客户反馈MESH建立异常,MAP的以太口插上一根网线使得以太口up后,MAP就能在AC上注册成功MESH-LINK建立问题,否则MESH-LINK建立时断时连。
第一步:MESH作为一个基础协议,一般而言不该出现问题,所以先检查配置,发现MAP(mesh access)配置一处错误,“portal-service enable”含义是当前MP使能入口功能,或者可以理解输入这段指令,AP角色就变成了MPP(mesh portal point)。
wlan ap map model WA2620i-AGN id 779
priority level 7
serial-id 219801A0CNC127000117
portal-service enable
radio 1
channel 149
mp-policy 1
mesh-profile 1
mesh peer-mac-address 70f9-6d11-4512
radio enable
radio 2
service-template 1 vlan-id 751
service-template 2 vlan-id 77
service-template 3
radio enable
第二步:按照客户的配置在实验室中尝试复现,正常配置MESH-LINK建立正常;MAP上配置了portal-service enable,MESH-LINK无法连接。按照客户的办法,在MAP上接一根网线,使接口up,这样MESH就能建立,而实验室的MAP接上有线后,meshlink起来一段时间又会断开,和客户现场现象不同。指导客户删除该命令后现象依旧是:MAP的以太口插上一根网线使得以太口up后,MAP就能在AC上注册成功MESH-LINK建立问题,否则MESH-LINK建立时断时连。
第三步:在客户现场的MAP上debug wlan mesh all,显示异常
*May 8 15:31:13:981 2015 map WMSH/7/Event: Channel 5 is detected in zero cfg state
*May 8 15:31:14:175 2015 map WMSH/7/Event: Channel 13 is detected in zero cfg state
*May 8 15:31:14:391 2015 map WMSH/7/Event: Channel 4 is detected in zero cfg state
*May 8 15:31:14:593 2015 map WMSH/7/Event: Channel 12 is detected in zero cfg state
*May 8 15:31:14:744 2015 map WMSH/7/Event: Channel 3 is detected in zero cfg state
*May 8 15:31:14:910 2015 map WMSH/7/Event: Channel 11 is detected in zero cfg state
*May 8 15:31:15:139 2015 map WMSH/7/Event: Channel 2 is detected in zero cfg state
*May 8 15:31:15:305 2015 map WMSH/7/Event: Channel 10 is detected in zero cfg state
*May 8 15:31:15:487 2015 map WMSH/7/Event: Channel 1 is detected in zero cfg state
*May 8 15:31:15:927 2015 map WMSH/7/Event: Channel 8 is detected in zero cfg state
*May 8 15:31:17:559 2015 map WMSH/7/Event: Channel scan ended for the radio: 1 in Mode: 2
*May 8 15:31:18:377 2015 map WMSH/7/Event: Channel 9 is detected in zero cfg state
*May 8 15:31:18:950 2015 map WMSH/7/Event: Channel scan ended for the radio: 2 in Mode: 1
*May 8 15:31:27:788 2015 map WMSH/7/Event: Reset sequence number generate method to diver
*May 8 15:31:27:788 2015 map WMSH/7/Event: Active scanning is started for radio 1 in mode 2
*May 8 15:31:27:789 2015 map WMSH/7/Event: Reset sequence number generate method to diver
*May 8 15:31:27:789 2015 map WMSH/7/Event: Active scanning is started for radio 2 in mode 4
*May 8 15:31:27:789 2015 map WMSH/7/Event: The neighbor 80f6-2e12-5300 is detected during zero configuration scan
*May 8 15:31:28:709 2015 map WMSH/7/Event: Channel scan ended for the radio: 1 in Mode: 2
*May 8 15:31:33:126 2015 map WMSH/7/Event: Channel scan ended for the radio: 2 in Mode: 4
*May 8 15:31:33:126 2015 map WMSH/7/Event: Reset sequence number generate method to diver
*May 8 15:31:33:126 2015 map WMSH/7/Event: Active scanning is started for radio 2 in mode 1
*May 8 15:31:38:490 2015 map WMSH/7/Event: Channel scan ended for the radio: 2 in Mode: 1
*May 8 15:31:47:851 2015 map WMSH/7/Event: Zero configuration state active scanning is finished
*May 8 15:31:47:851 2015 map WMSH/7/Event: MA 70f9-6d64-57e0 Kdk fsm move to idle.
