12-31-2023 09:01 PM
Why does ACI track loose nodes?
Fabric > Inventory > Fabric Membership > Unmanged Fabric Nodes
What is ACI doing with these, that it needs to treate these devices with a level of magnificence greater than any other connected, non-fabric device?
weylin
01-02-2024 11:30 AM
Unmanaged Fabric Nodes are devices that ACI detects via CDP/LLDP and create entries using their management IP address. It's not that it's detected as a Fabric Node, the name lets you know that it's not managed by the APIC. It's an outside device that ACI needs to be aware of.
In the case that you have a hypervisor behind those unmanaged devices, ACI should still be able to dynamically program VLANs as needed. This is the single hop away rule for VMM domains. It's the same if it was a blade switch chassis like a UCS Fabric Interconnect.
The only way to remove the unmanaged device would be either to unplug it from the fabric and wait a certain timeout period OR disable LLDP/CDP on the interface policy group related to the connected device.
01-04-2024 12:33 PM
thanks @Ruben Cocheno . Is it only for support to VMM Integration? I'm not using that, but I can concede the point there. What rules are used to decide whether a device gets listed as an unmanaged node? It's clearly not simply that it's detected in CDP/LLDP, I have a plethora of devices that participate in discovery protocols but are not listed as unmanaged nodes. And I have a number of nodes that, to me at least, clearly never could be VMM-managed (Dell switches; FWs; pre-UADP Catalysts; etc) but show up in the list.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide