Access Points Joining Random WLC

jake.stabl · ‎10-13-2010

My environment contains 4 seperate controllers. 1 controller in the datacenter which remote offices use with HREAP and discover this WLC using DNS We will call this Controller WLC1. 2 more controllers (WLC 2 and WLC 3) in the corporate headquarters to support >40 access points in a N+1 configuration and finally an anchor controller which doesnt matter in this post.

What I am experiencing is Access points in the Corporate office are using DHCP option 43 to discover the WLC2. every night or after long periods of time a random number of my access points are either rebooting or most likely sleeping. Then are attaching them selves to WLC 1 which is in the datacenter and is not supporting DHCP from that location which to clients make the wireless network seem to be down.

I think that with using a hybrid of DHCP and DNS is partly my issue but it shouldnt be. I have primed all the Access point to look at primary controller WLC1 and secondary controller WLC2 both by name and using the IP address of the respective management ports.

Can there be any explaination for this and why this seems to be so random?

WLCs are 5508

Corporate Access points are 1142

***EDIT

After some more investigation this isnt completely random rebooting the access points in question are connected to a new 4506-E chassis with the Sup 6-E while the other access points are connected to a 6509-E.. not sure how much this matters but I am seeing in the access point logs that about every 2-3 hours they are accessing DHCP again and sometime not successfully. Is there some EnergyWise or power savings happening on the 4500 possibly??

George Stefanick · ‎10-13-2010

yea, have you assigned first, second and third priority to your access points? Access points will fall back to other controllers if you dont "assign" them to controllers.

Depending on the code you're running. Drop into the GUI hit wireless--> select AP --> Look for this

Primary Controller Name
Secondary Controller Name
Tertiary Controller Name

"Satisfaction does not come from knowing the solution, it comes from knowing why." - Rosalind Franklin
___________________________________________________________

Nicolas Darchis · ‎10-13-2010

another nice tip : if your WLCs are in a mobility group, the APs will learn the ip addresses of the other controllers of the mobility group once they joined one WLC.

Nicolas.

jake.stabl · ‎10-14-2010

I understand that all and here is a screen shot (see attached) of an ap that was on WLC 2 last night at midnight when i wrote this orginally and now how jumped to WLC 2.. The major issue here is that WLC 1 does not give out DHCP because it is used for HREAP and remote office that have other local DHCP servers.

I am running code 6.0.199.4

Now yes all controllers WLC1,2,and3 are in the same mobilty group. Maybe I should move the WLC1 out of the mobility group and that might solve the issue. Here is a log too. See timing is 3 hours and then its trying to find another controller. I see a retransmission count exceeded here. What can cause that? I have 4 port LAG configured on both WLC2 and WLC3 and they are both on seperate 6500 switches. This access point is on a seperate floor connected to a 4500 with uplink to each of the 6500's

