L5030上配置了服务器负载均衡,想要实现将外网用户访问虚服务IP的流量分发给不同的实服务器,减轻服务器的业务负担。L5030旁挂在核心交换机上,现网的多台服务器在拓扑中用一台服务器体现
现场配置完服务器负载均衡后发现实服务组状态不可用,从图中可以看出状态不可用的实服务组中可用实服务器数为0。
健康性检测中有很多探测失败的系统日志,如果实服务器NQA探测失败会导致实服务器不可用。
1、于是根据不可用实服务组test-01进行排查,检查L5030和实服务组test-01中的两个服务器的连通性是否可达。查看实服务组状态显示两个实服务器状态均为探测失败。
[L5030]display server-farm
Server farm: test-01
Description:
Predictor: Hash address source IP
Proximity: Disabled
NAT: Enabled
SNAT pool: test02
Failed action: Keep
Active threshold: Disabled
Slow-online: Disabled
Selected server: Disabled
Busy action: Drop
Probe information:
Probe success criteria: At least 1
Probe method:
tcp-temp
Total real server: 2
Active real server: 0
Real server list:
Name State VPN-instance Address Port Weight Priority
test1 Probe-failed HLW 10.197.44.200 80 100 4
test2 Probe-failed HLW 10.197.44.201 80 100 4
ping测试发现L5030无法ping通10.197.44.200,但是可以ping通10.197.44.201。
[GZ-HXY-G160201-V07-L5000S-L-01]ping -vpn-instance HLW 10.197.44.200
Ping 10.197.44.200 (10.197.44.200): 56 data bytes, press CTRL_C to break
Request time out
Request time out
Request time out
Request time out
Request time out
--- Ping statistics for 10.197.44.200 in VPN instance HLW ---
5 packet(s) transmitted, 0 packet(s) received, 100.0% packet loss
[GZ-HXY-G160201-V07-L5000S-L-01]ping -vpn-instance HLW 10.197.44.201
Ping 10.197.44.201 (10.197.44.201): 56 data bytes, press CTRL_C to break
56 bytes from 10.197.44.201: icmp_seq=0 ttl=63 time=0.993 ms
56 bytes from 10.197.44.201: icmp_seq=1 ttl=63 time=0.865 ms
56 bytes from 10.197.44.201: icmp_seq=2 ttl=63 time=0.814 ms
56 bytes from 10.197.44.201: icmp_seq=3 ttl=63 time=0.828 ms
56 bytes from 10.197.44.201: icmp_seq=4 ttl=63 time=0.864 ms
--- Ping statistics for 10.197.44.201 in VPN instance HLW ---
5 packet(s) transmitted, 5 packet(s) received, 0.0% packet loss
2、由于现场采用的探测方式是TCP探测,所以需要排查是否是服务器端口不可达导致10.197.44.201可以ping通,但是探测失败。首先查看会话,看报文是从哪个接口上出去的,如果看不到报文数量的话需要通过session statistics enable开启会话统计功能,可以看到报文是从冗余口2出去的。
[L5030]display session table ipv4 destination-ip 10.197.44.201 verbose
Slot 1:
Initiator:
Source IP/port: 10.197.0.122/51188
Destination IP/port: 10.197.44.201/80
DS-Lite tunnel peer: -
VPN instance/VLAN ID/Inline ID: HLW/-/-
Protocol: TCP(6)
Inbound interface: InLoopBack0
Responder:
Source IP/port: 10.197.44.201/80
Destination IP/port: 10.197.0.122/51188
DS-Lite tunnel peer: -
VPN instance/VLAN ID/Inline ID: HLW/-/-
Protocol: TCP(6)
Inbound interface: Reth2
State: TCP_SYN_SENT
Application: HTTP
Rule ID: -/-/-
Rule name:
Start time: 2023-06-25 13:53:05 TTL: 2s
Initiator->Responder: 0 packets 0 bytes
Responder->Initiator: 0 packets 0 bytes
查看冗余口2的配置,可以看到Route-Aggregation1.1973的优先级高,则报文会从Route-Aggregation1.1973转发。
interface Reth2
description output-int
ip binding vpn-instance HLW
ip address 10.197.0.122 255.255.255.252
member interface Route-Aggregation1.1973 priority 100
member interface Route-Aggregation2.1973 priority 50
于是对Route-Aggregation1.1973进行抓包,发现反而和10.197.44.200之间有TCP报文交互,但是没有收到10.197.44.201的回应报文。3、由于抓包的结果比较奇怪,所以进一步对聚合口下的成员接口进行抓包。首先display link-aggregation verbose查看聚合口1中的成员接口,可以看到XGE1/1/0和XGE1/1/1为其成员接口。
Aggregate Interface: Route-Aggregation1
Aggregation Mode: Dynamic
Loadsharing Type: Shar
System ID: 0x8000, 9023-b46c-4400
Local:
Port Status Priority Oper-Key Flag
--------------------------------------------------------------------------------
XGE1/1/0 S 32768 3 {ACDEFG}
XGE1/1/1 U 32768 3 {ACG}
于是对两个接口进行抓包,发现XGE1/1/1接口上也抓取到了和实服务器10.197.44.200交互的报文。
但是根据查看的聚合口状态信息,XGE1/1/1接口是未选中状态的。那么可能就是交换机往本端未选中接口转发流量导致了该问题,将XGE1/1/1接口shutdown后ping服务器测试正常。
1、排查交换机侧往本端未选中端口发包的原因。发现两端聚合模式不匹配,交换机上是静态聚合,L5030上的是动态聚合,将聚合模式改成一致后正常。
L5030配置
#
interface Route-Aggregation1
description input-int
link-aggregation mode dynamic
#
交换机配置
#
interface Bridge-Aggregation33
description GZ-HXY-G160201-V07-H5030
port link-type trunk
undo port trunk permit vlan 1
port trunk permit vlan 972 to 973 1972 to 1973
#
#
interface Ten-GigabitEthernet1/2/0/43
port link-mode bridge
description GZ-HXY-G160201-V07-H5030-01_PORT1
port link-type trunk
undo port trunk permit vlan 1
port trunk permit vlan 972 to 973 1972 to 1973
port link-aggregation group 33
#
interface Ten-GigabitEthernet1/2/0/44
port link-mode bridge
description GZ-HXY-G160201-V07-H5030-01_PORT2
port link-type trunk
undo port trunk permit vlan 1
port trunk permit vlan 972 to 973 1972 to 1973
port link-aggregation group 33
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