How to fix Error 678 on Windows Server 2025 / Windows 11

by

Troubleshooting Guide: SQL Server Error 678 – Active Rowset for Partition ID Found at End of Batch

Why This Error Happens

This error occurs when SQL Server’s storage engine detects that an internal rowset associated with a specific partition ID was left active after a batch of operations completed, which typically points to improper exception handling during a transaction or bulk operation. It commonly surfaces during large index rebuilds, partition switching, or parallel query execution where a session was interrupted (via timeout, network failure, or manual cancellation) without allowing the engine to properly clean up its internal cursors. In some cases, this is also a symptom of underlying corruption in the partitioned table’s metadata or an incompatible driver/client library mishandling connection resets.

Step-by-Step Resolution Guide

Step 1: Identify the Offending Session and Terminate It

  • Open SQL Server Management Studio (SSMS) and navigate to Management > Activity Monitor.
  • Locate the Session ID (SPID) referenced in the error log tied to the partition operation.
  • Run the following to confirm the process is still active:
  SELECT session_id, status, last_request_start_time, blocking_session_id
  FROM sys.dm_exec_requests
  WHERE session_id = <SPID>;
  • If confirmed hung or orphaned, terminate it using:
  KILL <SPID>;

Step 2: Check for Partition-Level Corruption

  • Run DBCC CHECKTABLE against the affected table to rule out index/partition corruption:
  DBCC CHECKTABLE ('YourTableName') WITH ALL_ERRORMSGS, NO_INFOMSGS;
  • Review the SQL Server error log (default path: C:\Program Files\Microsoft SQL Server\MSSQL16.MSSQLSERVER\MSSQL\Log\ERRORLOG) for accompanying corruption warnings (Error 823, 824, or 5180).
  • If corruption is found, restore the affected partition from the last known good backup.

Step 3: Update SQL Server Cumulative Updates and Drivers

  • Confirm your instance is on the latest Cumulative Update (CU) for SQL Server 2025 via SSMS > Help > About.
  • Download and apply the latest CU from the official Microsoft Update Catalog.
  • On the client side, ensure the ODBC Driver or OLE DB Driver used by the application connecting to SQL Server is updated to the latest version compatible with Windows Server 2025 / Windows 11.
  • Restart the SQL Server service after patching:
  Restart-Service -Name MSSQLSERVER -Force

Step 4: Review and Adjust Query Timeout / Batch Settings

  • If the error correlates with long-running batch jobs, increase the command timeout setting in your application’s connection string (e.g., Connection Timeout=60 or higher).
  • For SSIS or scheduled jobs, edit the timeout property in SQL Server Agent > Job Steps > Advanced to allow sufficient completion time before forced termination.

Alternative or Verification Steps

  • Verify the fix: Re-run the original batch/partition operation and monitor sys.dm_exec_requests and the SQL Server error log to confirm Error 678 no longer appears.
  • Alternative solution: If the issue persists after patching and session cleanup, rebuild the affected partitioned index entirely using ALTER INDEX ... REBUILD PARTITION = ALL to force SQL Server to regenerate clean rowset metadata.