For a production environment (ADF as orchestrator, ADLS Gen2 as storage, Databricks for PySpark transformations), follow Microsoft-recommended best practices:

• Databricks → ADLS Gen2: Use Unity Catalog with Azure Managed Identity (via Access Connector) for direct, secure access without secrets or mounts. Avoid mounting in production (it's legacy and less secure/governable). If not using Unity Catalog yet, fall back to Service Principal + OAuth with secrets from Azure Key Vault.
• ADF → Databricks: Create a Databricks linked service using a Personal Access Token (PAT) stored in Azure Key Vault.
• ADF → ADLS Gen2: Use System-assigned Managed Identity or Service Principal (secrets in Key Vault).
• Key Vault: Yes, route secrets through Key Vault – it's essential for production security (never hardcode credentials).

Below are detailed, step-by-step instructions for a fully secure setup.

1. Prerequisites

• Azure subscription with Contributor/Owner access.
• Create an Azure Key Vault.
• Enable Unity Catalog on your Databricks workspace (recommended for production governance). If not possible yet, see the Service Principal fallback in section 2.

2. Databricks to ADLS Gen2 Access (Recommended: Unity Catalog + Managed Identity)

This is the modern, secretless approach (no Key Vault needed for storage access).

Step 1: Create an Azure Databricks Access Connector

• In Azure Portal → Search for "Databricks Access Connector" → Create.
• Note the Resource ID (e.g.,

/subscriptions/.../resourceGroups/.../providers/Microsoft.Databricks/accessConnectors/my-connector​

Step 2: Grant the Access Connector permission on ADLS Gen2

• Go to your ADLS Gen2 storage account → Access Control (IAM) → Add role assignment.
• Role: Storage Blob Data Contributor (or finer-grained if needed).
• Assign to: The Access Connector (search by name or use its Managed Identity Application ID).

Step 3: In Databricks, create a Storage Credential (Unity Catalog)

• In Databricks workspace → Catalog → Add → Storage credential.
• Type: Managed identity.
• Paste the Access Connector's Resource ID.
• Test the connection.

Step 4: Create an External Location (points to ADLS containers)

• Catalog → Add → External location.
• Select the Storage Credential above.
• Path 

abfss://<container>@<storage-account>.dfs.core.windows.net/<optional-folder>​

• Grant READ/WRITE permissions to users/groups as needed.

• Step 5: In PySpark notebooks

  • No mounts or configs needed.
  • Read/write directly:

df = spark.read.parquet("abfss://<container>@<storage-account>.dfs.core.windows.net/path/to/data")
df.write.format("delta").save("abfss://<container>@<storage-account>.dfs.core.windows.net/output")​

• Unity Catalog enforces governance (auditing, access controls).

Fallback if no Unity Catalog: Service Principal + Key Vault

• Register an App in Microsoft Entra ID → Note Client ID, Tenant ID, generate Client Secret.
• Grant the Service Principal Storage Blob Data Contributor on ADLS Gen2.
• Store Client ID, Secret, Tenant ID as secrets in Key Vault.
• In Databricks: Create Key Vault-backed secret scope (URL: https://#secrets/createScope in workspace).
• In notebooks, set Spark configs (no mount needed for production):

spark.conf.set("fs.azure.account.auth.type.<storage-account>.dfs.core.windows.net", "OAuth")
spark.conf.set("fs.azure.account.oauth.provider.type.<storage-account>.dfs.core.windows.net", "org.apache.hadoop.fs.azurebfs.oauth2.ClientCredsTokenProvider")
spark.conf.set("fs.azure.account.oauth2.client.id.<storage-account>.dfs.core.windows.net", dbutils.secrets.get(scope="<scope>", key="client-id"))
spark.conf.set("fs.azure.account.oauth2.client.secret.<storage-account>.dfs.core.windows.net", dbutils.secrets.get(scope="<scope>", key="client-secret"))
spark.conf.set("fs.azure.account.oauth2.client.endpoint.<storage-account>.dfs.core.windows.net", "https://login.microsoftonline.com/<tenant-id>/oauth2/token")​

3. ADF to Databricks Linked Service (Secure with Key Vault)

Step 1: Generate Databricks Personal Access Token (PAT)

• In Databricks → User Settings → Developer → Access tokens → Generate new token (no expiration for production).

Step 2: Store PAT in Key Vault

• Key Vault → Secrets → Generate/Import → Name: e.g., databricks-pat.

Step 3: Grant ADF access to Key Vault

• Enable System-assigned Managed Identity on your ADF (Properties tab).
• Key Vault → Access policies → Add → Principal: Your ADF's Managed Identity → Permissions: Get (secrets).

Step 4: Create Key Vault Linked Service in ADF

• ADF → Manage → Linked services → New → Azure Key Vault → Select your Key Vault.

Step 5: Create Databricks Linked Service

• Linked services → New → Azure Databricks.
• Workspace URL: e.g., https://adb-xxx.azuredatabricks.net
• Authentication: Access token.
• For the token: Select "Azure Key Vault" → Choose the Key Vault linked service → Secret name: databricks-pat.
• Cluster: Use an existing Job cluster or new (for production, use Job clusters or serverless).

4. ADF to ADLS Gen2 Linked Service (Secure)

• Linked services → New → Azure Data Lake Storage Gen2.
• Authentication: System-assigned Managed Identity (recommended, secretless) or Service Principal (store ID/Secret in Key Vault as above).
• Test connection.

5. Orchestrate with ADF Pipeline

• Create a pipeline.
• Add Databricks Notebook activity.
• Linked service: The one from step 3.
• Notebook path: e.g., /Users/yourname/my-notebook.
• Pass parameters if needed (e.g., file paths in ADLS).
• For input/output: Use ADLS linked datasets (abfss:// paths).
• Trigger: Schedule, Tumbling window, or Event-based (on new files in ADLS).

Production Tips

• Use Job clusters (not interactive) for cost/reliability.
• Enable ADF monitoring, alerts, and Git integration.
• Rotate secrets/PATs regularly.
• Network security: VNet-integrate Databricks and use Private Endpoints for ADLS/Key Vault if needed.
• If using Delta Lake tables on ADLS, register them in Unity Catalog for governance.