交换机端口错包排查方法.doc

一、重要性： 从网维排查的用户反应网速慢的故障中，相当一部分是由于交换机和交换机端口、交换机端口和用户路由器端口、交换机端口和光电转换器端口、交换机端口和用户网卡匹配不当，产生错包引起的。该问题虽然比较常见，但只要按交换机开局来配置，通常能降低故障发生率，即使还存在端口错包问题，也能通过改端口速度和双工状态来彻底解决。 二、交换机端口错包说明： 以华为交换机端口为例进行说明： NJ-A-GJXC-S3026C-1dis int eth 0/1 Ethernet0/1 current state : UP IP Sending Frames Format is PKTFMT_ETHNT_2, Hardware address is 000f-e21d-bedc Description: to gujiao_33-192.168.86.161 The Maximum Transmit Unit is 1500 Media type is twisted pair, loopback not set Port hardware type is 100_BASE_TX 100Mbps-speed mode, full-duplex mode Link speed type is force link, link duplex type is force link Flow-control is not enabled Port-flow-constrain has not been configured completely The Maximum Frame Length is 1536 Broadcast MAX-ratio: 100% PVID: 10 Mdi type: auto Port link-type: hybrid Tagged VLAN ID : , 99 Untagged VLAN ID : , 10, 458 Last 300 seconds input:0 packets/sec 11 bytes/sec Last 300 seconds output:7 packets/sec 520 bytes/sec Input(total):11864483 packets, 6684922188 bytes 66707 broadcasts, 9522 multicasts Input(normal):- packets, - bytes - broadcasts, - multicasts Input:0 input errors, 0 runts, 0 giants,- throttles, 0 CRC 0 frame,- overruns, 0 aborts, 0 ignored, - parity errors Output(total): 12008069 packets, 5178275457 bytes 3638352 broadcasts, 288120 multicasts, 0 pauses Output(normal): - packets, - bytes - broadcasts, - multicasts, - pauses Output: 0 output errors,- underruns, - buffer failures 0 aborts, 0 deferred, 0 collisions, 0 late collisions 0 lost carrier, - no carrier 如上红字所示： 输入错误数，input errors等于各种重要错误信息的总和。不同的产品可以根据具体情况增加其它参数，或减少不能实现的参数。 1.Runts: discarded packets that are smaller than the mediums minimum packet size. ( Receive Byte Count < 64 and NOT CRC Error orReceive Byte Count < 64 and Tag Packet and NOT CRC Error， 无论是否有vlan tag，数据段小于64字节，而且没有CRC校验错误的帧。) 2.Giants: discarded packets that are larger than the mediums maximum packet size. ( Receive Byte Count > 1518 and NOT CRC Error orReceive Byte Count > 1522 and Tag Packet and NOT CRC Error， 没有vlan tag，数据段大于1518字节，小于最大帧长度，而且没有CRC校验错误的帧，和有vlan tag，数据段大于1522字节，小于最大帧长度，而且没有CRC校验错误的帧。) 3.Throttles: discarded packets that are incomplete frames. ( 交换机察觉缓存或CPU过载，关闭接口接收器的情形称为 throttle，是cisco路由器上的一个概念，我们的交换机目前不具备这个功能，一般应该显示为不支持。) 4.CRC: discarded packets with checksum error. ( Receive Packet CRC error，帧长度在正常范围（不带tag，长度在64到1518之间，或带tag，长度在64到1522之间），而且CRC校验错，如果支持此项，则不支持奇偶校验错误项。) 5.Frame: discarded packets with frame pad/sequence/alignment error, out of frame etc. （不是整数字节，而是多17bit，因此不对齐，或乱序或空帧,而且CRC校验错误，但是不计入CRC错误。） 6.Overrun: the receiver hardware is unable to hand receiveddata to a hardware buffer because the input rate exceeds the receiversability to handle the data. （由于接口输入速率超过接受方处理能力，导致丢包，由于我们的交换机一般是线速转发，这项一般应该为0，只有部分交换机对上传CPU或三层线速转发的帧有接口带宽限制，或是通过ACL实现的带宽限制，因此被丢弃的帧，计入此项。） 