Error: cannot create mws credentials: invalid Databricks Account configuration

[MYB24]

Good Evening, 

I am configuring databricks_mws_credentials through Terraform on AWS.  I am getting the following error:

Error: cannot create mws credentials: invalid Databricks Account configuration
│
│ with module.databricks.databricks_mws_credentials.this,
│ on modules/aws-databricks-tf/main.tf line 128, in resource "databricks_mws_credentials" "this":
│ 128: resource "databricks_mws_credentials" "this" {
│

I have checked my account username, password and account-ID which are all correct. But i am getting the same error. Anything i am missing 

Below is my code: 

resource "databricks_mws_credentials" "this" {

#provider = databricks.mws

account_id = var.databricks_account_id

# role_arn = aws_iam_role.cross_account_role.arn

role_arn = var.databricks_role_arn

credentials_name = var.databricks_credentials

# depends_on = [aws_iam_role_policy.this]

}

module "databricks" {

source = "./modules/aws-databricks-tf"

region = var.region

databricks_vpc_id = element(data.aws_vpcs.vpc_list.ids, 0)

databricks_subnet_ids = data.aws_subnets.subnet_ids_private.ids # Private Subnet 1/2

databricks_security_group_ids = [aws_security_group.databricksNode-sg.id]

env = var.env

databricks_account_username = "my-username"

databricks_account_password = "mypassword"

databricks_account_id = "account-id"

databricks_root_storage_bucket = "root-bucket"

# cross_account_iam_role = "DatabricksRole"

databricks_role_arn = my-databricks-role # Instead of creating the Databricks Role via terraform, I have created it manually in the UI with relevant policies and trust relationship

databricks_credentials = "DatabricksCredentials"

#aws #terraform #Credentials #error

[Debayan]

Hi, Looks like formatting error , could you please follow the example code in here and let us know if this helps: https://registry.terraform.io/providers/databricks/databricks/latest/docs/resources/mws_credentials

[MYB24]

Hi @Debayan, this code was working before. i had deleted the mws credential file from my cloudformation stack and wnated to create a new one. when i tried to create a new mws credential and link it to a new workspace it would give me the above error. I have provided the whole file below which create the mws credential and workspace

resource "aws_kms_key" "this" {
enable_key_rotation = true
}

resource "aws_s3_bucket" "root_storage_bucket" {
bucket = var.databricks_root_storage_bucket
force_destroy = true
}

resource "aws_s3_bucket_server_side_encryption_configuration" "s3_bucket_encryption" {
bucket = var.databricks_root_storage_bucket

rule {
apply_server_side_encryption_by_default {
kms_master_key_id = aws_kms_key.this.arn
sse_algorithm = "aws:kms"
}
}
}

resource "aws_s3_bucket_versioning" "root_bucket" {
bucket = var.databricks_root_storage_bucket

versioning_configuration {
status = "Enabled"
}
}

resource "aws_s3_bucket_ownership_controls" "root_boot_ownership" {
bucket = var.databricks_root_storage_bucket
rule {
object_ownership = "BucketOwnerPreferred"
}
}

resource "aws_s3_bucket_acl" "root_bucket_acl" {
depends_on = [aws_s3_bucket_ownership_controls.root_boot_ownership]

bucket = var.databricks_root_storage_bucket
acl = "private"
}

resource "aws_s3_bucket_logging" "root_bucket_logging" {
bucket = var.databricks_root_storage_bucket

target_bucket = var.logging_target_bucket
target_prefix = var.logging_target_prefix
}

resource "aws_s3_bucket_public_access_block" "root_storage_bucket" {
bucket = aws_s3_bucket.root_storage_bucket.id
block_public_acls = true
block_public_policy = true
ignore_public_acls = true
restrict_public_buckets = true
depends_on = [aws_s3_bucket.root_storage_bucket]
}

data "databricks_aws_bucket_policy" "this" {
bucket = aws_s3_bucket.root_storage_bucket.bucket
}

