High Availability and System Administration
High availability
ACOS high availability is built on VRRP-A, A10's enhanced VRRP implementation. Peers form a redundancy group identified by a VRID and a device-id, exchange hellos, and elect a master by priority; preemption decides whether a recovered higher-priority peer takes the role back. Floating addresses and VIPs move with the master. Active-standby is the usual design; active-active is possible with careful traffic symmetry. Configuration sync keeps peers consistent, and forced failover is used for maintenance and testing.
Administration
Routine tasks include saving the running-config, backing up configuration and certificates, upgrading ACOS, managing administrator accounts and partitions, and exporting logs through syslog or SNMP. NTP matters: certificates, logs, and troubleshooting all depend on correct time. The aXAPI provides programmatic access.
Traps
Questions target split-brain from a broken heartbeat link, mismatched VRIDs or device-ids, asymmetric routing that breaks stateful failover, and a failover that works but leaves clients with stale ARP entries. Also watch for configuration saved on one peer but never synchronized.
Rule of thumb
After any HA change, verify three things: VRRP-A state on both peers, the floating address answering from the client side, and synchronized configuration. Worked example: a planned upgrade is performed by forcing failover to the standby, confirming the VIP answers, upgrading the now-standby unit, then failing back in a maintenance window and confirming sync.