WinCC v7.3 Archives Lost After Server Restart Recovery

Problem Overview

Symptom: After an uncontrolled server shutdown caused by a main power cut, a WinCC v7.3 SP3 Update 7 (2-server / 2-client, non-redundant) project reports missing trend data and lost taglogging/alarmlogging archives from the period spanning the restart window. The *.mdf and *.ldf files exist on disk but WinCC Trend control, OnlineTableControl, and Alarm Control return empty time ranges or "no data available" messages for the affected segment window.

This failure mode is most commonly observed in plants where:

• Multiple WinCC project copies (e.g., Project1 and Project2) exist on the same machine with an autostart shortcut pointed at only one of them.
• The UPS (uninterruptible power supply) is undersized for the runtime load, runs flat before operators intervene, and the servers hard-power-off.
• Archive backup is disabled in Taglogging and Alarmlogging editor configuration.
• SQL Server 2008 R2 / 2014 (the default instance bundled with WinCC v7.3) marks the databases as SUSPECT after abrupt shutdown, and the WinCC ArchiveManager cannot attach them at runtime startup.

Affected versions: Siemens WinCC V7.3, V7.3 SP1, V7.3 SP2, V7.3 SP3 with Update 7 installed. The same recovery flow applies to V7.4 and V7.5 with minor folder-name changes.

WinCC v7.3 Archive Database Architecture

Before restoring archives you must understand the file layout WinCC writes to disk. A runtime project stored at C:\Siemens\WinCC\Project1 contains:

Root Cause Analysis for Lost Archives

Five root causes have been observed in field deployments running WinCC v7.3 Update 7:

1. SQL Server SUSPECT state after power loss – When the OS loses power mid-transaction, the SQL Server database engine cannot complete the recovery phase. The database is marked SUSPECT in sys.databases, WinCC ArchiveManager cannot attach it, and the runtime window displays no trend data even though the file exists on disk.
2. Wrong project autostart – Operators or commissioning engineers duplicate the project folder to create a sandbox (e.g., Project2) but leave the Windows Startup shortcut or WinCC autostart list pointing at the wrong project. After the restart, the live archive database belongs to Project2; Project1 displays a gap.
3. Archive backup not configured – Taglogging Editor and Alarmlogging Editor both offer a "Backup" tab with destination path and schedule. If left blank, only the live *.mdf / *.ldf is written; no redundant copy exists to recover from.
4. UPS runtime exhausted – A typical server-grade UPS sized for 15 minutes of load cannot bridge a multi-hour outage. Once batteries deplete, the server hard-powers-off, skipping the Windows shutdown sequence and any controlled WinCC stop.
5. Server package mismatch – After project edits the server package on the client must be regenerated. If the client holds a stale package, trend requests are routed to segment indices that no longer exist on the server, producing empty trend windows for valid time ranges.

Recovery Procedure A – Reattach SQL Database from Backup Folder

When the Taglogging and Alarmlogging Backup tab is configured (recommended baseline), perform the following controlled restore. This procedure recovers archives without disturbing the running project.

Prerequisites

• Local Administrator rights on the WinCC server.
• SQL Server Management Studio (SSMS) version matching the WinCC-bundled instance (2008 R2 for WinCC v7.3 baseline, 2014 for V7.4+).
• Runtime may remain active if you are copying files into CommonArchiving; otherwise stop Runtime first.

Step-by-Step

1. Stop WinCC Runtime on the server. Close WinCC Explorer.
2. Open SSMS, connect to the WinCC instance (named instance WinCC by default).
3. Detach the corrupted database by right-clicking <Project>Tlg in Object Explorer and selecting Tasks → Detach. Tick "Drop Connections" and "Update Statistics".
4. Navigate to the configured backup folder (default ArchiveManager\Backup). Identify the *.mdf and *.ldf pair whose timestamp matches the gap window.
5. Copy the backup pair into the live ArchiveManager folder of Project1. Do not copy into CommonArchiving for the primary database – the live folder is the correct location for the runtime attach.
6. Re-attach the database in SSMS: Tasks → Attach → Add → select the restored .mdf. Confirm the .ldf path matches. Click OK.
7. Restart WinCC Runtime. Verify the trend window using WinCC OnlineTrendControl or the Taglogging editor's "Display" button.

Recovery Procedure B – Copy Segment Files into CommonArchiving

Used when the segment files exist on a peer server or duplicate project (e.g., Project2 on the same machine that was running during the outage). This is the documented Siemens approach for multi-segment runtime and external database coupling.

1. Verify both projects have identical configuration (same tag IDs, archive configuration, and segment sizes).
2. On the source project, locate the closed segment files (*.mdf / *.ldf) inside the ArchiveManager root. Do not descend into subfolders.
3. Copy these files into the destination project's CommonArchiving folder while the destination WinCC Runtime is active.
4. The ArchiveManager service detects the new files and automatically attaches them as additional segments. Trend windows across both segments return continuous data.
5. Clean up the copied files manually when the retention period expires; segments placed in CommonArchiving are not auto-deleted by the configured swap-out job.

