Forum Discussion
Power BI & Mirrored Databricks Catalog — Table Access Forbidden errors despite Unity Catalog access
Hi RomanTesolkin,
Thank you for reaching out to Microsoft Fabric Community.
Since only specific tables are failing while others sync successfully, looks like this is not general fabric or unity catalog issue.
- Please check whether the affected tables are external tables pointing to a different ADLS location and also verify that the underlying storage permissions and network access are available to fabric.
- Also compare the failing tables with a working one so that we can confirm that the affected tables are of a supported type for mirroring.
If these checks are valid and the issue persists, I recommend you to raise a Microsoft support ticket with the logs for further investigation.
Create a Fabric and Power BI Support Ticket - Power BI | Microsoft Learn
Thanks and regards,
Anjan Kumar Chippa
Hi Anjan,
thanks for the inputs.
all tables set this way are failing, i can query them in databricks without issues. Also i can query them in Power BI over standard Azure Databricks connector, however over mirroring, i get insufficient permissions issue.
Hi RomanTesolkin,
Since only specific tables fail, looks like those tables most likely have row filters or column masks applied in unity catalog. Databricks blocks those tables with row-level security or column masks from being mirrored to external systems like fabric, so you get TableAccessForbidden even though the required permissions are correct.
Remove the row filter or column mask from those and then retry the sync.
Thanks and regards,
Anjan Kumar Chippa
hi Anjan, as mentioned above all tables fail over mirroring, however, i can query them in Power BI over standard Azure Databricks connector and Databricks directly, no access issues. There are no masks applied, on the tables, as i tried to test on dummy excel file as well. It's being blocked on catalogue level fro some reason.
Hi RomanTesolkin,
Here the mirroring reads directly from the catalog, not through the databricks compute like the standard connector, so some catalog types are not supported for mirroring even though they can be queried normally through other methods.
- Since your entire catalog fails, check how it was created in unity catalog. Mirroring does not support lakehouse federated tables and delta sharing tables, or other unsupported table types, please check if your catalog is one of those types. So to use mirroring, the data should be stored in a supported standard unity catalog-managed catalog.
Please refer below documentation:
https://learn.microsoft.com/en-us/fabric/mirroring/azure-databricks-limitations
Thanks and regards,
Anjan Kumar Chippa