Title: Ericsson Signature
Hi,

Reading the README.1st file in PF_RING/userland/snort/pfring-daq-module it states that socket clustering can be used to distribute packets across multiple process instances. I assume that's how it is possible to get large throughput with snort as you run many instances. I tried running snort again without clustering but still no joy.

/usr/local/bin/snort  -A console --daq-dir=/usr/local/lib/daq --daq pfring --daq-var bindcpu=3 -i eth8:eth9 -Q

# ping 192.168.0.10
PING 192.168.0.10 (192.168.0.10) 56(84) bytes of data.
64 bytes from 192.168.0.10: icmp_seq=1 ttl=64 time=2276 ms
64 bytes from 192.168.0.10: icmp_seq=2 ttl=64 time=1276 ms
64 bytes from 192.168.0.10: icmp_seq=3 ttl=64 time=276 ms
64 bytes from 192.168.0.10: icmp_seq=4 ttl=64 time=1278 ms
64 bytes from 192.168.0.10: icmp_seq=5 ttl=64 time=1279 ms
64 bytes from 192.168.0.10: icmp_seq=6 ttl=64 time=1280 ms
64 bytes from 192.168.0.10: icmp_seq=7 ttl=64 time=1282 ms

//Norbert

On 06/03/14 20:34, Luca Deri wrote:
Why do you use a cluster?

On 03/06/2014 08:10 AM, Norbert Furlani wrote:
Hi,

I've downloaded and compiled the latest svn PF_RING and snort (2.9.6.0).

pf_ring kernel module loaded (ransparent_mode=0 min_num_slots=32768 enable_tx_capture=0)

$ cat /proc/net/pf_ring/info
PF_RING Version          : 5.6.3 ($Revision: exported$)
Total rings              : 2

Standard (non DNA) Options
Ring slots               : 32768
Slot version             : 15
Capture TX               : No [RX only]
IP Defragment            : No
Socket Mode              : Standard
Transparent mode         : Yes [mode 0]
Total plugins            : 0
Cluster Fragment Queue   : 0
Cluster Fragment Discard : 0


 and snort executed as follows (no rules):-

# /usr/local/bin/snort  -A console --daq-dir=/usr/local/lib/daq --daq pfring  --daq-var fast-tx=1 --daq-var clusterid=10,11 --daq-var bindcpu=3 -i eth8:eth9 -Q

I have one host connected on eth8 and one on eth9. When I ping one host from the other, I see the packet header in the console and the ICMP replies are received, however the RTTs are over 1000ms. ssh or ftp across the link is unusable.

By comparison, if I use afpacket daq, performance is excellent.

On a side note, if I invoke another instance of snort using pf_ring, I do not see a load balance of packets either.

Anyone have similar problems? I'm thinking of trying ubuntu 13.10 to see if it makes any difference.


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

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

--


line

NORBERT FURLANI
IT Begins With US


Ericsson Australia
ITTE Service Delivery Australia
6/153 Bertie Street
Port Melbourne 3207, Australia
Phone +61 3 9301 3239
Mobile +61 416 131 772
[email protected]
www.ericsson.com



Ericsson




This Communication is Confidential. We only send and receive email on the basis of the terms set out at www.ericsson.com/email_disclaimer


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

Reply via email to