Recovery Procedure C – Recreate Server Package and Rebuild Taglogging DB

Used when the database is so corrupted that SQL Server cannot attach it, or when the runtime reports "connection to server lost" in WinCC Control diagnostics. This procedure rebuilds the database files from scratch and relies on the backup folder for historical data.

1. On the server, open WinCC Explorer and stop the WinCC Server service.
2. Close WinCC Explorer on every client in the project. Wait for the WinCC Server service to reach "stopped" state.
3. Create a full project backup by copying the entire Project1 folder to Project1_BAK_<date>.
4. Delete or rename the following files inside the live project folder:
   • <Project1>Tlg.mdf
   • <Project1>Tlg.ldf
   • All files matching *.dc* in ArchiveManager subfolders
5. Open WinCC Explorer. Open the Server Data dialog and delete the existing server package for each client.
6. Regenerate the server package: right-click the server in the project tree, select "Server Data" → "Generate Server Package" for each client. Distribute the new package to the client machines via "Remote Desktop" or manually copy the resulting <Project>*.pck file.
7. Restart WinCC Runtime. The ArchiveManager creates fresh empty *.mdf / *.ldf files. Trend windows report empty for the rebuild window only.
8. Reapply Procedure A to merge any available backup segments back into the freshly created database.

Verification Checklist

Confirm recovery by running the following checks from the WinCC project tree before signing off the incident:

Prevention Strategy

Apply the following layered controls to eliminate recurrence.

1. UPS sizing – Size the UPS for at least 30 minutes at full server load (CPU + disk spin-up + monitor). Configure the bundled network shutdown agent to send a graceful Windows shutdown to each WinCC server when battery falls below 25%. Test the shutdown path quarterly by simulating mains loss.
2. Enable Archive Backup in both editors – Taglogging Editor → select archive → "Backup" tab. Set "Activate Backup" = yes, configure a destination path on a separate physical disk (not the same RAID volume as the live database), set schedule to daily at 02:00. Repeat for Alarmlogging Editor.
3. Stop Runtime cleanly during maintenance – Always use WinCC Explorer → "Stop Runtime" rather than logging off Windows or pressing the power button. Configure a Windows scheduled task that runs net stop "WinCC_Server" before any planned reboot.
4. Single project per machine – Do not duplicate the <Project> folder on the same machine. Use the WinCC Project Duplicator tool (or CCProjectMigrator.exe) for offline editing. Maintain only one active project per server with one autostart entry.
5. SQL Server health monitoring – Add a daily SQL Agent job that runs DBCC CHECKDB('<Project>Tlg') WITH NO_INFOMSGS and emails the operator on any reported corruption. Schedule outside production hours.
6. Disk redundancy – Place the live ArchiveManager folder on a RAID-1 or RAID-10 volume separate from the OS volume. Place the Backup folder on a separate physical disk or NAS share.
7. External archive access – For long-term analytics, configure the WinCC Archive Connector (part of DataMonitor) to expose archives through OLE DB to a reporting database. This provides an offline copy independent of WinCC runtime state. Refer to the official documentation: WinCC Archive Connector (RT Professional).

Diagnostic Flowchart

Field-Proven Pitfalls

• Detaching the live DB while Runtime is running produces "Archive connection lost" events for every connected client and may corrupt further segments. Always stop Runtime first.
• Copying segments into the live ArchiveManager root while subfolders are open causes the ArchiveManager to skip the file until the next service restart. Copy while Runtime is stopped or into CommonArchiving while Runtime is running.
• Mismatched segment sizes between two project copies produce silent rejection. Always verify the Taglogging editor → Archive Configuration → Segment size matches before performing a CommonArchiving copy.
• UPS without network shutdown agent – the batteries hold the server alive, but Windows never receives the shutdown command. When the batteries die, the OS crashes the same as an unprotected server.
• Deleting *.dc* files from a healthy database – these are compressed swap segments; deleting them removes compressed historical data that is not held in the main .mdf. Only delete when the goal is to rebuild the database from scratch (Procedure C).
• Network share for live archive – placing ArchiveManager on an SMB share introduces latency spikes that confuse the SQL Server recovery process after power-up. Always keep the live database on local direct-attached storage.

Standards and Reference Documentation

When planning recovery and prevention, cross-reference the following official Siemens documents:

• WinCC V7.3 SP3 Manual – "Working with Archives" section, covering Taglogging, Alarmlogging, and the ArchiveManager service.
• WinCC Information System help file WinCC_Information_System.chm – search keyword "CommonArchiving" for the segment reattach procedure.
• WinCC V7.3 Release Notes Update 7 – fixed bugs affecting archive attach after uncontrolled shutdown.
• Siemens Support Entry ID 109751706 – "What to do after a power failure with WinCC Runtime" (operator guide).
• WinCC DataMonitor Configuration Manual – Archive Connector configuration for OLE DB access: WinCC Archive Connector (RT Professional).
• SQL Server 2008 R2 / 2014 Books Online – "DBCC CHECKDB", "Database States", and "Attaching and Detaching Databases".

FAQ