High Availability, Backup, and System Maintenance
Domain 7 tests resilience and upkeep. The exam expects you to know how FortiNAC-F HA behaves, what a backup actually contains, and which maintenance tasks prevent avoidable outages such as certificate expiry, disk exhaustion, or unsupported upgrade paths.
Core rules
- FortiNAC HA is active-passive. The pair needs matching hardware or virtual resources, the same firmware, a dedicated heartbeat connection, a virtual IP, and synchronized configuration. The secondary should not be configured as an independent appliance.
- Failover moves the VIP and promoted services to the standby. Test failover deliberately; an untested HA pair is only an assumption.
- Backups should include configuration and the database, not just appliance settings. Store backups off-box, schedule them, and take one before every upgrade or major policy change.
- Restore to a compatible version. Restoring an older database into a newer firmware can fail or corrupt state; follow the documented upgrade and restore sequence.
- Maintenance includes NTP health, disk space, log rotation, certificate renewal, license validity, and firmware upgrade planning. Read release notes for model and HA caveats.
- For HA upgrades, follow the documented sequence, typically upgrading the standby first, failing over, and then upgrading the former active node.
Worked example
Before a firmware upgrade, an engineer backs up the primary, verifies that the secondary is synchronized, and confirms both nodes are licensed and have disk space. The engineer upgrades the standby, tests failover, then upgrades the former active node. After both are current, the engineer checks HA sync, RADIUS authentication, and a guest portal login. The reasoning: the backup protects against rollback, while the standby-first sequence preserves service during the change window.
Traps
- Backing up only appliance configuration and omitting the database.
- Running HA nodes on different firmware or models.
- Restoring a backup across incompatible versions.
- Ignoring disk space or certificate expiry until services fail.
- Upgrading both HA nodes at once and creating an unnecessary outage.