A hospital ACI fabric loses every APIC. A new leaf no longer completes discovery. An engineer opens Nexus Dashboard and expects it to register that leaf. Why does the leaf still fail discovery?
Select an answer to reveal the explanation.
Short Explanation
You cannot ask the building's security cameras to issue new badges. Leaf discovery in an ACI fabric belongs to the APIC cluster. Nexus Dashboard can show you the fabric, and it does not take over that job when the APICs are gone.
Full Explanation
Each APIC discovers the leaf it is cabled to, then the spines off that leaf, then the other leaves. That registration is an APIC function. Nexus Dashboard hosts services such as fabric control for NX-OS fabrics and insights. It is not a standby APIC, and opening it does not answer the new leaf's discovery. The dashboard is not in the forwarding path, so it is not dropping LLDP from the leaf. The infrastructure VLAN is the fabric VLAN the APIC and the switches use for discovery. Putting it in a tenant pool is the wrong place for it and does not bring a dead APIC cluster back. Hardware proxy is how a bridge domain handles unknown unicast for endpoints. It has nothing to do with registering a switch. Caveat: Nexus Dashboard Orchestrator can push tenant policy to APICs that are already up. It still needs those APICs. Operational check: restore APIC reachability and confirm the new leaf appears for registration there, not only as a tile on the dashboard. Insights on the dashboard can explain a fabric that the APICs are still managing. They cannot invent a controller cluster that is powered off.