KB3147012 - FIX: Large disk checkpoint usage occurs for an In-Memory optimized filegroup during heavy non-In-Memory workloads

Symptoms

Assume that you have a database that contains a memory-optimized filegroup and ordinary tables. Additionally, the database might contain in-memory online transaction processing (OLTP) memory-optimized tables. When you perform a Data Manipulation Language (DML) operation such as rebuild the index of the ordinary tables, the number of the checkpoint files keeps growing.

Resolution

We have released an update to reduce the disk footprint by reducing the In-Memory checkpoint files to 1 MB (megabytes) each. This update is included in the following cumulative update for SQL Server:

Cumulative Update 6 for SQL Server 2017

Cumulative update 8 for SQL Server 2016 SP1

Cumulative Update 6 for SQL Server 2014 Service Pack 1

Cumulative Update 1 for SQL Server 2016
To enable this update, you must enable trace flag 9929 at startup or in a user session. This trace flag has global-level effect. To enable trace flag 9929, use the DBCC TRACEON (9929, -1) command or use –T 9929 as a startup parameter. Because this traceflag is instance level, it affects the size of checkpoint files for all databases with a memory optimized filegroup on that instance. Be cautious when you use this traceflag.

For detailed information about how to enable or disable trace flags and for explanations of global and session-level trace flags, refer to the following topics in SQL Server Books Online:

• DBCC TRACEON (Transact-SQL)

• How to: Configure Server Startup Options (SQL Server Configuration Manager)

Status

Microsoft has confirmed that this is a problem in the Microsoft products that are listed in the "Applies to" section.

References

Learn about the terminology Microsoft uses to describe software updates.