RE: XML compile issues on Fedora Core 2

Hi Burton,

Thanks for that.  Deleted out all the problem variables from the inc
files and everything compiled fine.
But now when I run ntop I get the following:
Mon Oct  4 10:24:05 2004  **WARNING** Unable to load plugin
'/usr/local/lib/ntop/plugins/xmldumpPlugin.so'
Mon Oct  4 10:24:05 2004  **WARNING** Message is
'/usr/local/lib/ntop/plugins/xmldumpPlugin.so: undefined symbol:
BufferTooShort'
Mon Oct  4 10:24:05 2004  **WARNING** Unable to load plugin
'/usr/local/lib/ntop/plugins/snmpPlugin.so'
Mon Oct  4 10:24:05 2004  **WARNING** Message is
'/usr/lib/libnetsnmpagent.so.5: undefined symbol: hosts_ctl'

Where should these be defined?

Thanks
Craig 

> -----Original Message-----
> From: ntop-dev-bounces <at> unipi.it 
> [mailto:ntop-dev-bounces <at> unipi.it] On Behalf Of Burton M. Strauss III
> Sent: Saturday, September 25, 2004 12:30 AM
> To: ntop-dev <at> Unipi.IT
> Subject: RE: [Ntop-dev] XML compile issues on Fedora Core 2
> 
> Actually, the deleted variables should be REMOVED from the 
> xml .inc files,
> not the other way around.

RE: XML compile issues on Fedora Core 2

Sounds like a code problem - I replaced the buffer handling code (globally)
with a new version that handles the glibc changes.  Maybe that didn't make
it into the plugin - check any of the other modules for the syntax.

-----Burton

> -----Original Message-----
> From: ntop-dev-bounces <at> unipi.it [mailto:ntop-dev-bounces <at> unipi.it]On
> Behalf Of Craig Humphrey
> Sent: Sunday, October 03, 2004 5:57 PM
> To: ntop-dev <at> Unipi.IT
> Subject: RE: [Ntop-dev] XML compile issues on Fedora Core 2
>
>
> Hi Burton,
>
> Thanks for that.  Deleted out all the problem variables from the inc
> files and everything compiled fine.
> But now when I run ntop I get the following:
> Mon Oct  4 10:24:05 2004  **WARNING** Unable to load plugin
> '/usr/local/lib/ntop/plugins/xmldumpPlugin.so'
> Mon Oct  4 10:24:05 2004  **WARNING** Message is
> '/usr/local/lib/ntop/plugins/xmldumpPlugin.so: undefined symbol:
> BufferTooShort'
> Mon Oct  4 10:24:05 2004  **WARNING** Unable to load plugin
> '/usr/local/lib/ntop/plugins/snmpPlugin.so'
> Mon Oct  4 10:24:05 2004  **WARNING** Message is
> '/usr/lib/libnetsnmpagent.so.5: undefined symbol: hosts_ctl'
>
> Where should these be defined?

Ntop problem report PR_PVUJV96: example of missing stats

Hi again,

Oups! I forgot to include the sample page I mentioned in my problem report a
few minutes ago; here it is.

Save the .html and 3 .png files in the same directory, open the .html file.

Regards,

Marc.

_______________________________________________
Ntop-dev mailing list
Ntop-dev <at> unipi.it
http://listgateway.unipi.it/mailman/listinfo/ntop-dev

Ntop problem report PR_PVUJV96: Network Load Statistics missing first stats

** this problem is also present with 3.0 and 3.0.53
** (see "Problem description" below)

  n t o p   v e r s i o n  '2.2.97'  p r o b l e m   r e p o r t

From: Marc Mazuhelli

EMail: marc.mazuhelli <at> usherbrooke.ca

Date:  2004-10-04 13:31:19 GMT

Problem Report Id: PR_PVUJV96

----------------------------------------------------------------------------
Summary
Missing information in Network Load Statistics page

OS: RedHat Linux  version: 9

ntop from: source

Hardware:  CPU:           i686
           # Processors:  1 real, 2 simulated (Xeon)
           Memory:        1536 MB

Packets
Received:  1259096057
Processed: 1259096057 (immediately)
Queued:             0
Lost:               0 (queue full)

[FreeBSD] Ntop stuck in swread state?

Hi all,

As of late (the last 2 weeks or so) I have to restart NTop every few 
hours, or it crashes, with all available swap being eaten.

If I look at top for about 2 minutes, I can actually see the amount of 
swap space increasing rapidly. This is the results that I see in top:

last pid:  8449;  load averages:  0.03,  0.05,  0.07   up 32+20:16:41 
17:20:25
27 processes:  1 running, 26 sleeping
CPU states:  0.4% user,  0.0% nice,  5.8% system,  3.1% interrupt, 90.7% 
idle
Mem: 97M Active, 13M Inact, 36M Wired, 5868K Cache, 25M Buf, 528K Free
Swap: 512M Total, 488M Used, 24M Free, 95% Inuse, 1592K In, 128K Out

   PID USERNAME PRI NICE   SIZE    RES STATE    TIME   WCPU    CPU COMMAND
94609 root     -20    0   517M   107M swread 626:38  0.00%  0.00% ntop

Note the NTop process that seems to be stuck in the swread state.

