Forum Discussion
Scheduled Refresh Failure on Report Server (May 2025) - IDbCommand interface exception
Hi FLORENTINO_23,
Thank you for the update, and we are glad to hear that your refresh issue has been resolved.
Based on our understanding, while using SELECT * followed by “Remove Other Columns” will work, it is considered best practice to explicitly select only the required columns in your SQL query. This improves performance, reduces data transfer, ensures better query folding, and helps avoid unexpected schema changes. Your current approach is acceptable for small datasets. However, for production scenarios, defining columns at the source is recommended.
Should you have any further questions, please feel free to contact the Microsoft Fabric community.
Thank you.
Super UserHi - "This answer was generated with the assistance of AI and has been reviewed before posting."
This isn’t a known widespread bug specific to the May 2025 build of Power BI Report Server—your error pattern points much more toward a server-side execution failure, typically caused by timeouts, resource limits, or driver inconsistencies rather than a dataset logic issue. The key clue is “command has been canceled” cascading across IDbCommand and IDataReader, which usually happens when the server kills a long-running or blocked query, or when one step in a transaction fails and the rest get aborted. Since it works in Desktop, the difference is almost always environment-related: Report Server has stricter timeouts, lower memory ceilings, and depends on installed data providers (ODBC/OLE DB/SQL drivers) that may differ from your local machine. You should focus on (1) increasing CommandTimeout/ExecutionTimeout on the data source or server config, (2) checking SQL Server-side logs for query timeouts or deadlocks, (3) validating that the exact same driver version used in Desktop is installed on the Report Server, and (4) monitoring memory/CPU spikes during refresh—because Report Server will silently cancel queries under pressure. If the dataset involves multiple queries or merges, one failing step will cancel the entire transaction, which aligns exactly with your error chain.