[Flow-tools] flow-stat summary problem
David Mitchell
mitchell at ucar.edu
Thu Feb 21 09:48:30 EST 2008
Greetings,
I've hit what appears to be a bug in flow-stat. In our flow files
(exported from JunOS) the first and last time entries fall into two sets
of values. One is a pretty standard and reasonable value such as
1235577361, but some flows have values which are offset and are more in
the range of 3732739987. The result is that flow-stat generates a bogus
value for duration and hence average bandwidth:
Duration of data (realtime) : 300
Duration of data (1/1000 secs) : 2497164628
Average Kbits / second (flow) : 0.0433
Average Kbits / second (real) : 360.4539
I would write this off as being a problem with my flows, but flow-print
has no problem printing the flows with correct first and last time
values. flow-print calls the function ftltime() to massage the values,
as opposed to flow-stat which just uses the raw first and last values.
Should flow-stat be calling ftltime() as well? I haven't looked into
this too deeply, and most of the time I can just use the (realtime)
values. The little bit of poking around I did do though made me think
that this could be fixed, but I'm not sure where to get access to all
the values ftltime() needs in flow-stat. Is this a known bug though?
-David Mitchell
--
-----------------------------------------------------------------
| David Mitchell (mitchell at ucar.edu) Network Engineer IV |
| Tel: (303) 497-1845 National Center for |
| FAX: (303) 497-1818 Atmospheric Research |
-----------------------------------------------------------------
More information about the Flow-tools
mailing list