Issues Replicating Large CDC-Enabled Tables Using Microsoft Fabric SQL Server Database Mirroring

[spin1888]

Hi everyone,

We’re currently testing the Microsoft SQL Server (on-prem) Database Mirroring feature in Microsoft Fabric (Preview) and have encountered a critical blocker when attempting to mirror large tables with Change Data Capture (CDC) enabled.

---

What’s Working:

• CDC is properly enabled and validated 

• Copied on-prem table with the same schema but with 1000 records. CDC is replicating fine without issues

• Smaller tables in the same mirrored database replicate successfully using this feature

---

Issue Summary:

• Attempting to mirror large tables (e.g., 70+ million rows) with CDC enabled results in:

  • No rows being processed even after 8+ hours (as observed in the Monitor KQL database)

  • Eventual failure of the mirroring session with errors such as:

    • Timeout expired. The timeout period elapsed prior to completion of the operation.

    • InputValidationError

    • ErrorCode: InputValidationError

    • ArtifactId: [GUID]

    • 

---

Troubleshooting Steps Taken:

• Restarted the mirroring session multiple times

• Verified  indexing on the source table

• Monitored LSN progression and system resources during sync attempts

Despite these actions, large tables never complete the initial snapshot phase, and the mirroring pipeline ultimately fails.

---

Request:

• Has anyone else experienced this issue when replicating large CDC-enabled tables?

Any insights or shared experiences would be greatly appreciated.

[v-menakakota]

Hi  @spin1888 ,
Thanks for reaching out to the Microsoft fabric community forum. 

Please go through the below documentation which may help you in resolving the issue:

Limitations in Mirrored Databases From Azure SQL Managed Instance (Preview) - Microsoft Fabric | Mic...

As this feature is still in preview, there are known scalability and performance limitations, especially for very large datasets. These limits are actively being worked on by the product team. I hope you understand well.

 

If this post was helpful, please give us Kudos and consider marking Accept as solution to assist other members in finding it more easily.

Best Regards, 
Menaka Kota 
Community Support Team 

[spin1888]

Thank you @v-menakakota,

Unfortunately, I am not using the Mirrored Azure SQL Managed Instanced option. I am using the Mirrored SQL Server option (screenshot below) and meet all the conditions mentioned here:

https://learn.microsoft.com/en-us/fabric/database/mirrored-database/sql-server-limitations

I couldn't find any documentation or notes indicating scalability and performance limitations with very large datasets. Hopefully the product team comes up with a resolution soon. That feature is a game changer for those interested in moving from on-prem to Fabric.

[v-menakakota · Accepted solution]

Hi @spin1888 ,

I recommend raising a formal support ticket with Microsoft so the product team can review the issue in more detail. This will allow them to analyze backend logs related to your mirroring session, including the InputValidationError and timeout issues you mentioned.

Please refer below link on how to raise a contact support or support ticket. 
How to create a Fabric and Power BI Support ticket - Power BI | Microsoft Learn 

If this post was helpful, please give us Kudos and consider marking Accept as solution to assist other members in finding it more easily.
Best Regards, 
Menaka Kota 
Community Support Team