[j-nsp] Strange Packet Loss Problem

Christian Koch ckoch at globix.com
Mon Oct 23 08:38:21 EDT 2006


Hey Josef,

Real quick, here is  a regular ping which looks fine, and a snapshot of
system icmp stats

ckoch at core2.lhr3> ping 212.71.229.25                             
PING 212.71.229.25 (212.71.229.25): 56 data bytes
64 bytes from 212.71.229.25: icmp_seq=0 ttl=251 time=1.113 ms
64 bytes from 212.71.229.25: icmp_seq=1 ttl=251 time=1.063 ms
64 bytes from 212.71.229.25: icmp_seq=2 ttl=251 time=0.957 ms
64 bytes from 212.71.229.25: icmp_seq=3 ttl=251 time=0.995 ms
64 bytes from 212.71.229.25: icmp_seq=4 ttl=251 time=0.947 ms
64 bytes from 212.71.229.25: icmp_seq=5 ttl=251 time=1.429 ms
64 bytes from 212.71.229.25: icmp_seq=6 ttl=251 time=0.934 ms
64 bytes from 212.71.229.25: icmp_seq=7 ttl=251 time=1.017 ms
64 bytes from 212.71.229.25: icmp_seq=8 ttl=251 time=0.935 ms
64 bytes from 212.71.229.25: icmp_seq=9 ttl=251 time=0.973 ms
64 bytes from 212.71.229.25: icmp_seq=10 ttl=251 time=1.138 ms
64 bytes from 212.71.229.25: icmp_seq=11 ttl=251 time=1.003 ms
64 bytes from 212.71.229.25: icmp_seq=12 ttl=251 time=0.946 ms
64 bytes from 212.71.229.25: icmp_seq=13 ttl=251 time=1.172 ms
64 bytes from 212.71.229.25: icmp_seq=14 ttl=251 time=1.002 ms
64 bytes from 212.71.229.25: icmp_seq=15 ttl=251 time=1.009 ms
64 bytes from 212.71.229.25: icmp_seq=16 ttl=251 time=0.921 ms
64 bytes from 212.71.229.25: icmp_seq=17 ttl=251 time=1.009 ms
64 bytes from 212.71.229.25: icmp_seq=18 ttl=251 time=0.959 ms
64 bytes from 212.71.229.25: icmp_seq=19 ttl=251 time=0.973 ms
64 bytes from 212.71.229.25: icmp_seq=20 ttl=251 time=2.362 ms
64 bytes from 212.71.229.25: icmp_seq=21 ttl=251 time=1.034 ms
64 bytes from 212.71.229.25: icmp_seq=22 ttl=251 time=1.007 ms
64 bytes from 212.71.229.25: icmp_seq=23 ttl=251 time=0.913 ms
64 bytes from 212.71.229.25: icmp_seq=24 ttl=251 time=0.959 ms
64 bytes from 212.71.229.25: icmp_seq=25 ttl=251 time=1.021 ms
64 bytes from 212.71.229.25: icmp_seq=26 ttl=251 time=0.996 ms
64 bytes from 212.71.229.25: icmp_seq=27 ttl=251 time=0.966 ms
64 bytes from 212.71.229.25: icmp_seq=28 ttl=251 time=3.044 ms
64 bytes from 212.71.229.25: icmp_seq=29 ttl=251 time=0.940 ms
64 bytes from 212.71.229.25: icmp_seq=30 ttl=251 time=0.946 ms
64 bytes from 212.71.229.25: icmp_seq=31 ttl=251 time=1.054 ms
64 bytes from 212.71.229.25: icmp_seq=32 ttl=251 time=0.989 ms
64 bytes from 212.71.229.25: icmp_seq=33 ttl=251 time=0.973 ms
64 bytes from 212.71.229.25: icmp_seq=34 ttl=251 time=0.963 ms
64 bytes from 212.71.229.25: icmp_seq=35 ttl=251 time=1.028 ms
64 bytes from 212.71.229.25: icmp_seq=36 ttl=251 time=0.954 ms
64 bytes from 212.71.229.25: icmp_seq=37 ttl=251 time=0.965 ms
64 bytes from 212.71.229.25: icmp_seq=38 ttl=251 time=0.999 ms
64 bytes from 212.71.229.25: icmp_seq=39 ttl=251 time=1.003 ms
64 bytes from 212.71.229.25: icmp_seq=40 ttl=251 time=0.897 ms
64 bytes from 212.71.229.25: icmp_seq=41 ttl=251 time=1.126 ms
64 bytes from 212.71.229.25: icmp_seq=42 ttl=251 time=0.913 ms
64 bytes from 212.71.229.25: icmp_seq=43 ttl=251 time=1.058 ms
64 bytes from 212.71.229.25: icmp_seq=44 ttl=251 time=0.942 ms
64 bytes from 212.71.229.25: icmp_seq=45 ttl=251 time=1.030 ms
64 bytes from 212.71.229.25: icmp_seq=46 ttl=251 time=0.941 ms
64 bytes from 212.71.229.25: icmp_seq=47 ttl=251 time=1.047 ms
64 bytes from 212.71.229.25: icmp_seq=48 ttl=251 time=0.938 ms
64 bytes from 212.71.229.25: icmp_seq=49 ttl=251 time=0.954 ms
64 bytes from 212.71.229.25: icmp_seq=50 ttl=251 time=0.964 ms
64 bytes from 212.71.229.25: icmp_seq=51 ttl=251 time=0.976 ms
64 bytes from 212.71.229.25: icmp_seq=52 ttl=251 time=0.944 ms
64 bytes from 212.71.229.25: icmp_seq=53 ttl=251 time=1.006 ms
64 bytes from 212.71.229.25: icmp_seq=54 ttl=251 time=0.972 ms
64 bytes from 212.71.229.25: icmp_seq=55 ttl=251 time=1.585 ms
64 bytes from 212.71.229.25: icmp_seq=56 ttl=251 time=0.990 ms
64 bytes from 212.71.229.25: icmp_seq=57 ttl=251 time=11.160 ms
64 bytes from 212.71.229.25: icmp_seq=58 ttl=251 time=0.948 ms
64 bytes from 212.71.229.25: icmp_seq=59 ttl=251 time=21.018 ms
64 bytes from 212.71.229.25: icmp_seq=60 ttl=251 time=0.996 ms
64 bytes from 212.71.229.25: icmp_seq=61 ttl=251 time=30.768 ms
^C
--- 212.71.229.25 ping statistics ---
62 packets transmitted, 62 packets received, 0% packet loss
round-trip min/avg/max/stddev = 0.897/2.029/30.768/4.634 ms

