Racluster packet/byte counts off
Carter Bullard
carter at qosient.com
Wed Mar 26 18:16:12 EDT 2008
Hey Nick,
You may be truncating your pkts count because of a small field width.
Try "pkts:12" to see if your digits comes back.
The tools are suppose to put a '*' at the end of fields that have been
truncated
due to the width field being too small. I may need to fix that if this
is all there
is.
Carter
Nick Diel wrote:
> I have been merging records on mac address pairs and notice at some
> large packet count the merged records counts are off.
>
> Here is an example of the problem:
>
> [diel at lander-nic ~]$ racount -r /capture/dump-110707/pcap/*.argus.out
> - ether src X:X:X:X:X:52 or ether dst
> X:X:X:X:X:52
> ;
> racount records total_pkts src_pkts dst_pkts
> total_bytes src_bytes dst_bytes
> sum 411,639,887 *4,167,751,642* 3241106807
> 926644835 2,631,640,888,213 1886412015303 745228872910
>
>
> [diel at lander-nic ~]$ racluster -r
> /capture/dump-110707/pcap/*.argus.out -m smac dmac -L0 -s smac dir
> dmac bytes pkts - ether src X:X:X:X:X:52 or ether dst X:X:X:X:X:52
> SrcMac Dir DstMac TotBytes TotPkts
> X:X:X:X:X:52 -> X:X:X:X:X:d 779300
> *14,401*
> X:X:X:X:X:52 -> X:X:X:X:X:a
> 500060 *4,546*
> X:X:X:X:X:3c <-> X:X:X:X:X:52 2631632608 *41,677,326
>
> *The packet count from racluster is basically 1% of what racount is.
>
>
> Here it works on a smaller scale:
>
> [diel at lander-nic ~]$ racount -r
> /capture/dump-110707/pcap/anondump.0001.pcap.argus.out - ether src
> X:X:X:X:X:52 or ether dst X:X:X:X:X:52
> racount records total_pkts src_pkts dst_pkts
> total_bytes src_bytes dst_bytes
> sum 573210 5396094 4200746 1195348
> 3458510664 2561659100 896851564
>
>
> [diel at lander-nic ~]$ racluster -r
> /capture/dump-110707/pcap/anondump.0001.pcap.argus.out -m smac dmac
> -L0 -s smac dir dmac bytes pkts - ether src 7X:X:X:X:X:52 or ether dst
> X:X:X:X:X:52
> SrcMac Dir DstMac TotBytes TotPkts
> X:X:X:X:X:52 -> X:X:X:X:X:d 10700 22
> X:X:X:X:X:52 -> X:X:X:X:X:a
> 770 7
> X:X:X:X:X:3c <-> X:X:X:X:X:52 3458499194 5396065
>
>
> Nick
More information about the argus
mailing list