*May 8 15:31:47:851 2015 map WMSH/7/Event: ZeroCfg neighbor selection is triggered for 80f6-2e12-5300
*May 8 15:31:47:851 2015 map WMSH/7/Timer: Create check-link timer
*May 8 15:31:47:851 2015 map WMSH/7/Event: The selector for the link is 80f6-2e12-5300
*May 8 15:31:47:851 2015 map WMSH/7/Timer: Neighbor age timer is stopped for neighbor 80f6-2e12-5300
*May 8 15:31:47:851 2015 map WMSH/7/Timer: Random link initialization timer is started for neighbor 80f6-2e12-5300
*May 8 15:31:47:851 2015 map WMSH/7/Event: Peer Link Establishment process started for neighbor 80f6-2e12-5300
*May 8 15:31:47:852 2015 map WMSH/7/Fsm: LINK FSM Change state [Idle -> Listen] for neighbor 80f6-2e12-5300
*May 8 15:31:48:230 2015 map WMSH/7/Timer: Random link initialization timer expired for neighbor 80f6-2e12-5300
*May 8 15:31:48:230 2015 map WMSH/7/Frame_Send: Peer link open frame is being sent to neighbor 80f6-2e12-5300 in state Listen
*May 8 15:31:48:230 2015 map WMSH/7/Timer: Retry timer is started for neighbor 80f6-2e12-5300
*May 8 15:31:48:230 2015 map WMSH/7/Fsm: LINK FSM Change state [Listen -> OpenSent] for neighbor 80f6-2e12-5300
*May 8 15:31:48:232 2015 map WMSH/7/Frame_Rcvd: Peer link open frame is received from neighbor 80f6-2e12-5300
*May 8 15:31:48:232 2015 map WMSH/7/Frame_Send: Peer link confirm frame is being sent to neighbor 80f6-2e12-5300 in state OpenSent
*May 8 15:31:48:232 2015 map WMSH/7/Fsm: LINK FSM Change state [OpenSent -> OpenRcvd] for neighbor 80f6-2e12-5300
*May 8 15:31:48:238 2015 map WMSH/7/Frame_Rcvd: Peer link confirm frame is received from neighbor 80f6-2e12-5300
*May 8 15:31:48:238 2015 map WMSH/7/Timer: Retry timer is stopped for neighbor 80f6-2e12-5300
*May 8 15:31:48:238 2015 map WMSH/7/Fsm: LINK FSM Change state [OpenRcvd -> Established] for neighbor 80f6-2e12-5300
*May 8 15:31:48:299 2015 map WMSH/7/Event: Link information added successfully to driver for neighbor 80f6-2e12-5300
*May 8 15:31:48:439 2015 map WMSH/7/Event: Mesh security authentication bypassed since the device [70f9-6d64-57e0] or peer [80f6-2e12-5300] is in zero configuration state
*May 8 15:31:48:619 2015 map WMSH/7/Event: Temporary link is established with [80f6-2e12-5300]
*May 8 15:31:57:830 2015 map WMSH/7/Timer: Check link time age out
第四步:根据第三步所查看信息可知,AP已经信道扫描建立临时邻居后断开临时链路再次建立安全MESH-LINK时检测链路超时了。打开log信息查看发现接口反复up、down引起的MESH链路反复建立断开:
%Apr 28 18:02:39:056 2015 map IFNET/3/LINK_UPDOWN: Vlan-interface1 link status is UP.
%Apr 28 18:02:39:056 2015 map IFNET/5/LINEPROTO_UPDOWN: Line protocol on the interface Vlan-interface1 is UP.
%Apr 28 18:02:43:728 2015 map LWPC/6/LWPC_AP_UP:
Connection with AC 192.168.16.2 goes operational.
%Apr 28 18:02:44:894 2015 map IFNET/3/LINK_UPDOWN: WLAN-MESHLINK5 link status is DOWN.