ckoch at core2.lhr3> show system statistics icmp 
icmp:
        0 drops due to rate limit
        3747728 calls to icmp_error
        0 errors not generated because old message was icmp
        Output histogram:
                echo reply: 9220380
                destination unreachable: 3745300
                time exceeded: 1577
                time stamp reply: 472
        2 messages with bad code fields
        0 messages less than the minimum length
        338 messages with bad checksum
        2 messages with bad source address
        28 messages with bad length
        15 echo drops with broadcast or multicast destinaton address
        0 timestamp drops with broadcast or multicast destination
address
        Input histogram:
                echo reply: 50251
                destination unreachable: 5631
                source quench: 6
                routing redirect: 829
                #7: 2
                echo: 9220395
                time exceeded: 1883
                parameter problem: 1
                time stamp: 472
                time stamp reply: 1
                information request reply: 20
                address mask reply: 2
        9220852 message responses generated 

-----Original Message-----
From: Josef Buchsteiner [mailto:josefb at juniper.net] 
Sent: Saturday, October 21, 2006 9:27 AM
To: Christian Koch
Cc: juniper-nsp at puck.nether.net
Subject: Re: [j-nsp] Strange Packet Loss Problem

Christian,

          if it is a data corruption on certain bit patterns then
          you  should  see this problem also without rapid pings.
          In  order  to  find  out  where  this  happen  I  would
          certainly  perform  a  regular ping ( you may need more
          packets ) and watch the icmp statistics to confirm that
          it  is  data  corruption  in  case  you get icmp chksum
          errors. "show system statistics icmp"

          To find out if this is local or remote ( you could make
          some   statements  with  the  icmp  statistics  already
          however I prefer solid results)and you need to put this
          sonet link into local loop and encapsulation cisco-hdlc
          with no-keepalive. Then you perform the ping again with
          bypass  routing  and  interface  knob  to make sure the
          traffic  goes  out  and comes back in and you watch the
          packet  result  and  icmp  statistics. This way you will
          find  out  if this is local M40 or remote M40.