7.Aborts: Input abnormal frames that are discarded. ( 除其他错误之外，产品认为有必要统计的错误，例如前导码异常的帧，计入此项。) 8.Ignored: packets that are discarded because the interface hardware does not have enough internal buffers. ( Packetignored，由于接口内部buffer满，丢弃的帧，与由于主系统缓存空间缺乏，导致的丢弃帧不同。线速转发的帧，在多接口满带宽输入，单接口输出等情况下，由于输出接口的带宽不足，数据帧将内部缓存占满，导致从接口输入的帧在进入内部缓存之前被丢弃，以及进入内部缓存的帧超时无法输出，计入此项，上传到CPU的帧，由于CPU处理能力限制，toCPU的缓存满，导致被丢弃，也计入此项。) 9.Parity: Frames with parity error. ( Receive Packet parity error,如果支持此项，则不支持 CRC 错误项。) 输入丢弃数/输入错误数，output errors等于各种重要错误信息的总和。不同的产品可以根据具体情况增加其它参数，或减少不能实现的参数。 1.Underruns: There are no data in the output queue. ( Transmit under run，与 Overrun相反，输出接口的缓存从输出队列中取以太网帧时，没有帧，是一种非常少见的硬件异常。有的交换机就没有单独的接口输出缓存，与接口输出队列是同一块缓存。) 2.Buffer failures: Hardware does not have enough internal buffers. (内部缓存满，如果输出队列满，输出的以太网帧将在内部缓存中暂时存储，由于内部缓存满，导致帧丢弃。由于交换机对线速转发的数据帧发生的这种异常，认为只是到达内部缓存而没有到达出接口，是个输入帧，因此计入 Input IgnoredError，只有从CPU发出的帧，由于内部缓存满，导致帧丢弃，计入此项。) 3.Aborts: Output abnormal frames that are discarded. （ 在半双工模式下，由于冲突检测，延迟发送超过15次的帧，被丢弃，计入此项。除其他错误之外，产品认为有必要统计的错误，例如添加前导码异常的帧，也计入此项） 4.Deferred: The packets could not be sent out for collision under half-duplex mode. ( 半双工模式下，由于检测到载波正在被声明，当时没有发出的包，延时一次，计数加一。) 5.Collisions: Statistic of collision detection under half-duplex mode. ( 半双工模式下，在以太网帧数据部分的前64字节进入线路前，由于检测到冲突，当时没有发出的包。) 6.Late collisions: Statistic of collision with GE port sendingpackets less than 512 bytes and 10/100M port less than 64 bytes. ( 半双工模式下，在以太网帧数据部分的前64字节进入线路后，由于检测到冲突，当时没有发出的包。) 7.Lost carrier: Statistic of losing the carrier. （载波丢失，一般适用于串行WAN接口，发送过程中，每丢失一个载波，此计数加一，对于交换机，通常是由于线路中断造成。） 8.No carrier: Statistic of no carrier. （无载波，一般适用于串行WAN接口，当试图发送帧时，如果没有载波出现，此计数加一，对于交换机，通常是由于线路中断造成。） 如果“input errors”和“output errors”非“0”并且数字在增长，通常情况下可能存在端口匹配问题，必须进行端口状态调整。 其它厂商，cisco和中兴的二层交换机错包类型相似，可以类推。 三、错包问题排查方法 下面分别说下中兴、华为和CISCO二层交换机排查常用命令。 1中兴交换机端口排查方法，以Z2826S为例进行说明； 交换机端口原来配置,show run set port 1 description To_702 set port 1 auto disable set port 1 speed 10 set port 1 pvid 541 set port 1 macaddress on 1 set port 1 unit-statistics enable 交换机端口原来数据包统计 NJ-A-CFDS-Z2826S_2(1#7F)(cfg)#sh port 1 statistics PortId: 1 PortName: ReceivedFrames : 128998658 ReceivedBroadcastFrames: 1815 ReceivedBytes: 741746864 ReceivedMulticastFrames: 0 CrcError : 4934296 InPause: 0 InMACRcvErr: 0 Jabber : 2 Fragments: 6447228 UndersizeFrames: 0 Frames64Bytes: 1232457 Frames65_127Bytes: 84450478 Frames128_255Bytes : 7180773 Frames256_511Bytes : 2203985 Frames512_1023Bytes: 4738336 Frames1024_UpBytes : 29192629 OversizeFrames : 0 SendUnicastFrames: 127249246 SendBytes: 4270894316 SendNoneUnicastFrames: 1892514 SendFrames : 129141760 SendBroadcastFrames: 1035165 SendMulticastFrames: 857349 SendSingleCollision: 0 SendMultiCollision : 0 SendLateCollision: 0 SendExcessCollision: 0 SendCollision: 