Symptom
The APIC Cluster becomes unhealthy. APICs in the cluster will change to a Health State of: Data Layer Partially Degraded, Unknown, Data Layer Synchronization, or Fully Fit. Cluster communications are impaired. Leaf node switch or other node switches may became "inactive" status in the cluster.
Conditions
Enter VSH (Cisco iNX-OS Debug Shell) on a leaf node and enter the command "clear ip route vrf all *".
Workaround
To correct issue, reload the offending leaf node switch. On some occasions running a "setup-clean-config.sh" on the offending leaf node switch was necessary to fix the fabric heath.
