Forum Discussion
Sporadic ODBC Error: “Invalid SQL_DRIVER_ODBC_VER (03.80)” During Power BI Refresh (~20% Failures)
Hi,
I've got 2 more suggestions - if you're having an issue with the load from the driver, then it makes sense to reduce the load as much as possible. I do this by writing an SQL statement (under the optional setting when establishing the connection within Power BI).Might be worth limiting it to just the fields from the table that you need. Then repeat this for the other tables and have them as separate queries.
You should only have 1 query for each table that you're trying to retreive. If you need multiple queries for the same table, then make one base one and make several other referenced queries (this is much better for performance and is a best practice).
The second point was that if you're seeing that it's failing 1 in 10 times when you're directly calling data through the driver, then it looks like there's an issue with either the server or the connector.
Either way, what makes sense to do is:
1. Enable the automatic emails when a refresh fails.
2. Create a Power Automate flow, which has a trigger to run, based on that email (very simple to make this).
3. Then add a step in the Power Automate flow to refresh the semantic model. This will basically allow you to build a custom retry mechanism. You could even try the refresh multiple times and add delays in between. Then maybe send an email back to yourself if it doesn't get refreshed after x number of retries.
Community SupportHi Flávia , Thank you for reaching out to the Microsoft Fabric Community Forum.
You’re running into an intermittent driver-level compatibility issue, not a Power BI modeling problem. The error indicates that when the Power BI gateway queries the MongoDB ODBC driver for its version, the driver occasionally returns an invalid response. Since only some refreshes fail, this points to instability in the ODBC layer or how the gateway hosts it, not to your queries or credentials.
To confirm, test the driver directly on the gateway machine using ODBC Data Source Administrator or a simple pyodbc loop to repeatedly connect to MongoDB. If you see the same SQLSTATE 01000 warnings, the driver itself is the cause. In that case, try MongoDB Atlas SQL ODBC 1.5.x or 1.6.x, which are generally more stable than 2.x. Also, disable pooling (Pooling=No) and increase connection timeout in the DSN or connection string, both help when a driver fails during version negotiation. Check gateway logs under
C:\Program Files\On-premises data gateway\ -> MashupTraceLogs and GatewayPerformanceLogs for lines mentioning SQL_DRIVER_ODBC_VER or NativeError 444.
pbiuseruk was right, reducing concurrent refreshes and improving query folding can ease load on the gateway; buffering steps can enforce proper Power Query sequencing and occasionally reinstalling the gateway can clear cached driver states. Those steps improve refresh reliability.
If the issue continues after testing an older driver and disabling pooling, capture a failing trace and share it with MongoDB support. For more consistent reliability, you could also consider offloading data ingestion to Azure Data Factory or OneLake.
Power Query ODBC connector - Power Query | Microsoft Learn
Troubleshoot the on-premises data gateway | Microsoft Learn
Monitor and optimize on-premises data gateway performance | Microsoft Learn
Hi Flávia
I wanted to check if you had the opportunity to review the information provided by user. Please feel free to contact us if you have any further questions.
v-hashadapu thanks for your sugestions,
I tested the driver on the gateway machine — it fails around 10% of the time (intermittent connection failures).
I also tried reverting to the previous version, but it threw a different error: Unsupported cluster configuration: Unsupported cluster type. The driver is intended for use with MongoDB Atlas Data Federation.I’ll dig into that next and see if reducing MongoDB refresh frequency helps. Any other suggestions are welcome!
- v-hashadapu
Hi Flávia , Thank you for reaching out to the Microsoft Fabric Community Forum.
"Unsupported cluster configuration: Unsupported cluster type" error signals that the older drivers are incompatible with your deployment because they're designed for Atlas Data Federation, not standard Atlas clusters.
While reducing refresh frequency may alleviate some pressure, it won’t address the underlying compatibility gap. The best next step is to verify which cluster type your MongoDB Atlas instance uses and ensure you're installing the corresponding recommended ODBC driver; MongoDB maintains clear compatibility tables and tends to restrict newer drivers to specific cluster types. If cluster reconfiguration or driver alignment isn’t possible on your infrastructure, the most reliable long-term move is to shift ETL and reporting workloads away from direct ODBC connections, for example, ingesting data through a pipeline into Azure Data Lake or SQL for Power BI to consume. This sidesteps both random driver failures and compatibility drift issues entirely.
Ultimately, there’s no seamless client-side fix once driver and cluster compatibility are mismatched, so collaborating with your MongoDB admin or support to clarify cluster details and driver fit or pivoting architecture for more stable data access will deliver the most sustainable results.
MongoDB Atlas SQL interface connector - Power Query | Microsoft Learn
Error connect MongoDB Atlas SQL ODBC - MongoDB Atlas - MongoDB Developer Community Forums
Connect with ODBC Driver - Atlas - MongoDB Docs
