Service Principal Key Update, We couldn't load your warehouse An internal error occurred.

[zzthatcher]

Anyone else experiencing this?

Our most important warehouse and lakehouse were completely inaccessible via any sql analytics endpoint, which means all of our reports were inaccessible. And the pipelines that land data into the warehouse were broken.

It seems like it was brought on by simply updating the service pricipal key and then attempting an odbc connection query with the new key.

Pausing and restarting the capacity was the only way to resolve this.

[v-veshwara-msft]

Hi @zzthatcher ,
Thanks for usoing Fabric Community and sharing the details.

Based on your description, this behavior might be related to a known issue where:

Permissions applied to SQL analytics endpoint tables can be lost after schema changes, which may require them to be re-applied.

Reference – Known Issues

Although we can't say for certain that this is the root cause in your case, it's worth considering-particularly if there were any recent schema or security changes.

Restarting the capacity has helped in similar situations, as it can refresh metadata and restore access in some cases.

If the issue persists or recurs, we recommend raising a support ticket here  for a deeper investigation.

Let us know if you need further suggestions or have other questions.

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly and a kudos would be appreciated.

Best regards,
Vinay.

[zzthatcher]

Hi @v-veshwara-msft,

Thank you for the suggestion.

Maybe the underlying issue was similar. But this is a much worse result.

I am the capacity/workspace admin and creator of these fabric items, so it seems like a large flaw if my permissions were rejected for an entire lakehouse and warehouse endpoint because I attempted to access them outside of fabric with an existing service principal with a new secret value. To be clear, I was able to access them previously with this service principal. The only thing that changed is that I started using a new secret key. And that disabled all access. And it was all restored after restarting the capacity.

It also seems strange to me that a permissions issue would give me an error message "An Internal Error Occurred". But maybe that is because I should obviously be allowed to view the data as the creator of the item and workspace/capacity admin.

I raised the issue to support before I figured out how to resolve it on my own, so they are aware. 

[v-veshwara-msft]

Hi @zzthatcher ,

Thanks for the follow-up. Really helpful details.

Rotating a service principal’s secret shouldn't change its permissions, but in some cases there can be a brief disruption right after the update, especially if there are active sessions or cached tokens still using the old key. That could explain the "Internal Error" and why restarting the capacity helped reset things.

It's good that you already reported this through support. They’ll be able to dig deeper into what happened behind the scenes. Appreciate you bringing this up as these edge cases help improve the overall experience.

Let us know if you run into anything else.

If any insights come from the support team that could help others, feel free to share them here in the thread. Also, if any of the replies above were helpful, please consider marking as the Accepted Solution to assist others facing a similar issue.

Thank you.