The correct answer is A. Interface FastEthernet0/0 is configured in VRF CUST-A, so the ARP entry is also in that VRF.
In Cisco VRF implementations, ARP entries are maintained separately per VRF, not in the global routing table. Cisco documentation states that with Multi-VRF CE, ARP entries are learned in separate VRFs, and the user can display them for a specific VRF. Cisco also documents the command show ip arp vrf < vrf-name > to display the ARP table for a specified VRF.
That means if the interface belongs to VRF CUST-A, a normal global show ip arp 192.168.12.2 may not display the entry, even though ping vrf CUST-A 192.168.12.2 succeeds. The ARP entry exists, but it is stored under the CUST-A VRF ARP table, not the global ARP table.
B. When VRFs are used, ARP protocol is disabled in the global routing table. This is incorrect. ARP is not disabled; it is simply VRF-aware and tracked separately per VRF.
C. When VRFs are used, ARP protocol must be enabled in each VRF. This is incorrect. There is no separate requirement to manually enable ARP in each VRF for normal operation in this context.
D. The ping command must be executed in the global routing table. This is incorrect because Cisco supports pinging within a specific VRF, and the exhibit already shows the use of ping vrf CUST-A .
ENCOR exam point:
With VRF-aware services, always remember:
Submit