[Oisf-users] As I enabled the packet profiling, packet drop was generated.

이상철 leesc at ictis.kr
Tue Jun 6 10:23:30 UTC 2017


HI!

 

As I enabled the packet profiling, packet drop was generated.

The packet drop was confirmed by “packet.kernel_drop” in stats.log.

I want to always enable the packet profiling in my product, how can I
improve?

 

I generated 160 Kpps(packet per sec) traffic and lost 30% of the packets.
The size of each packet size is 64 bytes.

I have analyzed the source code and have found that packet drop will not
occur if the SCProfilingPrintPacketProfile() function is not executed.

How do I improve this function and are there any other improvement options?

 

Below is the hardware specification and suricata configuration information.

 

Hardware spec.

CPU : Intel(R) Xeon(R) CPU E5-2660 v3 @ 2.60GHz

      10 cores, 20 threads

Memory : 16G

Ethernet : Intel Corporation 82580 Gigabit ( 1Gbps)

 

 

Suricata Version : 3.2.1

OS - linux debian 8

excute command :  "suricata -c /etc/suricata/suricata.yaml --af-packet"

number of rules : 9000

suricata.yaml

af-packet:

  - interface: eth0

    threads: auto

    cluster-id: 99

    cluster-type: cluster_flow

    defrag: yes

    use-mmap: yes

    buffer-size: 64535

    copy-mode: ips

    copy-iface: eth1

 

  - interface: eth1

    threads: auto

    cluster-id: 98

    cluster-type: cluster_flow

    defrag: yes

    use-mmap: yes

    buffer-size: 64535

    copy-mode: ips

    copy-iface: eth0

 

packets:

    # Profiling can be disabled here, but it will still have a

    # performance impact if compiled in.

    enabled: yes

    filename: packet_stats.log

    append: yes

 

    # per packet csv output

    csv:

 

      # Output can be disabled here, but it will still have a

      # performance impact if compiled in.

      enabled: yes

      filename: packet_stats.csv

 

 

Sincerely

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openinfosecfoundation.org/pipermail/oisf-users/attachments/20170606/1c796b5d/attachment.html>


More information about the Oisf-users mailing list