Dataset refresh failure - DM_GWPipeline_Gateway_MashupDataAccessError

[dbeavon3]

Can someone translate this to English, on behalf of a coworker who is not a power bi admin:

Data source error: {"error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","pbi.error":{"code":"DM_GWPipeline_Gateway_MashupDataAccessError","parameters":{},"details":[{"code":"DM_ErrorDetailNameCode_UnderlyingErrorCode",
"detail":{"type":1,"value":"-2147467259"}},{"code":"DM_ErrorDetailNameCode_UnderlyingErrorMessage","detail":{"type":1,"value":"We're sorry, an error occurred during evaluation."}},{"code":"DM_ErrorDetailNameCode_UnderlyingHResult","detail":{"type":1,"value":"-2147467259"}}],"exceptionCulprit":1}}}
Table: GL Balance Sheet.
Cluster URI: WABI-WEST-US-C-PRIMARY-redirect.analysis.windows.net
Activity ID: 0671e40d-a364-4b97-8c8c-f84f5212694e
Request ID: 1882c29b-31ac-4f4e-9a7a-a2967902bea2
Time: 2023-08-15 13:02:55Z

They would like to know what we can do to allow users to see the underlying reason for these meaningless messages.  Is it reasonable to share the logs directory with them over the network? Eg:
C:\windows\ServiceProfiles\BuggyPbiService\AppData\Local\Microsoft\On-premises data gateway

[rico343]

In our case the logs on the gateway server showed there was not enough space on the disk.  Other reports worked fine, but one that is gathering a large amount of data started failing.  

[AleMaldonado]

la solucion esta en configurar el mismo nivel de privacidad en la puerta de enlace para todos los origenes en el servicio.
Editen las credenciales de origenes de datos y seleccionen el mismo para todos. encontraran las fuentes y la opcio de editar. 

Me paso recien y de esa manera lo pude solucionar.  Saludos 

[sulleyinoz]

This might be related to this issue reported a couple of days ago if it used to refresh and has now stopped with no changes.

https://community.fabric.microsoft.com/t5/Service/Power-BI-Service-Dataset-Refresh-Error-Expression-...

[dbeavon3]

It is impossible to say unless I can dig into the logs.  

Unfortunately when I'm using the Power BI service, all I can do is see the message above ("we're sorry something bad happened.").

[dbeavon3 · Accepted solution]

It was only possible to find the true underlying reason for this error message by reading the gateway logs.  Microsoft needs to make those available to the people who need them.  The admins in our environment do not necessarily want to participate in the troubleshooting of every failing dataset-refresh.

[jen15]

Same problem now in Nov 2023. My reports were refreshing well until some update recently has left me with same problem but no answers!

[dbeavon3]

Hi @jen15 

I've been working with Power BI for a few years.  Unfortunately I've come to accept that the enterprise gateway was built in a way that requires PQ developers to also be able to monitor and administer the gateway.

For many types of PQ failures, you must enable advanced logging and read the output, after your refresh has failed.

The best way to think of the Power Query language, is as a primitive programming language.  Unlike most programming languages which present errors to you directly, PG's error messages are trapped locally on the "gateway" server in the "advanced" logs.  If you go ahead and open a ticket with CSS about the failures, they will immediately ask your gateway administrator to enable advanced logs and share the output.    In the case that you are not the gateway administrator yourself, these cases take a lot longer than reasonable.

Until Microsoft finds a way to present PQ failure messages to dataset owners, you will need to be a gateway admin yourself, or work closely with a gateway admin to see the underlying errors