0 SendDefferTrans: 0 OutPause : 0 可以看到端口有大量“CrcError”错包，并且多次查看端口信息，可以看到错包在增长，该用户当前的上网速度肯定会慢，需要整改 更改端口匹配模式，改成端口自适应，主动宣告10M set port 1 auto enable set port 1 speedadvertise speed10 fullDuplex 更改后端口配置 set port 1 description To_702 set port 1 pvid 541 set port 1 speedadvertise speed10 FullDuplex set port 1 macaddress on 1 set port 1 unit-statistics enable 清除原端口1数据包统计，NJ-A-CFDS-Z2826S_2(1#7F)(cfg)#clear port 1 statistics ，再多次查看端口状态 NJ-A-CFDS-Z2826S_2(1#7F)(cfg)#sh port 1 statistics PortId: 1 PortName: ReceivedFrames : 2377ReceivedBroadcastFrames: 0 ReceivedBytes: 174972ReceivedMulticastFrames: 0 CrcError : 0 InPause: 0 InMACRcvErr: 0 Jabber : 0 Fragments: 0 UndersizeFrames: 0 Frames64Bytes: 4 Frames65_127Bytes: 2371 Frames128_255Bytes : 2 Frames256_511Bytes : 0 Frames512_1023Bytes: 0 Frames1024_UpBytes : 0 OversizeFrames : 0 SendUnicastFrames: 4293SendBytes: 6461573 SendNoneUnicastFrames: 6 SendFrames : 4299 SendBroadcastFrames: 4 SendMulticastFrames: 2 SendSingleCollision: 0 SendMultiCollision : 0 SendLateCollision: 0 SendExcessCollision: 0 SendCollision: 0 SendDefferTrans: 0 OutPause : 0 可以发现用户端口“CrcError”一直为“0”，此时用户上网才是正常的。 总结，只要2826端口有“CrcError”错误，一定要进行排查，直到“CrcError”错误为“0” 。 2华为二层交换机端口错包排查方法，以S2403H-EI为例进行说明 交换机端口原来配置 S2403EIdis cu int eth 0/1 interface Ethernet0/1 description to_101 speed 10 duplex full broadcast-suppression 5 mac-address max-mac-count 1 port access vlan 456 交换机端口原来数据包统计 S2403EI-2dis int eth 0/1 Ethernet0/1 current state : UP IP Sending Frames Format is PKTFMT_ETHNT_2, Hardware address is 000f-e22a-45fc Description: to_101 The Maximum Transmit Unit is 1500 Media type is twisted pair, loopback not set Port hardware type is 100_BASE_TX 10Mbps-speed mode, full-duplex mode Link speed type is force link, link duplex type is autonegotiation Flow-control is not enabled Port-flow-constrain has not been configured completely The Maximum Frame Length is 1518 Broadcast MAX-ratio: 5% PVID: 456 Mdi type: normal MAX mac-address learning count: 1 Port link-type: access Tagged VLAN ID : none Untagged VLAN ID : 456 Last 300 seconds input:29 packets/sec 5729 bytes/sec Last 300 seconds output:22 packets/sec 3705 bytes/sec Input(total):33659824 packets, 8999990068 bytes 16038 broadcasts, 175 multicasts Input(normal):33659823 packets, - bytes 16038 broadcasts, 175 multicasts Input:1102 input errors, 0 runts, 0 giants,- throttles, 112 CRC 0 frame,- overruns, 0 aborts, 2637 ignored, - parity errors Output(total): 36569492 packets, 12321086583 bytes 1404710 broadcasts, 5025215 multicasts, 0 pauses Output(normal): 36569492 packets, - bytes 1404710 broadcasts, 5025215 multicasts, - pauses Output: 0 output errors,- underruns, - buffer failures 19947 aborts, 0 deferred, 0 collisions, 0 late collisions - lost carrier, - no carrier 端口有1102个“input errors”，说明端口匹配有问题，需要进行排查。 