[netflow-tools] flows timers control

Sebastian Majkowski smajko at wp.pl
Mon May 4 19:24:47 EST 2009


alex k wrote:
>> Hi
>>
>> I use softflowd simultaneously with my router. I noticed that softflowd
>> is generating almost twice more softflow info than my cisco. I guess
>> thats because of timouts which I am able to set on cisco and use some
>> cache before sending netflow information to my collector.
>> Is it possible to set it also with softflowd? In other words:
>> - generate netflow information when flow is not active for X seconds (I
>> want to set X value, what is softflowd default?) (cisco: |ip flow-cache
>> timeout inactive X)|**
>> - generate netflow information when flow is active for Y seconds (and if
>> still active repeat this after another Y seconds (what is the default?))
>> (cisco: *ip flow-cache active-timeout Y)
>> I guess that softflowd timers are different than cisco ones, I probably
>> get more detailed info with softflowd but I am running out of resources
>> and just need to save on this.
>>
>> Any ideas?
>>
>> Regards
>>
>> S.M
>> *
>> _______________________________________________
>> netflow-tools mailing list
>> netflow-tools at mindrot.org
>> https://lists.mindrot.org/mailman/listinfo/netflow-tools
>>
>>     
>
> Hi S.,
>
> I'm not sure, if I understand your problem as I don't have experience with
> cisco routers. Running out of resources probably means memory (or cpu?).
> Softflowd has several timeouts. See "man softflowd", section "Timeouts".
> The one you are searching for might be "maxlife".
> Try something like "-t maxlife=5m". Then all flows will be expired after 5
> minutes (sent to collector).
> Alternatively you could reduce the maximum number of flows to concurrently
> track with "-m" or the track_level with for instance "-T proto" (less
> detailed, but less resources needed).
>
> Hope this helps.
>
> xela
>
>
>
>
>   
Hi Alex,

Thanks for this info. It looks that it works fine now, maxlife is what I 
needed (and expint). Files at collector seems to be similar in size.
But I still have some performance issues. My server is dedicated only 
for softflowd, and I process around 300Mbits/s
My CPu is AMD Opteron with 2 cores but only one core is used for 
softflowd process. My system is
netflow:~# uname -a
Linux netflow 2.6.26-2-amd64 #1 SMP Fri Mar 27 04:02:59 UTC 2009 x86_64 
GNU/Linux
Is it possible to use both cores?? Maybe different OS can do this?
btw, I cant use -T proto because I need full netflow info (TCP/UDP ports 
also) Take a look at my top chart:

Tasks:  51 total,   2 running,  49 sleeping,   0 stopped,   0 zombie
Cpu0  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  
0.0%st
Cpu1  : 71.9%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa, 13.0%hi, 15.1%si,  
0.0%st
Mem:   4064312k total,  1020408k used,  3043904k free,   106800k buffers
Swap:  9928128k total,        0k used,  9928128k free,   808820k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
25147 nobody    20   0 30036  16m 2748 R  100  0.4  19:56.48 softflowd
    1 root      20   0 10316  756  628 S    0  0.0   0:02.08 init
    2 root      15  -5     0    0    0 S    0  0.0   0:00.00 kthreadd
    3 root      RT  -5     0    0    0 S    0  0.0   0:00.00 migration/0
    4 root      15  -5     0    0    0 S    0  0.0   0:00.02 ksoftirqd/0
    5 root      RT  -5     0    0    0 S    0  0.0   0:00.08 watchdog/0
    6 root      RT  -5     0    0    0 S    0  0.0   0:00.08 migration/1
    7 root      15  -5     0    0    0 S    0  0.0   0:14.76 ksoftirqd/1
    8 root      RT  -5     0    0    0 S    0  0.0   0:00.16 watchdog/1
    9 root      15  -5     0    0    0 S    0  0.0   0:05.56 events/0
   10 root      15  -5     0    0    0 S    0  0.0   0:09.06 events/1
   11 root      15  -5     0    0    0 S    0  0.0   0:00.00 khelper
   46 root      15  -5     0    0    0 S    0  0.0   0:00.02 kblockd/0
   47 root      15  -5     0    0    0 S    0  0.0   0:00.28 kblockd/1
   49 root      15  -5     0    0    0 S    0  0.0   0:00.00 kacpid
   50 root      15  -5     0    0    0 S    0  0.0   0:00.00 kacpi_notify
  146 root      15  -5     0    0    0 S    0  0.0   0:00.04 ksuspend_usbd
  152 root      15  -5     0    0    0 S    0  0.0   0:00.00 khubd
  155 root      15  -5     0    0    0 S    0  0.0   0:00.00 kseriod
  203 root      20   0     0    0    0 S    0  0.0   0:00.00 pdflush
  204 root      20   0     0    0    0 S    0  0.0   0:02.52 pdflush
  205 root      15  -5     0    0    0 S    0  0.0   0:00.00 kswapd0
  206 root      15  -5     0    0    0 S    0  0.0   0:00.00 aio/0
  207 root      15  -5     0    0    0 S    0  0.0   0:00.22 aio/1
  743 root      15  -5     0    0    0 S    0  0.0   0:00.00 ata/0
  744 root      15  -5     0    0    0 S    0  0.0   0:00.00 ata/1
  745 root      15  -5     0    0    0 S    0  0.0   0:00.00 ata_aux
  812 root      15  -5     0    0    0 S    0  0.0   0:00.00 scsi_eh_0
  813 root      15  -5     0    0    0 S    0  0.0   0:00.00 scsi_eh_1
  933 root      15  -5     0    0    0 S    0  0.0   0:09.20 kjournald
 1008 root      16  -4 16512  756  488 S    0  0.0   0:00.12 udevd
 1863 daemon    20   0  8024  520  404 S    0  0.0   0:00.00 portmap
 1874 statd     20   0 10152  760  636 S    0  0.0   0:00.00 rpc.statd
 2097 root      20   0  180m 1728 1032 S    0  0.0   0:39.62 rsyslogd
 2111 root      20   0  3808  624  500 S    0  0.0   0:00.00 acpid
 2121 messageb  20   0 21096  536  344 S    0  0.0   0:00.00 dbus-daemon
 2137 root      20   0 48872 1180  676 S    0  0.0   0:02.60 sshd


Regards


Sebastian



More information about the netflow-tools mailing list