Controllerless Networks

Hello,

We are running into issues where IAPs that were installed years ago were never properly joined to the working cluster so they were never upgraded. This is causing major problems after recent network refreshes that are putting those old IAP into the correct VLAN, which is allowing them to try and join the cluster. 

I am parsing through the syslog data for one of the clusters and I see this trap. Not sure if this is the smoking gun I need to use to ID other clusters that may have this issue as well.

SNMP Trap Category is Hardware or SNMP Trap Category is Software (Normal) System Alert

Doesn't look like anything too descriptive. If you are worried about APs from other clusters joining the cluster on the correct VLAN, you could turn off the Auto-Join setting in the clusters. This way you would have to manually allow an AP to join the IAP cluster. 

We have entertained that idea as well. With the most recent scenario we had 4 IAP that were on really old code that never actually joined the cluster fully. If you tried to browse to the IP of one of these old IAPs it would redirect you to the VC VIP. In other instances one of these old APs will get put on the network for a variety of reasons(we are trying to eliminate them) and it will cornhole the whole cluster by downgrading it and wiping the config.

I would prefer turning off auto-join and that may be the ultimate solution. What I am trying to do now, is detect any live situations where we may have an AP somewhere that is not fully joined to the cluster because of a firmware mismatch. I was hoping the VC would send out a syslog message or something saying "Hey I have a problem child over here"