Read the FAQs at snapshot, the code or the back traffic.  While -A was
pulled as a command line parameter, the code is still present in
initialize.c, allowing you to - with a source change - revert to the prior
performance.  Look around line 400 in initialize.c.

Remember, you're running BOTH processes on the same box.  This means that
you are handling each packet at least twice through the TCP/IP stack and
then through the application programs.  I'm guessing that's why you see
40.5% user, 58.2% system.  What do you see when you shutdown ntop?

-----Burton


-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Frank
Raiser
Sent: Saturday, June 15, 2002 3:02 PM
To: [EMAIL PROTECTED]
Subject: Re: [Ntop] CPU load monitoring ftp


On Sat, Jun 15, 2002 at 02:55:55PM -0500, Burton M. Strauss III wrote:
> So, some comparison data is in order.
>
> (I used nntp, not ftp - it's easier to get a steady measurable load)
>
> CPU load (top): 0.7% to 1.5%
> Hardware: Pentium III 800 (100MHz FSB), ample free memory

I'm running on a mere Celeron 300.

> Configuration: running solely as a monitor (i.e. the monitoring box is NOT
> the NAT/Firewall/Gateway).

Similar. There's only the ftp server. No NAT, firewalling or gateway.

> Average network load: 1300 Kbps
> Network configuration: Cable Modem, capped at 1500 Kbps (so that's
> effectively pegged...)

Connected to a 100mbit LAN. Usually gives ftp downloads at a rate of
about 9MB/s. Slowed down to 5MB/s when ntop is running.

>
>   2:47pm  up 33 days, 20:20,  1 user,  load average: 0.00, 0.00, 0.00
> 34 processes: 31 sleeping, 3 running, 0 zombie, 0 stopped
> CPU states:  1.1% user,  1.3% system,  0.0% nice, 97.4% idle
> Mem:   125848K av,  121280K used,    4568K free,       0K shrd,   42196K
> buff
> Swap:  257032K av,    7700K used,  249332K free                   13916K
> cached

  9:59pm  up 6 days,  3:03,  1 user,  load average: 0.88, 0.21, 0.07
  46 processes: 41 sleeping, 5 running, 0 zombie, 0 stopped
  CPU states: 40.5% user, 58.2% system,  0.0% nice,  1.1% idle
  Mem:   255892K av,  252404K used,    3488K free,       0K shrd,   34552K
buff
  Swap: 1024120K av,    3292K used, 1020828K free                  143256K
cached

    PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
    12506 ftp       25   0  1196 1196   776 R    49.9  0.4   0:11 proftpd
    12505 ntop      25   0  8052 6204  1368 R    47.9  2.4   0:08 ntop
    12509 root      15   0  1024 1024   836 R     0.5  0.4   0:00 top

> If you are processor OR memory constrained, you are going to see a
> performance hit, of course...

I was supposing that 300Mhz should be able to handle this?
Is it possible to tweak ntop to improve performance? I don't really need
all the details (like I don't need protocol decoding, or tcp session
handling).
The stable 2.0 release had a -A switch that enabled me to do this, but I
haven't found anything comparable in the current version?

--
Raiser, Frank
_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://lists.ntop.org/mailman/listinfo/ntop

_______________________________________________
Ntop mailing list
[EMAIL PROTECTED]
http://lists.ntop.org/mailman/listinfo/ntop

Reply via email to