Inventory recoverable assets
For every line or facility, list PLCs, RTUs, HMIs, SCADA, historians, engineering workstations, drives, robots, safety systems, network devices, time and licence servers. Record owner, model, serial number, firmware, engineering software, project location and criticality.
The inventory must show not only that a copy exists, but which version matches the equipment currently running.
Capture the complete dependency set
Retain source projects, device uploads, libraries, firmware, parameters, recipes, certificates, keys and instructions. Secrets are stored separately and securely but have a coordinated recovery process. Legacy equipment may need an engineering-workstation image, installers and a physical adapter.
Create a new verified recovery point after every approved change, or the archive will quickly diverge from production.
Select a safe collection method
Active discovery and bulk polling can affect sensitive equipment. Agree methods with process owners and vendors, use maintenance windows and avoid untested agents on controllers. Where possible, collect through an engineering station or approved gateway.
Copies move to a separate storage domain with restricted access, immutable versions and an additional offline or off-site layer.
Connect backup to change management
A change record contains the asset, previous and new versions, reason, author, approval, validation and rollback plan. Project hashes and upload times establish correspondence with the archive. An unauthorized difference between device and repository should trigger investigation.
Test recovery in a safe environment
The exercise prepares compatible hardware or an emulator, loads firmware and logic, restores communications and validates I/O, interlocks, safe state and the process scenario. For network devices, verify addresses, VLANs, policy, time and redundancy.
OT, operations and safety owners jointly authorize return to production. An unknown copy should never be tested for the first time on a running line.
Measure actual readiness
Track coverage, age of the latest validated copy, version drift, test success, observed recovery time and availability of people, software and cables. A periodic exercise selects a random critical asset and completes the path to a working test system. That turns an archive into a recovery capability rather than a collection of files.