网线质量导致端口速率变低

发布时间:  2014-09-12 浏览次数:  366 下载次数:  8
问题描述
OCS Core Switch1和USAU Lan-Switch1之间以trunk连接,OCS Core Switch1的0/0/35口速率降为10M,导致业务过大的时候端口处理能力不够,同时业务并不切换到备份链路导致影响业务。

告警信息
处理过程
通过分析日志发现,环境中有多个端口还存在协商成百兆的情况, 其中包括OCS-SWITCH1的交换机如下端口,对端是ATAE设备,经过确认两端都是GE自协商,正常情况下应该协商成千兆。这和之前g0/0/35协商出10M的情况相类似。

GigabitEthernet0/0/32 current state : UP
Line protocol current state : UP
Description:port maybe some issue do not use it
Switch Port,PVID :    1,The Maximum Frame Length is 1600
IP Sending Frames' Format is PKTFMT_ETHNT_2, Hardware address is 781d-baa5-d7c1
Port Mode: COMMON COPPER
Speed :   100,  Loopback: NONE



在3月1日出问题之后,没有更换网线和端口,重新插拔了网线,协商出1000M后正常使用。 问题出现在3.12日,通过日志看出,3.11日出现过端口down/up,而端口速率重新协商之后会造成端口down/up,最后一次查询日志,端口是以10M的速率UP起来,说明发生过速率重新协商。



Mar 11 2013 23:58:03+05:30 PUN-OCS-CORE-LAN-SW-5352-1 %%01IFNET/4/IF_STATE(l)[252]:Interface GigabitEthernet0/0/35 has turned into UP state.
Mar 11 2013 23:57:28+05:30 PUN-OCS-CORE-LAN-SW-5352-1 %%01IFNET/4/IF_STATE(l)[253]:Interface GigabitEthernet0/0/35 has turned into DOWN state.
Mar 11 2013 23:54:38+05:30 PUN-OCS-CORE-LAN-SW-5352-1 %%01IFNET/4/IF_STATE(l)[256]:Interface GigabitEthernet0/0/35 has turned into UP state.
Mar 11 2013 23:54:32+05:30 PUN-OCS-CORE-LAN-SW-5352-1 %%01IFNET/4/IF_STATE(l)[257]:Interface GigabitEthernet0/0/35 has turned into DOWN state.
Mar 11 2013 23:53:35+05:30 PUN-OCS-CORE-LAN-SW-5352-1 %%01IFNET/4/IF_STATE(l)[258]:Interface GigabitEthernet0/0/35 has turned into UP state.
Mar 11 2013 23:53:25+05:30 PUN-OCS-CORE-LAN-SW-5352-1 %%01IFNET/4/IF_STATE(l)[259]:Interface GigabitEthernet0/0/35 has turned into DOWN state.
 





通过查看协商出100M端口的底层芯片寄存器信息,发现芯片寄存器发出了1000M能力码流,但是并未收到对端自协商能力。


3月14日凌晨对协商出百兆的端口和网线进行VCT测试,其测试结果如下:
[PUN-OCS-CORE-LAN-SW-5352-2-GigabitEthernet0/0/32]virtual-cable-test
Warning: The command will stop service for a while, continue [Y/N]?Y
Pair A length: 10meter(s)
Pair B length: 0meter(s)
Pair C length: 0meter(s)
Pair D length: 10meter(s)
Pair A state: Ok
Pair B state: Crosstalk
Pair C state: Crosstalk
Pair D state: Ok
 
 

[PUN-OCS-CORE-LAN-SW-5352-2-GigabitEthernet0/0/43]virtual-cable-test
Warning: The command will stop service for a while, continue [Y/N]?Y
Pair A length: 42meter(s)
Pair B length: 41meter(s)
Pair C length: 40meter(s)
Pair D length: 42meter(s)
Pair A state: Ok
Pair B state: Ok
Pair C state: Short
Pair D state: Short 
  

[PUN-OCS-CORE-LAN-SW-5352-2-GigabitEthernet0/0/31]virtual-cable-test
Warning: The command will stop service for a while, continue [Y/N]?Y
Pair A length: 11meter(s)
Pair B length: 8meter(s)
Pair C length: 8meter(s)
Pair D length: 11meter(s)
Pair A state: Ok
Pair B state: Crosstalk
Pair C state: Crosstalk
Pair D state: Ok
 


[PUN-OCS-CORE-LAN-SW-5352-2-GigabitEthernet0/0/7]virtual-cable-test
Warning: The command will stop service for a while, continue [Y/N]?Y
Pair A length: 10meter(s)
Pair B length: 10meter(s)
Pair C length: 10meter(s)
Pair D length: 10meter(s)
Pair A state: Ok
Pair B state: Crosstalk
Pair C state: Crosstalk
Pair D state: Ok
 

其中Crosstalk(交叉树)表明此双绞线在末端未正确连接。Short(短路)表明两个或更多导线一起短路。只有测试到的四个线序均为OK字样的时候才表明当前网线工作状态正常。
如上都说明由于这些网线目前连接状态不好,导致速率协商不上去。环境中网线有问题,会导致速率协商错误、错包、时通时断等现象。

从分析和测试结果看,因部分网线质量问题导致交换机的端口并未协商成1000M,有的协商成100M,有的协商成10M。

根因
 原因是该条链路上的网线质量有问题,影响两端设备的协商码流,导致速率协商不上去变为10M导致。

 
建议与总结
1 检查网线质量,更换质量好的网线。
2 将两台设备的组成trunk口的两个端口配置成强制千兆全双工模式,这样该端口就只能工作在1000M全双工模式下,如果发生链路质量较差的情况,端口无法协商出1000M,就down了,业务会切换到另外一条链路上,可以有效进行链路切换保护业务。


END