How to fix Error 3198 on Windows Server 2025 / Windows 11

by

Troubleshooting Guide: Error 3198 – I/O Was Resumed on Database

Why This Error Happens

Error 3198 is an informational message indicating that I/O operations on a database have resumed after being temporarily suspended, typically following a BACKUP or VSS-based snapshot operation using FREEZE/THAW semantics. This occurs when a snapshot provider (like a third-party backup tool or Windows Volume Shadow Copy Service) halts write activity momentarily to ensure a consistent point-in-time copy of the database files. While SQL Server logs this as a message rather than a true error, frequent or unexpected occurrences can point to misconfigured backup schedules, VSS writer timeouts, or storage subsystem latency causing the “thaw” phase to be delayed longer than expected.

Step-by-Step Resolution Guide

Step 1: Confirm the Trigger Source

  • Check the SQL Server Error Log and Windows Event Viewer (Application log) around the timestamp of Event ID 3198 to correlate it with a backup job or VSS snapshot request.
  • Run vssadmin list writers in an elevated Command Prompt on Windows Server 2025 to verify the SqlServerWriter service status is “Stable” and reports “No Error.”
  • Identify if the message coincides with a third-party backup agent (e.g., Veeam, Commvault) taking application-consistent snapshots.

Step 2: Validate VSS Writer and Service Health

  • Ensure the SQL Server VSS Writer service is running via services.msc or by executing:
  sc query SQLWriter
  • If stopped, restart it using:
  net start SQLWriter
  • Check that the service account running SQLWriter has appropriate permissions on the database file directories (typically C:\Program Files\Microsoft SQL Server\MSSQL17.MSSQLSERVER\MSSQL\DATA).

Step 3: Review Storage and I/O Latency

  • Use Performance Monitor (perfmon) to track the Avg. Disk sec/Transfer counter on the volume hosting your .mdf/.ldf files; values consistently above 15-20ms indicate I/O bottlenecks that can prolong the freeze/thaw cycle.
  • If using SAN or cloud-attached storage, confirm multipathing drivers and storage firmware are updated to the latest version compatible with Windows Server 2025.
  • Adjust backup software timeout settings to allow SQL Server sufficient time to complete the I/O freeze cycle without triggering cascading alerts.

Step 4: Adjust Backup Job Scheduling

  • Stagger overlapping backup jobs (SQL native backups and VSS-based snapshots) so they do not run concurrently against the same database, which can cause repeated freeze/thaw cycles.
  • Modify the backup schedule in SQL Server Agent Jobs or your third-party backup console to avoid peak transaction periods.

Alternative or Verification Steps

  • Verify resolution by monitoring the SQL Server Error Log for 24-48 hours to confirm Error 3198 appears only during expected, scheduled backup windows rather than at random intervals.
  • If the issue persists, run DBCC CHECKDB on the affected database to rule out underlying corruption, and consider switching the backup method from VSS-based snapshots to native SQL Server full/differential backups as a temporary workaround.