%Apr 28 18:02:44:894 2015 map WMSH/6/WMESH_PEER_LINK_CLOSED: Peer link is closed with neighbor 80f6-2e12-5300 for reason:LINK_CANCELLED.
%Apr 28 18:02:44:894 2015 map WMSH/6/WMESH_EXIT_ZERO_CFG_STATE: The device has exited zero configuration state.
%Apr 28 18:02:44:896 2015 map IFNET/3/LINK_UPDOWN: Vlan-interface1 link status is DOWN.
%Apr 28 18:02:45:101 2015 map IFNET/5/LINEPROTO_UPDOWN: Line protocol on the interface Vlan-interface1 is DOWN.
%Apr 28 18:02:48:542 2015 map WMSH/6/WMESH_AUTHENTICATOR: The device 70f9-6d64-57e0 is an authenticator.
%Apr 28 18:02:48:765 2015 map IFNET/3/LINK_UPDOWN: WLAN-MESHLINK6 link status is UP.
%Apr 28 18:02:48:765 2015 map WMSH/6/WMESH_PEER_LINK_ESTABLISH: Peer link is established with neighbor 80f6-2e12-5300 in MESH WLAN-LINK,on radio 1 with BSSID 70f9-6d64-57e0.
%Apr 28 18:02:48:771 2015 map IFNET/3/LINK_UPDOWN: Vlan-interface1 link status is UP.
%Apr 28 18:02:48:771 2015 map IFNET/5/LINEPROTO_UPDOWN: Line protocol on the interface Vlan-interface1 is UP.
%Apr 28 18:02:48:777 2015 map WMSH/6/WMESH_AUTH_SUCCESS: Authentication success with neighbor 80f6-2e12-5300 in MESH WLAN-LINK,on radio 1 with BSSID 70f9-6d64-57e0.
%Apr 28 18:03:15:143 2015 map IFNET/3/LINK_UPDOWN: WLAN-MESHLINK6 link status is DOWN.
%Apr 28 18:03:15:143 2015 map WMSH/6/WMESH_PEER_LINK_CLOSED: Peer link is closed with neighbor 80f6-2e12-5300 for reason:CLOSE_RCVD.
%Apr 28 18:03:15:146 2015 map IFNET/3/LINK_UPDOWN: Vlan-interface1 link status is DOWN.
%Apr 28 18:03:15:146 2015 map IFNET/5/LINEPROTO_UPDOWN: Line protocol on the interface Vlan-interface1 is DOWN.
%Apr 28 18:03:15:308 2015 map IFNET/3/LINK_UPDOWN: WLAN-MESHLINK7 link status is UP.
第五步: MESH-LINK链路和物理层以太口up、down都没有关系与VLAN接口有关,回忆MESH原理:设备上电启动后扫描信道建立临时MESH链路,触发DHCP获取IP地址,触发LWAPP发现过程,建立LWAPP隧道,通过LWAPP从AC获取配置,建立安全MESH链路,完成MESH节点配置。也就是说VLAN接口down了导致MESH链路断开,那么VLAN接口为什么会down呢?首先display int vlan 1查看发现vlan接口地址时有时无,查看接口地址状态发现状态不稳定:
Vlan-interface1 DHCP client information:
Current machine state: HALT
dis dhcp client interface vlan 1
Vlan-interface1 DHCP client information:
Current machine state: REQUESTING
Offered IP: 192.168.12.80 255.255.252.0, lease: 600 seconds
DHCP server: 192.168.12.1
Display AP Profile
-------------------------------------------------------------------------------
Model Number : WA2620-AGN-S
Serial-ID : 219801A0H69145Q01522
AP Address : Trying for IPv4 address
H/W Version : Ver.A
S/W Version : V100R005B09D027(279072128)
Boot Version : 3.02
Device State : Zero configuration state
Master AC:
Description : -NA-
AC Address : -NA-
State : BDisc
Transmitted control packets : 0
Received control packets : 0
Transmitted data packets : 0
Received data packets : 0
Latest AC IP address : 192.168.16.2
Tunnel Down Reason : Response Timer Expire
-------------------------------------------------------------------------------
Unicast static AC IPv4 address: Not Configured
Unicast static AC IPv6 address: Not Configured
-------------------------------------------------------------------------------
既然接口地址有问题引起MESH异常,那么我们就MAP配置vlan地址,AC地址以及一条指向网关的静态路由,很快mesh链路建立起来并且不会断开,初步判断是dhcp问题:
Peer Link Information
-------------------------------------------------------------------------------
Nbr-Mac(rssi) BSSID Interface Link-state Uptime(hh:mm:ss)
-------------------------------------------------------------------------------
80f6-2e12-5300(37) 70f9-6d64-57e0 WLAN-MESHLINK7 Active(T) 0: 5: 16
-------------------------------------------------------------------------------
第六步:在MAP上收集DHCP信息:
*May 21 11:19:17:006 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Resend DHCPREQUEST for enough times. Move to INIT state.
