[ISSUE] Issue with databricks_metastore_assignment resource

[francisco-ltech]

Configuration

resource "databricks_metastore_assignment" "assignment" {
  metastore_id         = var.metastore-id
  workspace_id         = azurerm_databricks_workspace.build.workspace_id
  default_catalog_name = "default"
}

Expected Behavior

Databricks workspace to be assigned to the given (existing) metastore

Actual Behavior

Error: cannot read metastore assignment: No metastore assigned for the current workspace.

Steps to Reproduce

Metastore already exists, new Databricks workspace is created, then assigned.

1. terraform plan
2. terraform apply

Terraform and provider versions

azurerm = {
source = "hashicorp/azurerm"
version = "3.45.0"
}
databricks = {
source = "databricks/databricks"
version = "1.11.1"
}

Debug Output

[DEBUG] provider.terraform-provider-databricks_v1.11.1: PUT /api/2.1/unity-catalog/workspaces/{id}/metastore

{
"default_catalog_name": "default",
"metastore_id": "uuid-value",
"workspace_id": int-value
}
< HTTP/2.0 200 OK
< {}: timestamp=2023-03-02T14:21:19.338Z
[DEBUG] provider.terraform-provider-databricks_v1.11.1: GET /api/2.1/unity-catalog/metastore_summary
< HTTP/2.0 404 Not Found
< {
< "details": [
< {
< "@type": "type.googleapis.com/google.rpc.RequestInfo",
< "request_id": "17e-8bf-4da-b23-b12",
< "serving_data": ""
< }
< ],
< "error_code": "METASTORE_DOES_NOT_EXIST",
< "message": "No metastore assigned for the current workspace."

Important Factoids

• The actual PUT request to the assignment reads a OK which indicates success, which indeed succeeds because I can see the metastore assigned to the workspace in the accounts.databricks portal.
• However, there seems to be a subsequent GET request that fails and I believe this makes TF not to update the state?
• If I terraform plan again, the plan indicates the assignment is yet to be created (again).

[nkvuong]

this is a known issue with the databricks_metastore_assignment resource if you're running it on a workspace that is not Unity Catalog enabled.

Engineering team is building support for metastore APIs at account-level, which should resolve this issue (and other similar ones)

[nkvuong]

fixed in #2182