更改端口匹配模式，改成端口指定10M，duplex自适应 duplex auto 更改后端口配置 interface Ethernet0/1 description to_101 speed 10 broadcast-suppression 5 mac-address max-mac-count 1 port access vlan 456 清除原端口1数据包统计，再多次查看端口状态，需要从“#”模式退到“>”模式 S2403EIquit <S2403EI> reset counters int eth 0/1 S2403EIdis int eth 0/1 Ethernet0/1 current state : UP IP Sending Frames Format is PKTFMT_ETHNT_2, Hardware address is 000f-e22a-45fc Description: to_101 The Maximum Transmit Unit is 1500 Media type is twisted pair, loopback not set Port hardware type is 100_BASE_TX 10Mbps-speed mode, full-duplex mode Link speed type is force link, link duplex type is autonegotiation Flow-control is not enabled Port-flow-constrain has not been configured completely The Maximum Frame Length is 1518 Broadcast MAX-ratio: 5% PVID: 456 Mdi type: normal MAX mac-address learning count: 1 Port link-type: access Tagged VLAN ID : none Untagged VLAN ID : 456 Last 300 seconds input:28 packets/sec 5804 bytes/sec Last 300 seconds output:27 packets/sec 4880 bytes/sec Input(total):33672483 packets, 9002467028 bytes 16038 broadcasts, 175 multicasts Input(normal):33672482 packets, - bytes 16038 broadcasts, 175 multicasts Input:0 input errors, 0 runts, 0 giants,- throttles, 0 CRC 0 frame,- overruns, 0 aborts, 0 ignored, - parity errors Output(total): 36580860 packets, 12322962080 bytes 1404792 broadcasts, 5025466 multicasts, 0 pauses Output(normal): 36580860 packets, - bytes 1404792 broadcasts, 5025466 multicasts, - pauses Output: 0 output errors,- underruns, - buffer failures 0 aborts, 0 deferred, 0 collisions, 0 late collisions - lost carrier, - no carrier 如果再多次查看端口状态，看到“input errors”为“0”，此时用户上网速度才正常。 3CISCO二层交换机端口错包排查方法，以C2950为例进行说明 交换机端口原来配置 2950#sh ru int fa0/3 interface FastEthernet0/3 description test switchport access vlan 100 no ip address duplex full speed 100 end 交换机端口原来数据包统计, 2950#sh int fa0/3 FastEthernet0/3 is up, line protocol is up (connected) Hardware is Fast Ethernet, address is 000d.bd98.1c43 (bia 000d.bd98.1c43) Description: test MTU 1500 bytes, BW 100000 Kbit, DLY 1000 usec, reliability 255/255, txload 1/255, rxload 1/255 Encapsulation ARPA, loopback not set Keepalive set (10 sec) Full-duplex, 100Mb/s input flow-control is off, output flow-control is off ARP type: ARPA, ARP Timeout 04:00:00 Last input never, output 00:00:01, output hang never Last clearing of "show interface" counters 21w6d Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0 Queueing strategy: fifo Output queue :0/40 (size/max) 5 minute input rate 0 bits/sec, 0 packets/sec 5 minute ouxtput rate 0 bits/sec, 0 packets/sec 330574221 packets input, 78353246 bytes, 0 no buffer Received 372 broadcasts, 0 runts, 0 giants, 0 throttles 3540 input errors, 3540 CRC, 0 frame, 0 overrun, 0 ignored 0 watchdog, 0 multicast, 0 pause input 0 input packets with dribble condition detected 307917156 packets output, 176359999 bytes, 0 underruns 0 output errors, 0 collisions, 0 interface resets 0 babbles, 0 late collision, 0 deferred 0 lost carrier, 0 no carrier, 0 PAUSE output 0 output buffer failures, 0 output buffers swapped out 端口有3540个“input errors”，说明端口匹配有问题，需要进行排查。 