*May 21 11:19:17:006 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: FSM state transfers (REQUESTING-->INIT) successfully.
*May 21 11:19:17:006 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: DHCPDISCOVER will be sent in 10 ms.
*May 21 11:19:17:016 2015 map DHCPC/7/DHCPC_PACKET:
Vlan-interface1: Send a DHCP packet...
Head: op (BOOTPREQUEST); htype (ETHERNET); hlen (6); xid (0x462a3522);
ciaddr (0.0.0.0); yiaddr (0.0.0.0); chaddr (70f9-6d64-57e0);
Options:
63 82 53 63 35 01 01 0C 1E 57 41 32 36 32 30 2D
41 47 4E 2D 53 2D 37 30 2D 66 39 2D 36 64 2D 36
34 2D 35 37 2D 65 30 32 04 C0 A8 0C 50 37 08 01
79 03 06 0F 21 2B 8A 39 02 04 80 3C 15 48 33 43
2E 20 48 33 43 20 57 41 32 36 32 30 2D 41 47 4E
2D 53 3D 07 01 70 F9 6D 64 57 E0 FF
*May 21 11:19:17:016 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Sending DHCPDISCOVER packet succeeded.
*May 21 11:19:17:016 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: FSM state transfers (INIT-->SELECTING) successfully.
*May 21 11:19:17:530 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Receive a packet.
*May 21 11:19:17:530 2015 map DHCPC/7/DHCPC_PACKET:
Vlan-interface1: Decode option 43:
80 07 00 00 01 C0 A8 10 02
*May 21 11:19:17:530 2015 map DHCPC/7/DHCPC_PACKET:
Vlan-interface1: Receive a DHCP packet...
Head: op (BOOTPREPLY); htype (ETHERNET); hlen (6); xid (0x462a3522);
ciaddr (0.0.0.0); yiaddr (192.168.12.80); chaddr (70f9-6d64-57e0);
Option: type (DHCPOFFER); mask (255.255.252.0); lease (600);
T1 (300); T2 (525); server (192.168.12.1); default router (192.168.12.1); Boot server (192.168.16.2)
*May 21 11:19:17:530 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Select 192.168.12.1 as the server.
*May 21 11:19:17:530 2015 map DHCPC/7/DHCPC_PACKET:
Vlan-interface1: Send a DHCP packet...
Head: op (BOOTPREQUEST); htype (ETHERNET); hlen (6); xid (0x462a3522);
ciaddr (0.0.0.0); yiaddr (0.0.0.0); chaddr (70f9-6d64-57e0);
Options:
63 82 53 63 35 01 03 0C 1E 57 41 32 36 32 30 2D
41 47 4E 2D 53 2D 37 30 2D 66 39 2D 36 64 2D 36
34 2D 35 37 2D 65 30 32 04 C0 A8 0C 50 36 04 C0
A8 0C 01 37 08 01 79 03 06 0F 21 2B 8A 39 02 04
80 3C 15 48 33 43 2E 20 48 33 43 20 57 41 32 36
32 30 2D 41 47 4E 2D 53 3D 07 01 70 F9 6D 64 57
E0 FF
*May 21 11:19:17:531 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Sending DHCPREQUEST packet succeeded
*May 21 11:19:17:591 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: FSM state transfers (SELECTING-->REQUESTING) successfully.