          Once  you  know  this  it  becomes kinda bit tricky. It
          could  be  the  FPC in question that is the trigger but
          also  other  FPCs  are  contributing in the shared data
          buffer  which  means  you would need to turn off one by
          one  and  see  if the error stops. If you are left with
          the  OC-48  then  I  would  suggest  to  replace  this
          particular FPC. To sanity check also other links of the
          router if possible.

          Check  also  the  message  log  if  you ever have seen
          anytime ECC error reported. Just one of them is already
          enough  to give you some hints as you would not need to
          do above isolation work

          hope this helps
          Josef

          PS:   yes  I  do  have  seen  such symptoms in the past
          and not only on one vendor.


          
Friday, October 20, 2006, 8:36:09 PM, you wrote:

CK>    
CK>    
CK> Hi All,
CK>  
CK>  I am experiencing a strange issue between 2 m40 core routers 
CK> connect  through an oc48 link
CK>  
CK>  I experience packet loss only when sending packets between 293 and 
CK> 599  bytes.
CK>  
CK>  I am also speaking to cold telecom about the issue to see if it's a

CK> problem with the sonet link, but as of now the link is not taking 
CK> any  errors or alarms.
CK>  
CK>  Anyone seen anything strange like this before?
CK>  
CK>  ckoch at core2.lhr3> ping rapid 212.71.229.25 count 10000 size 300  
CK> PING 212.71.229.25 (212.71.229.25): 300 data bytes
CK> 
CK>
!.!!.!!.!!!!.!.!....!!.!!!!!..!..!!.!!!..!!.!!....!!.!...!!!...!.!.!..!.
CK>  .!!...!....!..!.!.!!.!!.!!....!.!!!!!!..!..!.!!.!..!!.!!^C
CK>  --- 212.71.229.25 ping statistics ---
CK>  129 packets transmitted, 66 packets received, 48% packet loss  
CK> round-trip min/avg/max/stddev = 1.023/1.640/34.095/4.035 ms
CK>  
CK>  ckoch at core2.lhr3> ping rapid 212.71.229.25 count 10000 size 400  
CK> PING 212.71.229.25 (212.71.229.25): 400 data bytes  
CK> !!!!!!!!.!!!!!...!!^C
CK>  --- 212.71.229.25 ping statistics ---  20 packets transmitted, 15 
CK> packets received, 25% packet loss  round-trip min/avg/max/stddev = 
CK> 1.074/1.111/1.348/0.071 ms
CK>  
CK>  ckoch at core2.lhr3> ping rapid 212.71.229.25 count 10000 size 500  
CK> PING 212.71.229.25 (212.71.229.25): 500 data bytes  
CK> !!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!.!^C
CK>  --- 212.71.229.25 ping statistics ---
CK>  49 packets transmitted, 46 packets received, 6% packet loss  
CK> round-trip min/avg/max/stddev = 1.141/1.202/2.514/0.210 ms
CK>  
CK>  ckoch at core2.lhr3> ping rapid 212.71.229.25 count 10000 size 600  
CK> PING 212.71.229.25 (212.71.229.25): 600 data bytes
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK>  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!^C
CK>  --- 212.71.229.25 ping statistics ---
CK>  955 packets transmitted, 954 packets received, 0% packet loss  
CK> round-trip min/avg/max/stddev = 1.183/1.526/36.436/1.817 ms
CK>  
CK>  koch at core2.lhr3> ping rapid 212.71.229.25 count 10000 size 292  
CK> PING 212.71.229.25 (212.71.229.25): 292 data bytes
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK> 
CK>
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
CK>  !!!!!!!!^C!
CK>  --- 212.71.229.25 ping statistics ---
CK>  916 packets transmitted, 916 packets received, 0% packet loss  
CK> round-trip min/avg/max/stddev = 0.988/1.146/9.918/1.061 ms
CK>  
CK>  ckoch at core2.lhr3> ping rapid 212.71.229.25 count 10000 size 293  
CK> PING 212.71.229.25 (212.71.229.25): 293 data bytes  ..!!!.!^C
CK>  
CK>  Christian
CK>  _______________________________________________
CK>  juniper-nsp mailing list juniper-nsp at puck.nether.net  
CK> https://puck.nether.net/mailman/listinfo/juniper-nsp
CK>   
CK>   
CK>    
 




More information about the juniper-nsp mailing list