F5-CSE study materials
F5-CSE study resources

F5-CSE: High Availability, Synchronization, and Operational Maintenance

F5-CSE · Work through your exam, one topic at a time.

All guide sections

High Availability, Synchronization, and Operational Maintenance

Device service clustering

The HA domain tests how two or more BIG-IP devices share configuration and traffic. A trust domain establishes secure communication. A device group defines which devices sync configuration; sync-failover groups also fail over traffic groups. A traffic group owns floating self IPs and optionally a MAC masquerade address. Active/standby is the common model; active/active requires careful traffic-group design.

Synchronization and failover

Configuration sync can be full or incremental. tmsh run cm config-sync to-group <group> pushes the local configuration to the group. Synchronization is not the same as failover: a device can be in sync but standby, or active but out of sync. Failover can be triggered by network failover heartbeats, VLAN failure, gateway ICMP, hardware, or daemon failure. Network failover can use multicast or unicast. Auto failback returns traffic when the preferred device recovers.

Maintenance

Take a UCS backup before changes. tmsh save sys config writes running configuration to the saved configuration. Use qkview and iHealth for support. Upgrades should account for device-group compatibility and the traffic-group state; licensing and platform-specific settings are not ordinary synchronized configuration.

Worked example

The active device loses its uplink and the standby does not take over. tmsh show cm failover status shows both devices healthy from the HA perspective. The failed VLAN has VLAN failover disabled, and network failover is not configured, so the standby has no signal that the active is impaired. Enabling VLAN failover on the relevant VLAN or configuring network failover with a reachable peer address gives the standby a detectable failure condition. Gateway ICMP can also be used where appropriate.

Traps

Treating config sync as a backup. Assuming a standby device is automatically in sync. Forgetting that floating self IPs and MAC masquerade belong to traffic groups. Testing failover without checking whether the trigger is enabled. Making changes on both devices independently and then forcing a sync in the wrong direction.