Issue:
Found an inconsistency in error handling for a complex query. When using multiple GROUP BY clauses:
Error writing file '/tmp/MYqMk1xr' (Errcode: 28 - No space left on device)
Key observations:
- Query returns only 1,911 rows
- Removing one GROUP BY clause makes it work
- Metabase incorrectly marks it as
:error_type :invalid-query - The actual error is
SQLTransientConnectionException - Same query works directly in MySQL
Environment:
- Database: MySQL
- Deployment: Docker + Postgres
This suggests the "No space left" error might be masking a timeout or query optimization issue rather than an actual storage problem. Could Metabase's timeout settings be affecting how MySQL handles temporary tables for GROUP BY operations?