In Databricks, port 3306 is primarily used for communication with the Hive metastore, a MySQL database that stores metadata about Databricks objects (tables, schemas, etc.).

Security Groups and Network Configurations

When setting up Databricks within a virtual network (VNet), you’ll need to configure network security groups (NSGs) or firewall rules to allow traffic on port 3306:

• Inbound:
  • No inbound access is typically required on port 3306 from external sources, as the megastore is accessed internally by Databricks clusters.
• Outbound:
  • Allow outbound traffic on port 3306 from your Databricks clusters or workspaces to the meta store’s IP address or hostname.

Metastore Access

• Managed Metastore (Default): Databricks manages the megastore for you. You’ll need the meta store’s hostname, which you can find in your Databricks workspace settings or documentation.
• External Metastore: If you’re using an external metastore (e.g., your own MySQL database), you’ll need to configure connectivity on port 3306 to that database.

Additional Considerations

• Secure Cluster Connectivity (SCC): If you’re using SCC, you might need additional configurations to ensure secure communication with the megastore.
• Firewall Rules: If you have firewalls between your Databricks resources and the megastore, ensure they allow traffic on port 3306.
• Updates: Databricks is updating its control plane to use port 3306 for communication with the classic compute plane starting June 3, 2024. Ensure your network security groups do not block this port.