When you restore a database that contains memory-optimized tables in SQL Server 2016, the server stops responding (hangs).
When a large volume of work is performed on memory-optimized tables in a single transaction (including LOBs), a checkpoint that's run immediately after that transaction takes longer than expected to complete. This issue also occurs during restore for redo of the same transaction.
This update is included in Service Pack 1 for SQL Server 2016.
Note When this update is installed, the issue is resolved by increasing the size of an internal hash table tracking the LOBs to minimize the collisions.
Cumulative update information
The fix for this issue is included in the following cumulative update for SQL Server:
Each new build for SQL Server 2016 contains all the hotfixes and all the security fixes that were included with the previous build. We recommend that you install the latest build for SQL Server 2016.
Microsoft has confirmed that this is a problem in the Microsoft products that are listed in the "Applies to" section.
Learn about the terminology that Microsoft uses to describe software updates.