A customer’s infrastructure is set up to use both primary and secondary gateway clusters on the SSID profile cased on best practices. Why do they have an equal split of their 120 APs across the primary and secondary gateway clusters?

When a primary gateway cluster fails, all associated Access Points (APs) fail over to the secondary cluster. If cluster preemption is not enabled, these APs will not automatically return to the primary cluster even after it becomes available again. They will remain connected to the secondary cluster. The observed equal split of APs suggests that after the primary cluster was restored, a subsequent event (like a scheduled rolling reboot or a power outage affecting a subset of switches) caused half of the APs to restart. These rebooted APs, upon startup, would connect to their configured primary cluster, while the other half that did not reboot would remain on the secondary cluster, resulting in the 60/60 split.

A. The cluster type (heterogeneous) and node count influence AP load balancing within a single cluster, not the distribution between primary and secondary clusters.
B. If cluster preemption were enabled, all APs would have automatically moved back to the primary cluster once it was operational, leaving no APs on the secondary.
C. The cluster type (homogeneous) and node count influence AP load balancing within a single cluster, not the distribution between primary and secondary clusters.

1. ArubaOS 8.10.0.0 User Guide: In the section "Configuring a Cluster Profile," the cluster-preemption parameter is described. The guide states, "When this option is enabled, the APs will failback to the primary cluster when it is up." This directly implies that if the option is disabled (the default behavior), APs that have failed over will not automatically return to the primary cluster. This persistence on the secondary cluster is the necessary condition for the scenario described. (Reference specific to CLI commands under ap-cluster-profile).

2. Aruba Validated Solution Guide - Campus WLAN for Large Enterprises: In discussions on high availability and cluster design, these guides often recommend disabling preemption to prevent APs from "flapping" between clusters during intermittent primary cluster instability. It allows for a more controlled, manual, or scheduled return of APs to the primary cluster, which aligns with the scenario where a partial reboot would cause the observed split. (See sections on High Availability and Gateway Clustering).