Try this patch - I'm working on it to rationalize the drop counts.

It won't fix the drops (that's either NIC, libpcap or ntop simply not
accepting them fast enough from libpcap), but it should make the stats more
rational in both the PR_ and info.html/textinfo.html.

-----Burton


-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of David
Touitou
Sent: Friday, October 24, 2003 10:56 AM
To: [EMAIL PROTECTED]
Subject: Re: [Ntop] Dropped by the kernel




Luca Deri wrote:

> David,

Luca,

> tell me more about
> - how you use ntop (options)

The only option I'm using is -M -w 0 -W 0
I tried with -b and -n but did not notive any change.

Current command line is :
/usr/local/bin/ntop --interface fxp1 --user nobody --db-file-path
/var/db/ntop -M -w 0 -W 0

I've asked the rrdplugin to rrd everything.

> - whether you sniff only from network or also from NetFlow or similaria

Only network : I have a 3com switch that mirrors the traffic of one port
(uplink to the internet) to another one (fxp1 on the probe).

CPU usage (P3-733) went up to 41% at night (a bit more than 10Mbps
bandwidth and a lot of udp packets of game servers).

I'll re-launch ntop with syslog logging and stop it by monday morning in
order to get some stats from syslog (last one were 15 millions packets
with 0.4% of packets dropped by kernel and 0% by ntop).
I've understood that webstats were wrong.

David.

_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop

Attachment: BMS0382-queuedropcount.patch
Description: Binary data

Reply via email to