Hello,

I run vnstat-1.4 and netacct-mysql-0.78 to monitor bandwidth usage. The host in question has two interfaces and 99% of traffic goes through eth0, while eth1 is idle most of the time (it's my backup interface). VNSTAT shows a huge spike in usage for one day for eth1. Average for this interface is typically below 25mb per day and on this one day it's 8GB (4GB in and 4GB out)

day rx | tx | total
------------------------+-------------+--------------
24.06. 0 MB | 25.47 MB | 25.47 MB
25.06. 0 MB | 1.71 MB | 1.71 MB
26.06. 0 MB | 1.44 MB | 1.44 MB
27.06. 0 MB | 14.01 MB | 14.01 MB
28.06. 0 MB | 0.39 MB | 0.39 MB
29.06. 0 MB | 3.97 MB | 3.97 MB
30.06. 0 MB | 0.24 MB | 0.24 MB
01.07. 0 MB | 40.25 MB | 40.25 MB
02.07. 0 MB | 2.25 MB | 2.25 MB
03.07. 0 MB | 1.02 MB | 1.02 MB
04.07. 0 MB | 0.66 MB | 0.66 MB
05.07. 0 MB | 0.01 MB | 0.01 MB
06.07. 0 MB | 0.14 MB | 0.14 MB
07.07. 0 MB | 0.02 MB | 0.02 MB
08.07. 4,095 MB | 4,097 MB | 8,193 MB
09.07. 0 MB | 5.73 MB | 5.73 MB
10.07. 0 MB | 0.19 MB | 0.19 MB
11.07. 0 MB | 15.63 MB | 15.63 MB
------------------------+-------------+--------------
estimated 0 MB | 25 MB | 25 MB

The part that confounds me is that NETACCT shows no such spike for any of the IPs it's monitoring (and it's monitoring all valid IPs).

Any ideas? Or suggestions where I might look to get more info? Or anything else I could implement to help me track bandwidth usage?