe for databricks and databrick's workspace , i've added depends_on under databricks creation to make sure both resources groupes are created before databricks creation.

main.tf

module "databricks" {

source = "./modules/databricks"

name = var.databricks_name
resource_group_name = var.resource_group_name
location = var.resource_group_location
sku = var.databricks_sku
managed_resource_group_name = var.managed_resource_group_name

depends_on = [
module.resource_group_workspace , module.resource_group
]

}

my module un modules\databricks\main.tf :

resource "azurerm_databricks_workspace" "this" {

name = var.name
resource_group_name = var.resource_group_name
location = var.location
sku = var.sku

managed_resource_group_name = var.managed_resource_group_name

}

Terraform does not display any error but only success on creation.
Message I have on databricks : The workspace 'xx-xx-xxxxx' is in a failed state and hence cannot be launched. Please delete and re-create the workspace.

thank you for your help

[mycloud91]

I'm also facing the same issue. any one have update/solution on this. version for reference.

azurerm = {
  source  = "hashicorp/azurerm"
  version = "=2.64.0"
}
databricks = {
  source  = "databrickslabs/databricks"
  version = "0.5.1"
}

[mevenks]

Hi,

My working version :

terraform {
required_providers {
azurerm = {
source = "hashicorp/azurerm"
version = "~>3.2.0"
}

databricks = {
  source = "databrickslabs/databricks"
  version = "0.5.7"
}

}

backend "local" {
path = "./terraform.tfstate"
}

}

my naming convention :

variable "RESOURCE_GROUP_NAME" {
type = string
default = "XXX.POC.XX.XXXX.XXXX.01"
}
variable "MANAGED_RESOURCE_GROUP_NAME" {
type = string
default = "XXX.POC.XX.XXXX.XXXXXX.XXXXXX.01"
}

my module:

resource "azurerm_databricks_workspace" "this" {

name = var.name
resource_group_name = var.resource_group_name
location = var.location
sku = var.sku
managed_resource_group_name = var.managed_resource_group_name
public_network_access_enabled = true

custom_parameters {
no_public_ip = true
virtual_network_id = var.virtual_network_id
private_subnet_name = var.private_subnet_name
private_subnet_network_security_group_association_id = var.private_nsg
public_subnet_name = var.public_subnet_name
public_subnet_network_security_group_association_id = var.public_nsg
}

tags = {
Environnement = "POC"
Support = "Architecture Technique SI"
}

}

[mycloud91]

@mevenks Thanks for your update! I have resolve the issue. (cause: used existing managed group name for newly created data bricks resource). I hope different managed group name is required for each data bricks if incase we use same virtual network.

[caldempsey]

experienced this issue. every time terraform wants to make a change to the databricks workspace by deleting the workspace it errors out complaining it doesn't know what to do with the managed resource group. @mycloud91 did you have a solution that lets you define a managed resource group for the databricks cluster (rather than it creating its own)?