Forum Discussion
Memory issue after data table doubles in size
Hi am_nr, Apologies for missing you last comment.
It’s possible the issue is coming from the SQL Server side. When Power BI refreshes, it pushes queries back to the database, so if SQL Server is already under memory pressure that can cause the refresh to fail. I’d recommend testing the same query in SSMS or to check for indexing or resource bottlenecks on the database.
thanks,
Prashanth
Hi am_nr
That error isn’t fixed by the “maximum MB” options.They don’t give VertiPaq more RAM. It means your model got too big to process in memory after the fact table doubled. Shrink the model: remove unused columns, move calculated columns to Power Query/SQL, keep only numeric keys in the fact (move long text to dimensions), reduce precision/high-cardinality fields (e.g., round decimals, split DateTime), and keep Auto date/time off. If you must keep all 60M rows, use Incremental Refresh and/or switch the big fact to DirectQuery/aggregations so only recent/summary data is imported. Also note: with a Pro license the published dataset must be ≤1 GB (compressed),a multi-GB model needs PPU/Premium or further reduction. For Desktop refresh, ensure roughly 2–3× the model size is available as free RAM.
