Correct me if I am wrong, your idea of stats collection is based on
iptables. This is beautiful by itself as it is directly related to firewall
rules.
On the other hand, if one's purpose is to measure qdisc/classses, wouldn't
it be better to use 'tc' , much like those explained in HTB user guide
chapter 7 ?
Examples, tc -s -d qdisc show dev eth0
tc -s -d class show dev eth0
etc.
In any case, whether is using tc or iptables, it will require a fork to
collect
results. If I need to run it continuously, any words of advise regarding
avoiding excess load ?
Regards.
----- Original Message -----
From: "Stef Coene" <stef.coene@docum.org>
To: <mingching.tiew@redtone.com>; <lartc@mailman.ds9a.nl>
Sent: Saturday, January 25, 2003 1:37 AM
Subject: Re: [LARTC] tools for traffic monitoring
> I wrote some scripts for long term graphing and a script for real time
> logging. You can find them all on www.docum.org. (gui : long term
logging,
> monitorl : real time monitoring). They work fine for me.
>
> > Appreciate someone could point me where in the '/proc' file system where
> > these data could be gathered.
> There is no tc interface for the /proc interface.
>
> Stef
>
> --
>
> stef.coene@docum.org
> "Using Linux as bandwidth manager"
> http://www.docum.org/
> #lartc @ irc.oftc.net
>
>
>
