How Is a Server Learned on a vPC Peer Leaf in Cisco ACI?

Answer Correct answer: D — Leaf2 learns S1 as an on-peer endpoint because S1 is attached to Leaf1 through a port channel while Leaf1 and Leaf2 are vPC peers.

An engineer configures two Cisco ACI leaf switches called Leaf1 and Leaf2 as VPC peers. The server called S1 is connected to Leaf1 with a port channel interface. How is S1 learned in the endpoint table of Leaf2?

  1. remote endpoint
  2. bounce entry endpoint
  3. local endpoint
  4. on-peer endpoint Correct Answer

Community Votes

D
67%
A
33%

67% of anonymous learners picked answer D. Votes are pick records left by other test-takers — they are not the verified answer.

Community Insight

The question tests whether you can separate local, remote, on-peer, and bounce endpoint types in the ACI endpoint table; the trap is assuming that any endpoint not learned locally by Leaf2 must simply be a remote endpoint.

When server S1 attaches to Leaf1 through a non-vPC port channel while Leaf1 and Leaf2 are vPC peers, Leaf2 programs S1 in its endpoint table as an on-peer endpoint rather than a local or plain remote entry. This page confirms why option D is the correct 300-620 answer and how ACI classifies synchronized vPC endpoint entries.

Choosing A (remote endpoint) because Leaf2 never saw S1's traffic in its own data plane — but entries synced from a vPC peer leaf are specifically flagged as on-peer, a distinct subtype of remote.

Community Discussion (3 comments)

prospio971 👍 2 Selected: A
It is A. In a Cisco ACI fabric with vPC configured between two leaf switches (Leaf1 and Leaf2), when a server (S1) is connected to one of the leaf switches (e.g., Leaf1), the endpoint learning process works as follows: Local Endpoint Learning on Leaf1: Since S1 is directly connected to Leaf1, its MAC and IP addresses are learned as local endpoints on Leaf1. Remote Endpoint Learning on Leaf2: In a vPC configuration, the two leaf switches (Leaf1 and Leaf2) operate as vPC peers but do not share a dedicated peer link. Instead, the ACI fabric itself acts as the medium for communication between the peers. Leaf1 communicates the endpoint information (MAC and IP of S1) to the spine switches using the COOP (Council of Oracle Protocol) database. The spine then propagates this information to Leaf2. On Leaf2, S1 is learned as a remote endpoint because it is not directly connected to Leaf2 but is reachable through its vPC peer (Leaf1).
prospio971 👍 1 Selected: D
When two Cisco ACI leaf switches (Leaf1 and Leaf2) are configured as vPC peers, and a server (S1) is connected to Leaf1 via a port channel, S1 is learned in the endpoint table of Leaf2 as an "on-peer endpoint."
mdriraa 👍 3 Selected: D
The on-peer endpoint is a variant of the remote endpoint. These endpoints are remote endpoints that point to a port that is not part of a virtual Port Channel (vPC), also called an orphan port, on a vPC peer leaf. They are special endpoints because they are remote, but they are learned through vPC synchronization on the control plane instead of through data-plane learning from the actual traffic. As a result of vPC synchronization, on-peer endpoints have MAC address, IP address, and EPG information, unlike other remote endpoints, which have either bridge domain (MAC address) or VRF (IP address) information.

Comments & Corrections

No comments yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Submitted for moderation before publishing. Keep it helpful and respectful.

Expert Analysis

Why the Answer Is Correct

Leaf1 and Leaf2 form a vPC pair, yet S1 is attached only to Leaf1 through a port channel, so from the fabric's perspective S1 sits on a non-vPC (orphan-style) port of a vPC peer leaf. Leaf1 learns S1 locally in its own endpoint table through data-plane learning, then synchronizes that MAC and IP address to Leaf2 across the vPC peer link using COOP control-plane messaging. Because the entry arrives from a vPC peer instead of from an unrelated remote leaf, APIC records it on Leaf2 as an on-peer endpoint — the special synchronized entry that points back to a port on the peer leaf. That on-peer label is exactly the state the question describes, which makes option D the correct classification over the plain remote endpoint option.

Why the Other Options Are Wrong

A (remote endpoint) is the deliberate distractor: remote entries are learned from leaf switches that are not vPC peers, and the scenario explicitly states that Leaf1 and Leaf2 are vPC peers, which is precisely why the on-peer subtype applies instead. C (local endpoint) would require S1 to be attached to Leaf2 itself, either directly or through a member link of a vPC port channel spanning both leaves, but the question says the port channel is on Leaf1 only. B (bounce entry endpoint) describes a transient entry created after an endpoint moves or its MAC flaps so traffic is not black-holed while it is re-learned; nothing in the scenario indicates a move, so it cannot be the steady-state classification for S1.

Community Comment Notes

Most learners who picked D follow mdriraa's reasoning that the on-peer endpoint is "a variant of the remote endpoint", reserved for entries learned through "vPC synchronization on the control plane" rather than data-plane learning, which matches S1 sitting on a non-vPC port of a vPC peer. prospio971 first argued for A as a remote entry, then posted the D explanation himself, showing how easily the remote versus on-peer distinction trips people up when they focus only on the fact that Leaf2 did not learn S1 locally. Roughly a third of voters stayed with A because a synchronized entry looks remote at first glance; the deciding detail in the stem is the explicit vPC peering between Leaf1 and Leaf2.

Exam Strategy

Before answering, map each option to a one-line definition: local = attached to this leaf (or a member link of a vPC spanning both leaves), remote = learned via COOP from a non-vPC-peer leaf, on-peer = synchronized from a vPC peer, bounce = transient entry after a move or flap. Underlining the phrases "vPC peers" and "connected to Leaf1" in the stem usually collapses the choice to on-peer within seconds.

Frequently Asked Questions

Why is S1 not simply a remote endpoint in Leaf2's endpoint table?

Plain remote entries come from leaf switches that are not vPC peers; because Leaf1 is Leaf2's vPC peer, the synchronized entry is flagged on-peer, a special variant of remote.

When would S1 show up as a bounce entry endpoint instead of an on-peer endpoint?

A bounce entry is a temporary entry created after an endpoint move or MAC flap to avoid black-holing traffic until the endpoint is re-learned; S1 is stable here, so it is not bounce.

Related Analysis

← Back to 300-620 Study Guide