★ AP offline query method

Universal

1.   The first step need to check when the AP is offline

We first observe whether the AP is down because it has rebooted or not. So how do we determine this?

1)        Check the display wlan ap name xxx verbose on the AC

By checking the display wlan ap name xxx verbose on the AC and observing the system up time and online time, if they are consistent, it means that the AP has rebooted frequently recently and we need to further check the reboot reason; if they are not consistent, we need to focus on the tunnel down reason field and check the specific reason for the drop.

If the AP is offline due to restart, we need to pay attention to the specific cause, which can be viewed through the following fields:

2)      What should I do if I confirm that the AP has rebooted?

Analyze the causes of AP reboots further:

l  Power on: Check whether the AP upstream POE switch ports are powered off. If some aps are powered off, check the cable length, cable quality, and power of the POE switch. If possible, use the local power supply or POE module to power the switch and check whether the power on is still restarted.

l  Kernel exception soft reboot: If there is an abnormal reboot, call h3c local hotline for technical support.

1.   How to check the cause of AP offline

AC side:

[AC] display wlan ap statistics tunnel-down-record

AP name    AP ID      Tunnel down at    Tunnel Down Reason

ap1           2        08-30/14:56:00    Processed join request in Run state

AP side:

[AP] display logbuffer

%Mar30 15:58:17:774 2021 LONGI-A1-10FAP13 CWC/4/CWC_AP_DOWN: Master CAPWAP tunnel to AC 172.30.230.3 went down. Reason: Deleted AP IP address.

2.   Common AP offline causes

The following are the causes of AP offline, the more common causes are: Failed to retransmit message, Processed join request in Run state, Neighbor dead timer expired, these are usually caused by unstable link and packet loss.

3.   How to check whether the AC - AP link is faulty

1)      Pay attention to the number of echo packets lost by the AP

Collect display wlan ap all verbose to view AP details. The number of Lost echo responses displayed in the statistics shows the number of offline APs. Based on the Capwap tunnel mechanism, after the AP sends Echo heartbeat packets, the controller needs to respond with Ack packets. If the AP receives no Echo response packets more than three times, the AP considers that the link is faulty. If the number of Lost echo responses is large, attention should be paid to the AC-AP link problem.

2)      View AP link detection parameters

The wlan ap-link-test command can test the link between the AC and the AP for packet loss.

       [AC]probe

       [AC-probe]wlan ap-link-test 192.168.238.7

       Transmitted packets : 1

       Received packets : 1

       Packet loss : 0.00%

       Transmitted packet length : 128

       Received packet Length : 128

       Out-of-order packets : 0

       Min RTT(ms) : 1 Average

       RTT(ms) : 1

       Max RTT(ms) : 1

       AP last reboot reason : Power on

       AP up time : 1 weeks, 6 days, 8 hours, 42 minutes

3)      Pay attention to the link quality of the AP wired port

      [AP]display interface

      Input: 0 input errors, 0 runts, 0 giants, - throttles 0 CRC, - frame, 0 overruns, 0 aborts - ignored, - parity errors

CRC: Link quality check, generally the presence of CRC in the interface indicates that the link quality of the interface is not good and need to pay attention to whether the count is increasing.

Solution: If the number continues to increase, replace the link (network cable) and check whether the uplink switch port has error packets

Overruns: If the number of AP broadcast and multicast traffic on the wire-line interface is much larger than the number of unicasts and this parameter continues to grow, pay attention to it.

Solution: Streamline the AP wired port VLAN, configure Layer 2 isolation, and isolate the port.

4)      Whether the AC and AP cross the public network

To ensure the stability of public/private network links, can ping large packets to test the link stability. For example, ping 1000 packets of 1500 byte:

      ping -s 1500 -c 1000 192.168.238.7

      Ping 192.168.238.7 (192.168.238.7): 1500 data bytes, press CTRL_C to break

      1500 bytes from 192.168.238.7: icmp_seq=0 ttl=255 time=1.069 ms

      1500 bytes from 192.168.238.7: icmp_seq=1 ttl=255 time=0.975 ms

      1500 bytes from 192.168.238.7: icmp_seq=2 ttl=255 time=1.015 ms

4.   Check if the APs are offline in bulk

Run the following command to check whether the AP tunnel disconnects at the same time:

[AP] display wlan ap statistics tunnel-down-record

AP name AP ID Tunnel down at Tunnel Down Reason

ap1 1 09-01/15:26:28 Processed join request in Run state

ap2 2 09-01/15:26:28 Processed join request in Run state

ap3 3 09-01/15:26:28 Processed join request in Run state

ap4 4 09-01/15:26:28 Processed join request in Run state

ap4 5 09-01/15:26:28 Failed to retransmit message 

1)      If there is a situation where the APs are dropping out in bulk, focus on the whole network. Whether the wireless and multicast reports of the AP wired port are too large. (e.g., the number of wireless and multicast files is twice that of unicast).

Solution: Streamline the VLAN of the AP; the local forwarding uplink switch port is isolated.

2)      Interface flapping occurs on the AP upstream access switch. If STP TC packets are generated on the interface, all devices connected to the core and enabled with STP refresh ARP entries after receiving TC packets and send a large number of ARP request packets. If all VLANs are enabled on the interface connected to the core and AC, the AC receives a large number of ARP packets. The CPU increases and the AP goes offline.

Solution: Only necessary VLANs are enabled for the link between the core and AC. Add BPDU protection to edge ports on the AP upstream POE switch.