Re: SQL Endpoint from Lakehouse is SUSPECT

HarderT

Hello,

Since the May 22nd update, some of the Lakehouse SQL endpoints are showing as SUSPECT in SSMS.

I can see the data in Lakehouse, but when I switch to the SQL endpoint, I don't see any data.

Is there a solution?

v-dineshya

Hi @HarderT ,

Thank you for reaching out to us on the Microsoft Fabric Community Forum.

Please try below option to fix the issue.

1. Refresh the Lakehouse Metadata

Go to the Lakehouse experience in Microsoft Fabric, Open the affected Lakehouse. Click "Settings" > "SQL Endpoint".
Try turning the SQL Endpoint off and on again to reset the connection. Alternatively, click "Refresh metadata" if that option exists.

2. Verify Permissions

Ensure that the user or service principal connecting via SSMS, Has Viewer or above permission on the Lakehouse workspace.
Is assigned a Fabric role that permits access to SQL endpoints. You can check this in Microsoft Fabric / Power BI Service > Workspace > Access.

3. Test with Power BI or Azure Data Studio

Sometimes SSMS lacks compatibility with new authentication models, Try connecting using Azure Data Studio or Power BI Desktop using the same endpoint. If Power BI works, the issue is likely specific to SSMS or the AAD token not refreshing correctly.

4. Manually Recreate the SQL Endpoint

If metadata corruption is suspected, Create a new Lakehouse or duplicate the existing one. Check if the new SQL Endpoint works as expected.

Please refer community threads.

Solved: SQL endpoint lags behind lakehouse - Microsoft Fabric Community

Solved: Schema on lakehouse sql endpoint - Microsoft Fabric Community

Solved: Lakehouse tables not showing up in SQL Endpoint - Microsoft Fabric Community

Solved: Is Case-insensitivity in Lakehouse SQL endpoint po... - Microsoft Fabric Community

Solved: SQL Endpoint Slow To Reflect Changes In Lakehouse - Microsoft Fabric Community

Solved: Re: Connecting to SQL end point in Fabric from SSM... - Microsoft Fabric Community

Solved: Re: Can't connect to lakehouse sql endpoint from P... - Microsoft Fabric Community

If my response has resolved your query, please mark it as the "Accepted Solution" to assist others. Additionally, a "Kudos" would be appreciated if you found my response helpful.

Thank you

HarderT

Hi @v-dineshya ,

thanks for the advice.

Option 1:

I dont have the option "Refresh Metadata" or connection? Do I have to have PowerBI-Admin rights for this?

Option 2:

Im admin of the workspace, so I cant be a permission problem. At least at the workspace level.

Option 3:

I tried with PowerBI Desktop and I got the following error:
Warning: Fatal error 615 occurred at May 27 2025 11:30AM. Note the error and time, and contact your system administrator. Login failed for user '<token-identified principal>'

Option 4:

That's my last resort, but it means a lot of work because we're already using Fabric productively and I'll have to change a lot of things. That's why I'm looking for a perhaps simpler solution.

Many thanks for your help!

v-dineshya

Hi @HarderT ,

Please follow below steps.

1. Use Azure Data Studio with Azure Active Directory - Universal with MFA

SSMS does not handle token refresh well, especially in Fabric/AAD contexts.

open Azure Data Studio, Try connecting using Authentication type: Azure Active Directory - Universal with MFA

Server: your-sql-endpoint-url (from Fabric Lakehouse). If it connects and shows your tables, the issue is SSMS specific.

2. Manually Refresh SQL Endpoint Metadata (with Admin Rights)

You mentioned not seeing the “Refresh Metadata” option this might be restricted to Fabric Admins or require certain roles.

Try this workaround to fix:

Go to Fabric > Lakehouse, From the Lakehouse Explorer panel, create a dummy table or modify a schema slightly
(e.g create/drop a test table). This sometimes forces a metadata refresh under the hood, similar to how auto-refresh gets triggered in dataflows. Then try reconnecting via SSMS or Power BI Desktop.

If my response has resolved your query, please mark it as the "Accepted Solution" to assist others. Additionally, a "Kudos" would be appreciated if you found my response helpful.

Thank you