Header Only - DO NOT REMOVE - Extreme Networks

C25 Controller Not Seeing APs Online


I have a C25 Wireless controller that is not seeing APs online and giving them their networks. All APs are powered by the switch and have power, and the Domain Controller is finding all APs through DHCP and giving them an address, but for some reason the controller is not recognizing them as online. The APs are a mixture of 3825i and 3935i APs. I have tried rebooting the Controller and that did not solve the issue. Again, the APs are getting power and a IP address from the Domain Controller, I can ping the APs and get a reply, but the Controller is not picking them up.

EDIT: These APs have also already been joined to the controller. We had a power outage over the weekend and only one AP came back on the controller. Switch is working fine, just APs not joining to their controller.

6 replies

Userlevel 5
Did somebody accidentally disable AP-registration on the ESA port of the controller that the AP's connect to?
Not that I am aware of. The Controller never turned off with the power outage and was never altered.
Userlevel 3
This article (see below) provides the best information we have regarding this scenario (power outage/AP discovery) and the possible reasons your APs may not reconnect. Reviewing this document may highlight one of the possible reasons at play in your environment:

https://gtacknowledge.extremenetworks.com/articles/Solution/After-power-loss-an-Access-Point-no-long...
Userlevel 3
Hawkins, Bruce wrote:

This article (see below) provides the best information we have regarding this scenario (power outage/AP discovery) and the possible reasons your APs may not reconnect. Reviewing this document may highlight one of the possible reasons at play in your environment:

https://gtacknowledge.extremenetworks.com/articles/Solution/After-power-loss-an-Access-Point-no-long...

Here's the most relevant information from that document:

CauseIf an Access Point loses power (outage, switch upgrade, configured in staging area and powered down to be deployed, etc) the acinfo.conf file that contains the last known Controller IP address is removed on boot up. In this situation, the AP may not be able to get back to it's Primary Controller if the IP address on the AP is on a different subnet than the Controller and:

  • There is no static IP address set
  • The DNS options for the Controller IP have changed or are not resolvable on the AP subnet
  • DHCP options have changed or are incorrect
If the AP is in the same subnet, it will resolve the controller address through SLP. If the AP reboots, but doesn't power down, it will keep it's acinfo.conf file, and that has the last known AC (access controller) IP.


ResolutionTo avoid this, the recommended configuration is to set the controller IP's in the Controller Search List in the AP configuration screen once it is on it's approved controller
How to assign a static IP address to an IdentiFi Access Point over the EWC WebGui
This can also be done via the AP CLI:
How to Assign a Static IP Address to an Access Point via AP Shell

A new DNS entry or DHCP option can be set.
How To Configure an AP to find the IdentiFi Wireless Controller from a DNS server entry.
How To Configure an AP to Find the IdentiFi Wireless Controller with DHCP Option 78 on a Windows Server
Userlevel 5
I would verify, also are the AP's in the same vlan as the controller? If not is the default route configured in the controller? If you take a trace on the ESA port are you seeing the AP's reaching the controller and does the controller respond?
Thank you both for your help. I believe it was kind of related to what Bruce mentioned. We have our APs on different scopes, but have them finding the controllers using scope options. After Bruce linked that article I decided to take a look there. Nothing seemed wrong with any of the set up but I decided to erase the scope option for that scope and add it back in from scratch. As soon as I did that the Controller picked everything back up again. Not sure how the DHCP Server and Controller were having issues as neither of them were effected by the power outage since they run on backups in that event. What ever the reason was, that seemed to fix it.

Reply