**********LOG from Access Point***********

*Oct 14 02:41:21.000: %CAPWAP-5-DTLSREQSEND: DTLS connection request sent peer_ip: 10.104.210.21 peer_port: 5246 This Is correct
*Oct 14 02:41:21.000: %CAPWAP-5-CHANGED: CAPWAP changed state to
*Oct 14 02:41:21.826: %CAPWAP-5-DTLSREQSUCC: DTLS connection created sucessfully peer_ip: 10.104.210.21 peer_port: 5246
*Oct 14 02:41:21.827: %CAPWAP-5-SENDJOIN: sending Join Request to 10.104.210.21
*Oct 14 02:41:21.827: %CAPWAP-5-CHANGED: CAPWAP changed state to
*Oct 14 02:41:21.911: %CAPWAP-5-CHANGED: CAPWAP changed state to CFG
*Oct 14 02:41:22.047: %CAPWAP-5-CHANGED: CAPWAP changed state to UP
*Oct 14 02:41:22.052: %LWAPP-3-CLIENTEVENTLOG: Received AP Syslog IP Address(255.255.255.255) configuration.
*Oct 14 02:41:22.173: %CAPWAP-5-JOINEDCONTROLLER: AP has joined controller nwcp01-1211-02b
*Oct 14 02:41:22.254: %LWAPP-3-CLIENTEVENTLOG: SSID Roswifi added to the slot[0]
*Oct 14 02:41:22.255: %DOT11-4-NO_HT: Interface Dot11Radio0, Mcs rates disabled on vlan 1 due to WMM is not enabled
*Oct 14 02:41:22.257: %LWAPP-3-CLIENTEVENTLOG: SSID RosGuest added to the slot[0]
*Oct 14 02:41:22.259: %LWAPP-3-CLIENTEVENTLOG: SSID Roswifi added to the slot[1]
*Oct 14 02:41:22.261: %LWAPP-3-CLIENTEVENTLOG: SSID RosGuest added to the slot[1]
*Oct 14 02:41:22.263: %DOT11-4-NO_HT: Interface Dot11Radio1, Mcs rates disabled on vlan 1 due to WMM is not enabled
*Oct 14 02:41:22.263: %DOT11-4-NO_HT: Interface Dot11Radio1, Mcs rates disabled on vlan 2 due to WMM is not enabled
*Oct 14 02:41:22.264: %DOT11-4-NO_HT: Interface Dot11Radio0, Mcs rates disabled on vlan 2 due to WMM is not enabled
*Oct 14 02:41:22.274: %WIDS-6-ENABLED: IDS Signature is loaded and enabled
*Oct 14 05:57:40.164: %CAPWAP-3-ERRORLOG: Retransmission count for packet exceeded max(CAPWAP_ECHO_REQUEST
., 3)
*Oct 14 05:57:40.164: %CAPWAP-3-ERRORLOG: GOING BACK TO DISCOVER MODE
*Oct 14 05:57:40.164: %DTLS-5-SEND_ALERT: Send WARNING : Close notify Alert to 10.104.210.21:5246 *** This is correct
*Oct 14 05:57:40.206: %WIDS-6-DISABLED: IDS Signature is removed and disabled.
*Oct 14 05:57:40.207: %CAPWAP-5-CHANGED: CAPWAP changed state to DISCOVERY
*Oct 14 05:57:40.207: %CAPWAP-5-CHANGED: CAPWAP changed state to DISCOVERY
*Oct 14 05:57:50.229: %CAPWAP-3-ERRORLOG: Go join a capwap controller
*Oct 14 05:57:49.000: %CAPWAP-5-DTLSREQSEND: DTLS connection request sent peer_ip: 10.98.10.20 peer_port: 5246 **This is the controller it should never get on.
*Oct 14 05:57:49.000: %CAPWAP-5-CHANGED: CAPWAP changed state to
*Oct 14 05:57:49.844: %CAPWAP-5-DTLSREQSUCC: DTLS connection created sucessfully peer_ip: 10.98.10.20 peer_port: 5246
*Oct 14 05:57:49.845: %CAPWAP-5-SENDJOIN: sending Join Request to 10.98.10.20
*Oct 14 05:57:49.845: %CAPWAP-5-CHANGED: CAPWAP changed state to JOIN
*Oct 14 05:57:49.930: %CAPWAP-5-CHANGED: CAPWAP changed state to CFG
*Oct 14 05:57:50.078: %CAPWAP-5-CHANGED: CAPWAP changed state to UP
*Oct 14 05:57:50.078: %LWAPP-3-CLIENTEVENTLOG: Received AP Syslog IP Address(255.255.255.255) configuration.
*Oct 14 05:57:50.206: %CAPWAP-5-JOINEDCONTROLLER: AP has joined controller nep00WLC00e
*Oct 14 05:57:50.293: %LWAPP-3-CLIENTEVENTLOG: SSID Roswifi added to the slot[0]
*Oct 14 05:57:50.298: %LWAPP-3-CLIENTEVENTLOG: SSID RosGuest added to the slot[0]
*Oct 14 05:57:50.303: %LWAPP-3-CLIENTEVENTLOG: SSID Roswifi added to the slot[1]
*Oct 14 05:57:50.309: %LWAPP-3-CLIENTEVENTLOG: SSID RosGuest added to the slot[1]
*Oct 14 05:57:50.414: %WIDS-6-ENABLED: IDS Signature is loaded and enabled

Nicolas Darchis · ‎10-14-2010

Sounds like a network disconnection between the ap and wlc so the AP had to go on another WLC. If you had AP fallback enabled and only the primary WLC configured in the high availability tab of the AP, the AP would go back to WLC1 as soon as the connectivity was back up. Here it stays on WLC2 until WLC2 fails, then it would go back to WLC1.

Hope it helps.

Nicolas