FortiManager 7.6: Revision Control and Safe Rollback
Revision control matters in FortiManager because a policy installation can affect dozens or hundreds of FortiGate devices. A useful revision history answers who changed what, why it changed, which device set received it, and how operations could restore the prior behavior. Merely saving a configuration is not a dependable rollback strategy unless the dependencies and observed traffic state are understood. Fortinet’s current 7.6 Administrator assessment has moved to the NSE 6 exam pathway; product administration concepts remain essential.
On this page
Create a revision before a risky operation
Record the starting policy package, relevant shared objects, administrator, change ticket and device groups. Capture the package’s installation status and any device-local differences. A good revision label tells reviewers what decision changed—such as a new restricted application route—not just “backup 12.” Lock or coordinate concurrent edits so two administrators do not unknowingly approve incompatible versions.
Compare proposed changes with the last installed state
Use available comparison and install-preview mechanisms to distinguish the working package from what devices currently enforce. Review rule sequence, changed services, source and destination scope, object definitions, new logging and any exceptions. A benign-looking shared-object change may expand access across several packages. Separate formatting or naming changes from security-significant actions when explaining the diff to an approver.
Plan a controlled installation
Stage the change on a representative subset if the operating model permits it. Define what success looks like in network terms: expected connections work, forbidden paths stay blocked and critical monitoring continues. Record installation errors separately from traffic failures, because a configuration can install successfully yet still violate service intent. Keep the approved change window, contacts and reversal criteria visible to the operator.
Revert safely when behavior deteriorates
Identify the prior known-good revision and verify whether a rollback will also undo important intervening local or emergency changes. Confirm dependent objects still exist and device versions support the target policy. Restore the intended package under controlled approval, reinstall where necessary and test the original failed service. If an incident is still active, contain risk first; blindly replaying an old configuration can create a second outage.
Make the evidence useful later
Preserve revision identifiers, device list, installation result, validation tests and operator notes in the change record. After recovery, determine whether the failure came from intent, an overlooked dependency, a software incompatibility or a review gap. Change the test or approval procedure accordingly. A reviewer should be able to reconstruct the blast radius without relying on a single administrator’s memory. Refer to Fortinet’s current FortiManager documentation and exam scope when validating product-specific behavior.