Question 560
A fabric uses two equal-cost underlay next hops yet packets continue toward a failed path following a port outage. Which initial check would most directly distinguish the likely cause?
Explanation
Correct Answer
This checks the relevant evidence first: check interface state, routing convergence and forwarding entries following the simulated spine-link loss. The underlying concept is that path redundancy only improves availability when failed underlay next hops are withdrawn or detected promptly. After confirming the finding, a controlled remediation is to correct underlay failure detection and routing convergence within the approved topology and timer policy.
Incorrect Answers
The option “Inspect installed ECMP next hops and compare per-link counters across multiple distinct flows” is useful in a different setting: An operator finds two equal-cost fabric paths but a single large traffic flow consistently traverses only one link. The case being tested is: A fabric uses two equal-cost underlay next hops yet packets continue toward a failed path following a port outage. The relevant inspection is: Check interface state, routing convergence and forwarding entries following the simulated spine-link loss. The most discriminating first check is “Check interface state, routing convergence and forwarding entries following the simulated spine-link loss”.
“Inspect routing tables and underlay advertisements for each intended VTEP address” is relevant to a different condition: A newly installed rack can exchange underlay packets but VTEP loopback routes are missing from the remote leaf routing tables. The stated question instead concerns: A fabric uses two equal-cost underlay next hops yet packets continue toward a failed path following a port outage. The relevant inspection is: Check interface state, routing convergence and forwarding entries following the simulated spine-link loss. The most discriminating first check is “Check interface state, routing convergence and forwarding entries following the simulated spine-link loss”.
“Inspect underlay routing adjacency, interface addressing and reachability to the spine or leaf loopback endpoints” is associated with a distinct issue: A data center team adds a leaf switch but the new node cannot reach any spine loopback addresses although local links are up. By contrast, the actual problem is: A fabric uses two equal-cost underlay next hops yet packets continue toward a failed path following a port outage. The relevant inspection is: Check interface state, routing convergence and forwarding entries following the simulated spine-link loss. The most discriminating first check is “Check interface state, routing convergence and forwarding entries following the simulated spine-link loss”.