VMware NSX
This situation can arise if there are multiple movements of the impacted IP address between 2 VMs. For example, in case of frequent failover between VMs when the 2 VMs are in high availability and the impacted IP address is a VIP address.
This is a rare corner case triggered by an upstream data pattern originating from CCP.
Workaround:
Restart cfgagent service on the host:
/etc/init.d/nsx-cfagent restart