更改端口匹配模式，改成端口双工状态duplex自适应 duplex auto 更改后端口配置 2950#sh ru int fa0/3 interface FastEthernet0/3 description test switchport access vlan 100 no ip address speed 100 end 清除原端口1数据包统计，再多次查看端口状态 2950#clear counters fa0/3 Clear "show interface" counters on this interface confirm 回车确认 sw4-idc#sh int fa0/3 FastEthernet0/3 is up, line protocol is up (connected) Hardware is Fast Ethernet, address is 000d.bd98.1c43 (bia 000d.bd98.1c43) Description: HeHai-SyslogServer-218.104.84.114_5#B03 MTU 1500 bytes, BW 100000 Kbit, DLY 1000 usec, reliability 255/255, txload 1/255, rxload 1/255 Encapsulation ARPA, loopback not set Keepalive set (10 sec) Full-duplex, 100Mb/s input flow-control is off, output flow-control is off ARP type: ARPA, ARP Timeout 04:00:00 Last input never, output 00:00:01, output hang never Last clearing of "show interface" counters 00:01:40 Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0 Queueing strategy: fifo Output queue :0/40 (size/max) 5 minute input rate 0 bits/sec, 0 packets/sec 5 minute output rate 105000 bits/sec, 1 packets/sec 4278 packets input, 3722881 bytes, 0 no buffer Received 0 broadcasts, 0 runts, 0 giants, 0 throttles 0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored 0 watchdog, 0 multicast, 0 pause input 0 input packets with dribble condition detected 5454 packets output, 4793504 bytes, 0 underruns 0 output errors, 0 collisions, 0 interface resets 0 babbles, 0 late collision, 0 deferred 0 lost carrier, 0 no carrier, 0 PAUSE output 0 output buffer failures, 0 output buffers swapped out 多次查看端口状态，看到“input errors”为“0”，此时用户上网速度才正常。 四、端口错包排查常用命令 1中兴交换机常用命令 2826(cfg)#show run#查看设备配置文件# 2826(cfg)#sh port 1 statistics#查看端口数据统计信息# 2826(cfg)#set port 1 auto enable#改成端口自适应# 2826(cfg)#set port 1 speedadvertise speed10 fullDuplex#改端口主动宣告10M全双工# 2826(cfg)#clear port 1 statistics #清除原端口1数据包统计# 2华为二层交换机常用命令 S2403EIdis cu int eth 0/1 #查看端口配置# S2403EIdis int eth 0/1#查看端口数据统计信息# S2403EIint eth 0/1#进入0/1口# Speed 10 #端口速度指定为强制10M# duplex auto#双工自适应# S2403EIquit <S2403EI> reset counters int eth 0/1#清除原端口1数据包统计# S2403EIdis int eth 0/1 3.CISCO二层交换机端口错包排查方法，以C2950为例进行说明 2950#sh ru int fa0/3 #查看端口配置# 2950#sh int fa0/3#查看端口数据统计信息# 2950（config）#int fa0/3 #进入fa0/3口# Speed 10#端口速度指定为强制10M# duplex auto #双工自适应# 2950#clear counters fa0/3#清除原端口数据包统计# Clear "show interface" counters on this interface confirm 回车确认 请代维在网络开通和维护时，尽量减少因为端口匹配的故障。定期主动远程登入到交换机，查下端口统计信息，如果发现端口有错包，尽快解决，直到端口没有错包产生。