resource "aws_s3_bucket_policy" "root" {
bucket = aws_s3_bucket.root_storage_bucket.id
policy = <<POLICY
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Principal": {
"AWS": "arn:aws:iam::account-id:root"
},
"Action": [
"s3:GetObject",
"s3:GetObjectVersion",
"s3:PutObject",
"s3:DeleteObject",
"s3:ListBucket",
"s3:GetBucketLocation"
],
"Resource": [
"arn:aws:s3:::${var.env}databricks/*",
"arn:aws:s3:::${var.env}atabricks"
]
}
]
}
POLICY
}

resource "databricks_mws_networks" "this" {
provider = databricks.mws
account_id = var.databricks_account_id
network_name = var.databricks_network_name
vpc_id = var.databricks_vpc_id
subnet_ids = flatten(var.databricks_subnet_ids)
security_group_ids = var.databricks_security_group_ids
}

resource "databricks_mws_storage_configurations" "this" {
provider = databricks.mws
account_id = var.databricks_account_id
bucket_name = aws_s3_bucket.root_storage_bucket.bucket
storage_configuration_name = var.databricks_root_storage_bucket
}

resource "databricks_mws_credentials" "this" {
#provider = databricks.mws
account_id = var.databricks_account_id
# role_arn = aws_iam_role.cross_account_role.arn
role_arn = var.databricks_role_arn
credentials_name = var.databricks_credentials
# depends_on = [aws_iam_role_policy.this]
}

# resource "databricks_mws_workspaces" "this" {
# provider = databricks.mws
# account_id = var.databricks_account_id
# aws_region = var.region
# workspace_name = var.workspace_name
# # deployment_name = var.workspace_name

# credentials_id = databricks_mws_credentials.this.credentials_id
# storage_configuration_id = databricks_mws_storage_configurations.this.storage_configuration_id
# network_id = databricks_mws_networks.this.network_id
# }

[MYB24 · Accepted solution]

Managed to fix the issue by updating the provider.tf while. Had to create a Service Principle token and add that into my provider.tf file. 

provider "databricks" {

alias = "accounts"

host = "https://accounts.cloud.databricks.com"

client_id = "service-principle-id"

client_secret = "service-principle-secret"

account_id = "databricks-account-id"

}
reference: https://registry.terraform.io/providers/databricks/databricks/latest/docs#special-configurations-for...

[TMD]

just to add a context for probably the underlying issue requiring an account level service principal (with OAuth).

I experienced the same issue while using username and password as in the case how TF provider was configured for existing workspaces created prior to 11/2023.  Looks like Databricks is expecting/enforcing an account level service principal for TF provider for new workspaces after 11/23.

[Alexandre467]

Hello, I'm facing a similaire Issue. I try to update my TF with properly authentification and I have this error ?! 

╷
│ Error: cannot create mws credentials: failed visitor: context canceled
│ 
│   with databricks_mws_credentials.this,
│   on main.tf line 8, in resource "databricks_mws_credentials" "this":

How you add context and what it is ?

[TMD]

Hello  Alexandre467,

Not sure what your issue is.  The reference to "context" in my earlier reply was referring to the situation as described in the second paragraph. 

If you add a little bit of details as to what you have configured and the issue, then I can comment on:

FYI, here're the workspace and account level provider configs as examples. Please pay attention to the Databricks Terraform provider documentation as to which one you need for the resource you are trying to deal with:

provider "databricks" {
  # Accunt Level resources such as workspaces
  alias         = "mws"
  host          = var.databricks_host     // https://accounts.cloud.databricks.com
  account_id    = var.databricks_account_id
  client_id     = var.databricks_tfe_client_id
  client_secret = var.databricks_tfe_client_secret
}

provider "databricks" {
  # workspace level but using account level credential
  alias         = "databricks-ws"
  host          = var.databricks_domain_for_workspace    // https://xyz-ws_name.cloud.databricks.com
  client_id     = var.databricks_tfe_client_id
  client_secret = var.databricks_tfe_client_secret
}