Forum Discussion
Model Refresh Errors after June 2026 Gateway Upgrade Resolved Themselves?
I updated my Dev on Premise Data Gateway to the June 2026 Release on Monday, tested by refreshing my main Dev model in the Power BI Service through the gateway, and it succeeded.
I waited until today and confirmed that my Dev Gateway was still good and all connections were still good, and proceeded to update Prod.
The prod update was successful, but upon trying to refresh my main prod model, it failed with generic:
| Data source error: | {"error":{"code":"DM_GWPipeline_Gateway_DataSourceAccessError","pbi.error":{"code":"DM_GWPipeline_Gateway_DataSourceAccessError","parameters":{},"details":[],"exceptionCulprit":1}}} |
I went through all of the normal troubleshooting with my buddy, CoPilot, and confirmed that:
- No errors with the install.
- Gateway service network test all good, all green checkboxes.
- Service side showed gateway updated, online, and all connections good.
- Re-entered credentials for a few connections just to see if that helped.
- Confirmed that many paginated reports, and other semantic models using the prod gateway and connections worked fine.
- Published a new semantic model using the same gateway and connection as my main prod model, and it refreshed service side fine.
- Refreshed prod model locally in Power BI Desktop and it was fine.
- Re-published it to service and failed again with same error above.
- Tried refreshing the equivalent Dev model against the Dev Gateway that was upgraded on Monday and worked, and it now gave the same error.
- Exported Gateway logs including Request ID associated with the failures. Gave them all to Co Pilot and it determined it was some type of Mashup issue, but didn't know exactly what. It suggested either rolling back to a prior gateway version, or starting to disable queries in my model until I found the broken one.
I was getting ready to do one of these painful options, so I tried refreshing my prod model a fifth time in the service, and it suddenly worked, despite no changes to connection, credentials, or the .pbix file.
I then hit refresh on my dev model in the service and it suddenly worked too.
So now I'm uncertain if I'm going to have intermittent issues like this, or if this was just a temporary Microsoft glitch?
Co Pilot claims there were no known outages or issues that would cause this.
I guess I just need to keep an eye on it, but thought I would post here if others have similar issue.
Hi Karl-D1,
From what you've described, the gateway installation appears to have completed successfully, other semantic models and paginated reports continued to refresh successfully, and the affected semantic models eventually refreshed without any configuration changes .There is currently no published Microsoft Known Issue that describes intermittent DM_GWPipeline_Gateway_DataSourceAccessError failures following the June 2026 On-premises Data Gateway upgrade.
Since the issue resolved on its own, it appears to have been a transient condition rather than an ongoing gateway configuration issue.I'd recommend monitoring future refreshes to see if the behavior recurs. If it does, please capture the Activity ID, Request ID, Correlation ID (if available), Timestamp (UTC), and the gateway logs, then open a Microsoft Support ticket so the backend diagnostics can be investigated.
Create a Fabric and Power BI Support Ticket - Power BI | Microsoft Learn
Thank you.
10 Replies
Hi Karl-D1,
From what you've described, the gateway installation appears to have completed successfully, other semantic models and paginated reports continued to refresh successfully, and the affected semantic models eventually refreshed without any configuration changes .There is currently no published Microsoft Known Issue that describes intermittent DM_GWPipeline_Gateway_DataSourceAccessError failures following the June 2026 On-premises Data Gateway upgrade.
Since the issue resolved on its own, it appears to have been a transient condition rather than an ongoing gateway configuration issue.I'd recommend monitoring future refreshes to see if the behavior recurs. If it does, please capture the Activity ID, Request ID, Correlation ID (if available), Timestamp (UTC), and the gateway logs, then open a Microsoft Support ticket so the backend diagnostics can be investigated.
Create a Fabric and Power BI Support Ticket - Power BI | Microsoft Learn
Thank you.
Hi,
It looks like you already covered all the right troubleshooting steps. Since both your Dev and Prod models started refreshing successfully without any changes, it doesn't seem like there's an issue with your gateway installation or the semantic model itself.
This was most likely a temporary Power BI Service or gateway-related issue. I'd keep an eye on the refresh history over the next few days. If the same error starts happening again, then it's worth opening a Microsoft support ticket and sharing the gateway logs you already collected.
Hopefully it was just a one-time glitch, but it would be helpful to know if anyone else who upgraded to the June 2026 gateway has experienced the same issue.
Hope this helps!
Thanks!
Worked fine Friday, and through the weekend. Refresh failed in prod early this morning. I then tried manual refresh in the service and it failed with same errors from last Thursday. I then tried refreshing Dev (totally separate gateway, connections, database, network, etc.) and it is failing again this morning too.
...and now it is working again. I didn't change any settings or touch either of the failing models, and they started refreshing again.
We are experiencing the same issue. We run two Capacities and two clusters of on-prem data gateways. Following successful updates on both clusters on Friday (17/07) we have been experiecing multiple model refresh failures across both clusters.
- Different models each day
- Different gateway data sources (many models use the same source, most refresh some fail)
This morning a model refresh failed TEN times before succeeding on the 11th attempt. This is significantly impacting our ability to deliver our products.
Error sample:
Data source error: 0xC1450012: {"error":{"code":"DM_GWPipeline_Gateway_DataSourceAccessError","pbi.error":{"code":"DM_GWPipeline_Gateway_DataSourceAccessError","parameters":{},"details":[],"exceptionCulprit":1}}} 0xC112001C: The operation was canceled.. The exception was raised by the IDbCommand interface. 0xC11C0006: The current operation was cancelled because another operation in the transaction failed.
Hi pixmeister ,
Thanks for letting me know I'm not alone here.
I ended up opening a support case.
I don't fully understand how all of these settings work, or what this really changes, so I don't want to encourage you to do this unless you are comfortable with it.Below is what I was advised to do. I've not had a refresh error on the two gateways since doing this on Monday. However, even when I was experincing the problem, I had a couple days when it didn't happen. So if I make it to Monday the 27th without any refresh errors, it would lead me to belive that this resolved the issue.
Suggestions:
To mitigate this issue, we suggest updating the Gateway configuration by enabling the following setting:
"NextResultSkipped = True" ------ intially it is set to False, need to change it to True and save the changes to the file.
This setting should be added to the configuration file located at: C drive or whichever drive your gateway is installed in ----C:\Program Files\On-premises data gateway\Microsoft.PowerBI.DataMovement.Pipeline.GatewayCore.dll.config
Thanks so much for posting this. I've had a support ticket open for days but the support has not been great. I implemented this fix and have not had a refresh failure since. You'd think MS would have a knowledge base for this kind of thing and, your engineer having provided the resolution, it would be a short hop for my engineer to find the same info. Oh well.
Fabric Community wins the day
I just had to roll back our gateway servers to before the June update. I did the other workarounds to fix data import reports for Oracle but none of the workarounds I could find fix DQ to Oracle. I can fix it in power BI desktop when I turn off 'Enable using Bundled Oracle Managed ODP Provider for Import Mode' & for Direct Query Mode but when the report it published to Fabric it breaks again ... I dont see any option in Fabric Admin Portal to turn this off
