Will, 
I recently had this same trouble with pf_ring snaplen. 
The API function does not work properly if you didn't set the correct 
"bucket_len" arg when you load the ring module . 
The default BUCKET_LEN value is set to 96 or 128 bytes , probably for capture 
optimization

so, to correct this strange issue, try it: 

rmmod ring
insmod path_to_ring/ring.ko bucket_len=1600

and restart your apps , it will be OK

regards,
Diogo Sato 

www.predialnet.com.br
PredialNet Wireless ISP



----- Mensagem Original -----
De: Will Metcalf <[EMAIL PROTECTED]>
Para: [email protected]
Data: Monday, 28 De July De 2008 18:18
Assunto: [Ntop-dev] strange issue with latest pf_ring

> I'm having a weird issue with the latest pf_ring that allows user
> specified snaplen/caplen to tell pf_ring what the bucket length should
> be.  Everything works fine until I restart one of the apps argus which
> has a smaller snaplen than snort and daemonlogger both with a snaplen
> of 1515.  Once I do this both deamonlogger and snort start to only
> capture 96 bytes of traffic that is specified as the argus snaplen.
> If I change the argus snaplen to to 1515 everything is fine, but I
> don't want to do that as the boxes are overtaxed as it is, and I need
> to roll the argus files via cron daily....
> 
> Regards,
> 
> Will
> _______________________________________________
> Ntop-dev mailing list
> [email protected]
> http://listgateway.unipi.it/mailman/listinfo/ntop-dev
> 

_______________________________________________
Ntop-dev mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev

Reply via email to