Forum Discussion
Refresh Failures
- 1 year ago
Hi,
I found a workaround that resolved the issue for me.In the query code, look for queries having the `[Implementation="2.0"]` parameter:
`= Value.NativeQuery(GoogleBigQuery.Database([Implementation="2.0"])`
and replace them with
`= Value.NativeQuery(GoogleBigQuery.Database()`
by deleting `[Implementation="2.0"]`
It's working fine.
Hi VC ,
The error “Cannot execute b__0 after 5 tries. The exception was raised by the IDbCommand interface” usually means the service tried to run a query against your data source but couldn’t complete it, and kept retrying. Since your desktop refresh works but service refresh does not, this points to a gateway/credentials/network issue rather than the query itself.
What to check and how to fix:
Identify the data source and connection type
What is the data source (SQL Server, Oracle, etc.)?
Are you using an On-premises data gateway? If yes, which edition (Standard or Personal) and is it online?
Check the Power BI refresh history
In the Power BI service, open the failing dataset and view Refresh history.
Copy the full error message and any inner exception details. This often points to a credential or connectivity problem.
Gateway health and connectivity (if on-premises data)
Ensure the gateway is online and the host machine has network access to the data source.
Verify the gateway is up to date with the latest version.
In the gateway config, verify the data source connection is enabled and the credentials are correct.
Check firewall/network rules between the gateway machine and the data source (port opening, VPN, etc.).
Data source credentials in the service
Go to the dataset in Power BI Service > Settings > Data source credentials.
Re-enter credentials and test the connection. Make sure you use the correct authentication type (Windows, Basic, OAuth, etc.) and that the account used by the gateway has access to the data source.
Permissions of the gateway service account
The account used by the gateway must have read access to the data source and any underlying objects (tables, views, stored procedures).
If a service principal or managed identity is used, ensure it has the necessary database permissions.
Timeout and workload considerations
Large or slow queries can time out or fail under service load. If possible, review the query workload:
Simplify heavy queries or break large loads into smaller steps.
If you’re using DirectQuery, consider optimizing the underlying queries or using a cached import where feasible.
Some gateways allow a command/query timeout setting; if available, consider increasing it.
Please mark this post as solution if it helps you. Appreciate Kudos.