刘文粟,某局点UNIS S6600XPG 接入现网后导致网络延时变大
组网及说明
问题描述
接入6600XP后，核心ping cisco设备有较大延时
接入前网络中延时：
<7800XP-IRF>dis ip int bri | in 50
Vlan50 up up 192.x.x.5 /25 --
<S7800XP-IRF>ping 192.x.x.6
Ping 192.x.x.6 ( 192.x.x.6 ): 56 data bytes, press CTRL+C to break
56 bytes from 192.x.x.6 : icmp_seq=0 ttl=254 time=5.087 ms
56 bytes from 192.x.x.6 : icmp_seq=1 ttl=254 time=1.617 ms
56 bytes from 192.x.x.6 : icmp_seq=2 ttl=254 time=3.368 ms
56 bytes from 192.x.x.6 : icmp_seq=3 ttl=254 time=1.658 ms
56 bytes from 192.x.x.6 : icmp_seq=4 ttl=254 time=4.395 ms
接入后网络中延时：
[YGNL18F-OA-HXSW-S7800XP-IRF]ping 192.x.x.6
Ping 192.x.x.6 (192.x.x.6): 56 data bytes, press CTRL+C to break
56 bytes from 192.x.x.6: icmp_seq=0 ttl=254 time=181.926 ms
56 bytes from 192.x.x.6: icmp_seq=1 ttl=254 time=42.038 ms
56 bytes from 192.x.x.6: icmp_seq=2 ttl=254 time=47.680 ms
56 bytes from 192.x.x. 6: icmp_seq=3 ttl=254 time=20.431 ms
56 bytes from 192.x.x.6: icmp_seq=4 ttl=254 time=61.232 m
过程分析
1、7800XP 和思科互联的接口配置
#
interface Bridge-Aggregation1
port link-type trunk
port trunk permit vlan all
stp port bpdu-filter enable
#
2、最早就是聚合情况下中断的，为了规避光电转换器的问题，选择聚合单线路测试，而且现在如此高的流量感觉明显不太正常
7800XP 和思科互联接口流量较大，配的聚合，暂时只有一根线使用。
Ten-GigabitEthernet1/0/31
Current state: UP
Line protocol state: UP
Description: Ten-GigabitEthernet1/0/31 Interface
Bandwidth: 1000000 kbps
Peak input rate: 27293925 bytes/sec, at 2023-02-21 15:12:29
Peak output rate: 108555826 bytes/sec, at 2023-02-25 15:37:10
Last 300 seconds input: 247 packets/sec 47129 bytes/sec 0%
Last 300 seconds output: 524942 packets/sec 77191232 bytes/sec 70%
Vlan-interface50
Current state: UP
Line protocol state: UP
Bandwidth: 1000000 kbps
Maximum transmission unit: 1500
Last 300 seconds input rate: 151 bytes/sec, 1208 bits/sec, 2 packets/sec
Last 300 seconds output rate: 28 bytes/sec, 224 bits/sec, 0 packets/sec
Input: 18822 packets, 1510940 bytes, 3482 drops
Output: 1854 packets, 90884 bytes, 0 drops
看设备CPU也正常
===============display cpu===============
Slot 1 CPU 0 CPU usage:
2% in last 5 seconds
3% in last 1 minute
3% in last 5 minutes
Slot 2 CPU 0 CPU usage:
2% in last 5 seconds
2% in last 1 minute
2% in last 5 minutes
3、
接上去的66 XP只有一个接口UP的，其他的都是堆叠和MAD口，不像有环路
===============display interface brief===============
XGE1/0/45 UP 10G(a) F(a) T 1
XGE1/0/46 UP 10G(a) F(a) A 1 IRF-MAD
XGE1/0/47 UP 10G(a) F(a) -- --
XGE1/0/48 UP 10G(a) F(a) -- --
XGE2/0/46 UP 10G(a) F(a) A 1 IRF-MAD
XGE2/0/47 UP 10G(a) F(a) -- --
XGE2/0/48 UP 10G(a) F(a) -- --
互联接口发出的流量很大，按理说这个设备没有流量才对
Ten-GigabitEthernet1/0/45
Current state: UP
Line protocol state: UP
Bandwidth: 10000000 kbps
Loopback is not se
PVID: 1
MDI type: Automdix
Port link-type: Trunk
VLAN Passing: 1(default vlan), 10-11, 14-17, 19-22, 30-31, 34, 38, 40-43
45-48, 50
VLAN permitted: 1(default vlan), 10-11, 14-17, 19-22, 30-31, 34, 38, 40-43
45-48, 50
Trunk port encapsulation: IEEE 802.1q
Last 300 seconds input: 53 packets/sec 4643 bytes/sec 0%
Last 300 seconds output: 6544012 packets/sec 500480791 bytes/sec 51%
5、查看接口出入确实流量异常，被打满，堆叠口和BFD口流量也很大，开始怀疑堆叠口和检测口环了
6、再次检查配置
Mad检测的vlan是4000，接口放的是vlan1，IRF接口pvid也是1，接入网络的接口PVID也是1，导致VLAN1流量环路了
#
interface Vlan-interface4000
mad ip address 1.1.1.1 255.255.255.252 member 1
mad ip address 1.1.1.2 255.255.255.252 member 2
#
#
interface Ten-GigabitEthernet1/0/46
port link-mode bridge
description IRF-MAD
undo stp enable
#
解决方法
规范配置，修改mad检测口放通vlan
内容来源：知了社区，基于知识共享署名-相同方式共享3.0中国大陆许可协议CRM论坛（crmbbs.com）——一个让用户更懂CRM的垂直性行业内容平台,CRM论坛致力于互联网、客户管理、销售管理、SCRM私域流量内容输出5年。 如果您有好的内容，欢迎向我们投稿，共建CRM多元化生态体系，创建CRM客户管理一体化生态解决方案。,某局点UNIS S6600XPG 接入现网后导致网络延时变大
