Cisco Cisco Aironet 1522 Lightweight Outdoor Mesh Access Point 디자인 가이드

When you are manually allocating IP addresses, Cisco recommends that you make IP addressing changes 
from the furthest mesh access point child first and then work your way back to the RAP. This also applies 
if you relocate equipment. For example, if you uninstall a mesh access point and redeploy it in another 
physical location of the mesh network that has a different addressed subnet.

Another option is to take a controller in Layer 2 mode with a RAP to the location with the misconfigured 
MAP. Set the bridge group name on the RAP to match the MAP that needs the configuration change. 
Add the MAP’s MAC address to the controller. When the misconfigured MAP comes up in the mesh 
access point summary detail, configure it with an IP address.

Despite the DHCP fallback mechanism, there is still a possibility that a mesh access point can become 
stranded, if any of the following conditions exist:

There is no DHCP server on the network.

There is a DHCP server on the network, but it does not offer an IP address to the AP, or if it gives a 
wrong IP address to the AP (for example, on a wrong VLAN or subnet).

These conditions can strand a mesh access point that is configured with or without a wrong static IP 
address or with DHCP. Therefore, it is necessary to ensure that when a mesh access point is unable to 
connect after exhausting all DHCP discovery attempts or DHCP retry counts or IP gateway resolution 
retry counts, it attempts to find a controller in Layer 2 mode. In other words, a mesh access point 
attempts to discover a controller in Layer 3 mode first and in this mode, attempts with both static IP (if 
configured) or DHCP (if possible). The AP then attempts to discover a controller in Layer 2 mode. After 
finishing a number of Layer 3 and Layer 2 mode attempts, the mesh access point changes its parent node 
and re-attempts DHCP discovery. Additionally, the software exclusion-lists notes the parent node 
through which it was unable to obtain the correct IP address. 

Depending on the mesh network design, it is entirely possible that a node finds another node “best” 
according to its routing metric (even recursively true), yet it is unable to provide the node with a 
connection to the correct controller or correct network. It is the typical honeypot access point scenario 
caused by either misplacement, provisioning, design of the network, or by the dynamic nature of an RF 
environment exhibiting conditions that optimize the AWPP routing metric for a particular link in a 
persistent or transient manner. Such conditions are generally difficult to recover from in most networks 
and could blackhole or sinkhole a node completely, taking it out from the network. Possible symptoms 
include, but are not limited to:

A node connects to the honeypot, but cannot resolve the IP gateway when configured with static IP 
address, or cannot obtain the correct IP address from DHCP server, or cannot connect to a WLAN 
controller.

A node ping-pongs between a few honeypots or circles between many honeypots (in worst-case 
scenarios).

Cisco mesh software tackles this difficult scenario using a sophisticated node exclusion-listing 
algorithm. This node exclusion-listing algorithm uses an exponential backoff and advance technique 
much like TCP sliding window or 802.11 MAC.