Need help again, cannot get 1130 APs to join

[gary.hostetler]

They see the controller, it downloads the software but then on reboot I get these errors:

LWAPP_CLIENT_ERROR_DEBUG: spamHandleJoinTimer: Did not rec

ieve the Join response

*Mar 1 00:00:58.622: LWAPP_CLIENT_ERROR_DEBUG: No more AP manager IP addresses

remain.

*Mar 1 00:00:58.655: %SYS-5-RELOAD: Reload requested by LWAPP CLIENT. Reload Re

ason: DID NOT GET JOIN RESPONSE.

I cannot ping the ip address of the ap-management port, maybe I'm not supposed to. This is hooked into a 4506 core switch and I have it set to Dynamic Auto on the port. At one point I could get the AP's that were connected to the 4506 to stay associated but not the ones outside the switch.

thanks

Gary

[Scott Fella]

So you have the WiSM module on the 4506? If not, what WLC do you have? How many ap's have joined the WLC? The error looks like you maxed out the port.

You can't ping the ap-manager interface only the management interface. Is this a new install or you are adding to an existing?

[gary.hostetler]

I'm not sure what a WiSM module is but I get 802.q choices on the vlan on the gig port that I am connected to. I have added a 2nd controller which I want to migrate everyone to but I do not want to have to go to all the AP's and associcate with the new name.

I actually had several of the 1130 AP's that found the controller but they were all hooked into the the 4506. None of the ones outside of that hooked in. They all have green on the circle of the AP.

I guess what I really want to do is get them all on the new one and see what happens.

I do have the first contrller in Layer 2 at the moment because in Layer 3 one of the APs cannot see the controller.

I have the 2nd one up in Layer 3 and it sees the 4 AP's but when a laptop connects it cannot get an ipaddress from the AP.

thanks for any help, pretty desparate here.

Gary

[gary.hostetler]

I fixed the laptop not getting an IP address from the new controller that is in layer 3 as I had the wrong IP address in for the dhcp server so the laptops were getting the wrong IP. So if I can get the currrent laptops to look at the new controller without touching each one I get can get this puppy out of the way.

I have enough troble figuring out one controller let along 2.

Gary

[Scott Fella · Accepted solution]

What are you consoled into to configure the controller? What model controller do you have?

Well it looks like you have things setup wrong. This is what I would do first. Work on one controller at a time. Take off the second controller.

1) You need to have the management and ap-manager on the same subnet and vlan tagged as '0' not as what vlan they are on.

2) They will be 802.1q trunk on the switch with native vlan (whatever your management vlan is)

3) Configure the controller for layer 3

4) Configure DHCP on the switch so that the ap's get an IP

ip dhcp excluded-address 10.100.6.1 10.100.6.50

!

ip dhcp pool AP_Staging

network 10.100.6.0 255.255.255.0

default-router 10.100.6.1

option 60 ascii "Cisco AP c1130"

option 43 ascii "IP address of the management interface"

5) Configure the service port to an ip address that is not on any subnet and don't plug thi into your network. Use a 192.168.x.x or just 1.2.2.1. Service port has to be configured if you have a WiSM!

6) When ou create an ssid in the WLAN, make sure you identify which interface the ssid will be mapped to, or else it will be on the subnet your management interface is on.

7) On the interface you created for the ssid, that is where you specify the DHCP address for that interface.

Then let me know how things are after that, because there is more!

[gary.hostetler]

I have the 4402 for 50 controllers. I am using the web based gui to get into the controllers. Management and ap are on the same subnet and both are tagged 0. I have 802.1q on the 4506 gig port and and I am using Network Assitant for that. It is layer 3.

When you say configure dhcp on the switch do I need that since I have a Microsoft dhcp server sitting on my network. Or is this to use only with 4402.

Service port is 192.168.1.2

The access points get an ip but from the real network.

Interesting stuff.

Gary

[Scott Fella]

The access points if in a different subnet from the management and ap-manager, then you need to configure option 60 & 43 on the dhcp server itself. this way the accesspoint knows the IP of the controller. Or you can use DNS to point the ap's to the controller. Here is a good doc that might explain it better. If you put the ap's on the same subnet... by just changing the vlan on the switch port and then doing a shut and no shut on the port to bounce the ap. As long as the ap gets an ip address on that subnet, then the ap's will find the controller. You then can switch the port back to whatever vlan and then bounce the port. The ap will store the ip oif the controller and you will be good to go. Or else you can look at this doc.

http://www.cisco.com/en/US/tech/tk722/tk809/technologies_tech_note09186a00806c9e51.shtml#topic2

[Scott Fella]

Also make sure you configure LAG on the controller and that you etherchannel all the ports to the switch. You only get 25 ap's per port if I remember correctly.

[gary.hostetler]

All is well but I'm not sure why. I have lwapp-cisco-controller pointed in dns to my WLC, I set the LAG to enabled and I rebooted the WLC and every stinikin on of them show up immmediately. I had the 4 1130s sitting on the 2nd controller but it was set for layer 3 adn the 8 1010s were on the first controller. I shut down the switch port on the 2nd controller and rebooted the first one into layer 3 and all but one 1010 and all the 1130s showed up in less than a minute with the REG status.

I have one 1010 that will not hook into layer 3 but I can live with that.

I even bumped the remaning controller's switchport and all 11 came right back up.

I'm not even using the dhcp 43 option.

thanks a million

Gary

[Scott Fella]

when lag is set to enabled, a reboot is required. as long as the ap and the ap-manager in the same subnet, only an ip is required for the ap. DNS and option 43 is requied if the ap is on a different subnet.

[gary.hostetler]

I am assuming that subnet is the building subnet as in 10.228.x.x which the subnet of the whole building. Now with that in mind how do I get the 2nd controller up so that it's APs are the same group as the first controller so that I can move all the APs over and then shut down the first one.

Also what is LAG and what does it do for me?

Gary

[Scott Fella]

Gary,

The 10.228.x.x is the whole building, but can be broken down into more subnets which requires L3 routing between subnets.

Example:

Vlan 2 10.228.2.x /24

Vlan 3 10.228.3.x /24

Vlan 4 10.228.4.x /24

Vlan 5 10.228.5.x /24

.

.

.

Vlan 254 10.228.254.x /24

Vlan 255 10.228.255.x /24

Vlan 2 (Management-Network Devices, Controllers and AP's not more than 20 ap's per subnet)

Vlan 3 (Wireless Data)

Vlan 4 (Wireless Voice)

Vlan 5 (Wired Data)

Vlan 6 (Wired Voice)

Vlan 254 (Wireless Guest)

You get the idea.....

LAG is Link Aggression and should not be used unless both gig ports are connected to the switch via etherchannel.

[gary.hostetler]

I don't need to get that fancy yet on the VLANS. I only have about 12 APs on there now but will be increasing to 40-50. I will be adding a 2nd gig etherchannel tomorrow so I will leave LAG on.

thanks

Gary

[Scott Fella]

Just in case you have issues connecting to the controller via the GUI, it is because you what an interface on the controller that is also on the same subnet as users on the wired LAN. Just an fyi.....