*May 21 11:19:17:731 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Receive a packet
*May 21 11:19:17:831 2015 map DHCPC/7/DHCPC_PACKET:
Vlan-interface1: Decode option 43:
80 07 00 00 01 C0 A8 10 02
*May 21 11:19:17:971 2015 map DHCPC/7/DHCPC_PACKET:
Vlan-interface1: Receive a DHCP packet...
Head: op (BOOTPREPLY); htype (ETHERNET); hlen (6); xid (0x462a3522);
ciaddr (0.0.0.0); yiaddr (192.168.12.80); chaddr (70f9-6d64-57e0);
Option: type (DHCPACK); mask (255.255.252.0); lease (600);
T1 (300); T2 (525); server (192.168.12.1); default router (192.168.12.1); Boot server (192.168.16.2)
*May 21 11:19:18:432 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Begin to detect IP address conflict via ARP.
*May 21 11:19:18:552 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Sending ARP request for address (192.168.12.80) succeeded.
*May 21 11:19:18:692 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Move to BOUND state if no ARP reply is received in 1500 milliseconds.
*May 21 11:19:19:036 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Receive no ARP reply for 192.168.12.80, so begin to use the address.
*May 21 11:19:19:036 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: FSM state transfers (REQUESTING-->BOUND) successfully.
*May 21 11:19:19:117 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Notify route module to add the default gateway: destination (0.0.0.0), mask (0.0.0.0), nexthop (192.168.12.1).
*May 21 11:19:19:317 2015 map DHCPC/7/DHCPC_EVENT:
Vlan-interface1: Sending ARP request for address (192.168.12.1) succeeded.
%May 21 11:19:35:086 2015 map IFNET/3/LINK_UPDOWN: WLAN-MESHLINK19 link status is DOWN.
%May 21 11:19:35:086 2015 map WMSH/6/WMESH_PEER_LINK_CLOSED: Peer link is closed with neighbor 80f6-2e12-5300 for reason:LINK_CANCELLED.
%May 21 11:19:35:087 2015 map WMSH/6/WMESH_EXIT_ZERO_CFG_STATE: The device has exited zero configuration state.
%May 21 11:19:35:089 2015 map IFNET/3/LINK_UPDOWN: Vlan-interface1 link status is DOWN.
从DHCP的debug看,报文有请求有回复似乎也不是DHCP的问题!为了保险起见在DHCP中继和MPP的中间有线做了镜像抓包。
发现dhcp请求和回应均是成对出现,与此前某局点MESH相同现象dhcp server的安全机制不允许相同client在短时间内连续申请IP地址引起的问题不一样。继续观察发现一个很奇怪的现象,DHCP中继回复的报文有的是广播有的是单播,打开1270号报文dhcp discover继续观察,发现报文标志位要求广播回复:
打开1326号报文dhcp offer观察,中继设备回复的报文为广播:
打开1327号报文DHCP request观察,标志位要求广播:
打开1328号报文DHCP ACK,DHCP relay回复报文为广播:
而之后的DHCP ACK却变成了单播包:
这是因为MESH-LINK建立需要两次DHCP申请拿地址过程,第一次dhcp中继设备回复报文是MAP要求的广播ACK,而第二次回复的ACK确实单播的,导致VLAN接口异常,LWAPP无法建立,从而不能建立稳定的MESH连接。为了进一步核实想法,在实验室搭建了测试环境,使用我司交换机作为DHCP中继,抓包显示来往的DHCP报文均为广播:
第七步:查看RFC2131可知DHCP ACK报文“broadcast or unicast, based on bootp flags.”综上可知,客户MESH建立异常的现象是由于中继设备DHCP协议实现不标准引起。
使用H3C交换机做中继或者将MAP的DHCP服务器放到AC上通过二层拿地址。
该案例暂时没有网友评论
✖
案例意见反馈
亲~登录后才可以操作哦!
确定你的邮箱还未认证,请认证邮箱或绑定手机后进行当前操作