On 7/8/08 7:25 AM, "Steve Clark" <[EMAIL PROTECTED]> wrote:

> James Lay wrote:
>> On 7/8/08 6:17 AM, "Andrei-Florian Staicu" <[EMAIL PROTECTED]> wrote:
>> 
>> 
>>> Hello all,
>>> 
>>> I seem to have a small problem when running ntop 3.3.6, listening on
>>> interfaces other than eth0.
>>> For example:
>>> /usr/bin/ntop -u nobody --sticky-hosts --use-syslog=local6 -i eth0
>>> will run for ages, while
>>> /usr/bin/ntop -u nobody --sticky-hosts --use-syslog=local6 -i eth1
>>> on the same machine
>>> will run for 5-30 seconds and yield Segmentation fault.
>>> The same is true for another machine: on eth0 it works, on eth3 I get
>>> Segmentation fault.
>>> The machines have different hardware configurations; one of them is a
>>> Slackware 12.1 and the latter is a Slackware 12.0.
>>> 
>>> Do you have any ideas on why this could happen?
>>> If you need any more information, such as the logs, please let me know
>>> and I will post them, although I don't see differences between the logs
>>> between the two runs.
>>> 
>>> Thanks in advance.
>>> 
>>> 
>>> _______________________________________________
>>> Ntop mailing list
>>> [email protected]
>>> http://listgateway.unipi.it/mailman/listinfo/ntop
>> 
>> 
>> 
>> Do interfaces eth1 or eth3 have a valid IP assigned to them?
>> 
>> 
>> _______________________________________________
>> Ntop mailing list
>> [email protected]
>> http://listgateway.unipi.it/mailman/listinfo/ntop
>> 
>> 
> In my experiencing running NTOP on FreeBSD I have found it to be a great tool,
> but it seems to be very fragile.
> 
> My $.02
> Steve


I'm don't believe the OP was running a FreeBSD box, but my experience with
nTop on FreeBSD lead me to run nTop with on Linux +PF_RING.  My distance
varied with FreeBSD 7.0 RELEASE (i386 and amd64) +libpcap  and a Intel PT
quad port card.

It seems it only ran for about 13-20 hours and then just "gave up".  The
binary would still be running, and the web pages would still be active, but
the libpcap dropped packets skyrocketed up to about 90%.  The listening card
would still be in promisc, and off of a mirrored port.  If I stopped the
daemon, and restarted it, it would start working immediately.

My mileage also varied using the Intel quad cards under -RELEASE and
-CURRENT, so I switched from what I was raised on for this instance.  I have
seven sensors, and nTop has been up on each for 10+ days now after the
latest OS patch.

Regards,
David

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

Reply via email to