Pcap to Argus Data Oddness

Peter Van Epp vanepp at sfu.ca
Fri May 11 13:03:59 EDT 2007


On Fri, May 11, 2007 at 09:01:34AM +0800, CS Lee wrote:
> Hey Peter,
> 
> Greetings. I have sent this pcap file to Carter but it doesn't show problem
> to him when he generates argus data from the pcap attached. However when I
> tried with openbsd, freebsd and gentoo box to convert the pcap to argus
> data, the saddr and daddr becomes odd. Here's the data shown in tcpdump and
> argus in comparison -
> 
> sudo tcpdump -ttttnnr icmpshell.pcap
> reading from file icmpshell.pcap, link-type EN10MB (Ethernet)
> 2006-09-01 01:36:27.511558 arp who-has 192.168.0.119 tell 192.168.0.250
> 2006-09-01 01:36:27.511600 arp reply 192.168.0.119 is-at 00:0a:e4:35:ea:8e
> 2006-09-01 01:36:27.511776 IP 192.168.0.250 > 192.168.0.119: ICMP echo
> reply, id 60165, seq 0, length 39
> 2006-09-01 01:36:27.514256 IP 192.168.0.119 > 192.168.0.250: ICMP echo
> reply, id 60165, seq 256, length 90
> 2006-09-01 01:36:32.497343 IP 192.168.0.250 > 192.168.0.119: ICMP echo
> reply, id 60165, seq 0, length 39
> 2006-09-01 01:36:32.499682 IP 192.168.0.119 > 192.168.0.250: ICMP echo
> reply, id 60165, seq 512, length 90
> 2006-09-01 01:36:34.829683 IP 192.168.0.250 > 192.168.0.119: ICMP echo
> reply, id 60165, seq 0, length 37
> 2006-09-01 01:36:39.166263 IP 192.168.0.250 > 192.168.0.119: ICMP echo
> reply, id 60165, seq 0, length 37
> 2006-09-01 01:36:41.219730 IP 192.168.0.250 > 192.168.0.119: ICMP echo
> reply, id 60165, seq 0, length 43
> 2006-09-01 01:36:41.222004 IP 192.168.0.119 > 192.168.0.250: ICMP echo
> reply, id 60165, seq 768, length 41
> 2006-09-01 01:36:59.823108 IP 192.168.0.250 > 192.168.0.119: ICMP echo
> reply, id 60165, seq 0, length 52
> 2006-09-01 01:36:59.824779 IP 192.168.0.119 > 192.168.0.250: ICMP echo
> reply, id 60165, seq 1024, length 547
> 2006-09-01 01:36:59.824803 IP 192.168.0.119 > 192.168.0.250: ICMP echo
> reply, id 60165, seq 1280, length 547
> 2006-09-01 01:36:59.824825 IP 192.168.0.119 > 192.168.0.250: ICMP echo
> reply, id 60165, seq 1536, length 547
> 2006-09-01 01:36:59.824844 IP 192.168.0.119 > 192.168.0.250: ICMP echo
> reply, id 60165, seq 1792, length 420
> 2006-09-01 01:37:04.820877 arp who-has 192.168.0.119 tell 192.168.0.250
> 2006-09-01 01:37:04.820910 arp reply 192.168.0.119 is-at 00:0a:e4:35:ea:8e
> 
> After converting it into argus data, it looks like this -
> 
> ra -nnr icmpshell.arg
>   01:36:27.511558  e         2054      192.168.0.250          who
> 192.168.0.119               2        2          120           84   CON
>   01:36:27.511776  e            1      250.0.168.192           ->
> 119.0.168.192               6        0          451            0   ECR
>   01:36:27.514256  e            1      119.0.168.192           ->
> 250.0.168.192               1        0          124            0   ECR
>   01:36:32.499682  e            1      119.0.168.192           ->
> 250.0.168.192               1        0          124            0   ECR
>   01:36:41.222004  e            1      119.0.168.192           ->
> 250.0.168.192               1        0           75            0   ECR
>   01:36:59.824779  e            1      119.0.168.192           ->
> 250.0.168.192               1        0          581            0   ECR
>   01:36:59.824803  e            1      119.0.168.192           ->
> 250.0.168.192               1        0          581            0   ECR
>   01:36:59.824825  e            1      119.0.168.192           ->
> 250.0.168.192               1        0          581            0   ECR
>   01:36:59.824844  e            1      119.0.168.192           ->
> 250.0.168.192               1        0          454            0   ECR
>   21:15:13.099275             man                  0
> 0                       37      1       17       10           37
> 1493508   STP
> 
> The address 192.168.0.250 becomes 250.0.168.192 and 192.168.0.119 becomes
> 119.0.168.192. Since I read that you are running argus on various platform,
> then I think it would be good to verify with you, I just want to confirm if
> this odd behavior only happens to me. Hereby I attach with the pcap and
> argus data.
> 
> Thanks.
> 
> 
> 
> -- 
> Best Regards,
> 
> CS Lee<geekooL[at]gmail.com>

	You are correct, clients.rc.44 has an endian problem:

FreeBSD 6.2 on Intel:

2.0.6:

%argus_bpf -r icmpshell.pcap -w icmpshell.argus2
%ra -r icmpshell.argus2
11 May 07 15:30:38           man  229.97.122.203  v2.0                   1 0     0        0         0            0           STA
31 Aug 06 10:36:32          icmp   192.168.0.119        ->   192.168.0.250       1        0         124          0           ECR
31 Aug 06 10:36:27          icmp   192.168.0.119        ->   192.168.0.250       1        0         124          0           ECR

3.0 clients.rc.44 on FreeBSD (wrong, note the arp is correct just icmp isn't
correctly converted from network byte order):

%ra3 -r icmpshell.argus3 -n
   10:36:27.511776  e         icmp      250.0.168.192           ->      119.0.168.192               2        0          146            0   ECR
   10:36:27.511558  e          arp      192.168.0.250          who      192.168.0.119               1        1           60           42   CON
   10:36:27.514256  e         icmp      119.0.168.192           ->      250.0.168.192               1        0          124            0   ECR
   10:36:32.499682  e         icmp      119.0.168.192           ->      250.0.168.192               1        0          124            0   ECR

clients.rc.44 on Mac OS 10 on PowerPC:

test4:~ vanepp$ ra3 -r icmpshell.argus3 -n
   10:36:27.511776            icmp      192.168.0.250           ->      192.168.0.119               2        0          146            0   ECO
   10:36:27.511558             arp      192.168.0.250          who      192.168.0.119               1        1           60           42   CON
   10:36:27.514256            icmp      192.168.0.119           ->      192.168.0.250               1        0          124            0   ECO
   10:36:32.499682            icmp      192.168.0.119           ->      192.168.0.250               1        0          124            0   ECO

	It appears the the endian macros aren't being invoced correctly on 
the Intel boxes for at least icmp.

Peter Van Epp / Operations and Technical Support 
Simon Fraser University, Burnaby, B.C. Canada



More information about the argus mailing list