I might add that I have not seen the behavior in NTop until about 2 
weeks ago when it just started eating away on memory.

This is a relatively small network, and NTop is installed in a bridge 
configuration.

Now, my question: Is there any known problems that can cause this sort 
of behavior? I know that the version of NTop that I run is a bit older, 
but, it did not have these problems before.

RE: [FreeBSD] Ntop stuck in swread state?

There were changes between pre2 and the two release candidates and then the
actual 3.0 release - whether those matter, I don't really recall.  But we
don't support pre anything, obviously.

Basically, swap usually means more hosts - and once you overfill real
memory, ntop's performance sucks. It can be bad enough to not allow
processes on timers to complete, backing them up, deadlocking, etc.

1. Are you using --set-pcap-nonblocking?
2. The real question is what changed in your network so that you are seeing
many more hosts.  The info.html and textinfo.html pages show the number of
hosts...

-----Burton

> -----Original Message-----
> From: ntop-dev-bounces <at> unipi.it [mailto:ntop-dev-bounces <at> unipi.it]On
> Behalf Of Jaco van Tonder
> Sent: Monday, October 04, 2004 10:41 AM
> To: ntop-dev <at> Unipi.IT
> Subject: [Ntop-dev] [FreeBSD] Ntop stuck in swread state?
>
>
> Hi all,
>
> As of late (the last 2 weeks or so) I have to restart NTop every few
> hours, or it crashes, with all available swap being eaten.
>
> If I look at top for about 2 minutes, I can actually see the amount of
> swap space increasing rapidly. This is the results that I see in top:

New ntop commit (author deri)

Update of /export/home/ntop/ntop
In directory unknown:/tmp/cvs-serv12250

Modified Files:
	admin.c globals-defines.h globals-structtypes.h initialize.c 
	main.c pbuf.c prefs.c 
Log Message:
Added packet sampling into ntop for allowing ntop to run on places
where there are too many packets to process. Sampling can be enabled
via a new command line flag or via web configuration.

New ntop commit (author deri)

Update of /export/home/ntop/ntop
In directory unknown:/tmp/cvs-serv27307

Modified Files:
	traffic.c 
Log Message:
Added fix for host matrix and IPv6

Ntop problem report PR_M9ACAPX : stats missing on Summary->Network load graphs

  n t o p   v e r s i o n  '3.0.053'  p r o b l e m   r e p o r t

From:  Marc Mazuhelli

EMail: marc.mazuhelli at usherbrooke.ca

Date:  Tue Oct 12 13:11:20 2004

Problem Report Id: PR_M9ACAPX

----------------------------------------------------------------------------
Summary
stats missing on Summary->Network load graphs

OS(uname): sysname(Linux) release(2.4.21-20.ELsmp) version(#1 SMP Wed Aug 18
20:46:40 EDT 2004) machine(i686)

ntop from: source

Hardware:  CPU:           i686
           # Processors:  1 real, 2 simulated with SMP kernel (Xeon)
           Memory:        1536 MB

Packets
Received:  1259450838
Processed: 1259450837 (immediately)
Queued:             0
Lost:               0 (queue full)
Queue:     Current: 0 Maximum: 0

Help: ntop for windows

Hi,
I'm sorry for my request because, maybe many people ask before the 
same think but in you faq I can't find nothing about, then I think to 
write you.

I'm from Italy,
for testing I had installed ntop on win98 but when open my browser on 
http://127.0.0.1:3000 the server don't work and ntop send me a 
message:

C:\Programmi\ntop-Win32>ntop
10/13/04 12:25:54  ntop v.3.0 MT
10/13/04 12:25:54  Configured on (null), built on 22/03/03 08:43:08 PM.
10/13/04 12:25:54  Copyright 1998-2004 by Luca Deri <deri <at> ntop.org>
10/13/04 12:25:54  Get the freshest ntop from http://www.ntop.org/
10/13/04 12:25:54  Initializing ntop
10/13/04 12:25:58  Resetting traffic statistics for device IBM 10_100 
EtherJet C
ardBus Adapter
10/13/04 12:25:58  DLT: Device 0 [IBMC] is 1, mtu 1514, header 14
10/13/04 12:25:58  Initializing gdbm databases
10/13/04 12:25:58  **FATAL_ERROR** ....open of ./prefsCache.db failed: Can't be
writer
10/13/04 12:25:58  1. Is another instance of ntop running?
10/13/04 12:25:58  2. Make sure that the use you specified can write 
in the targ
et directory

The directory ntop-Win32 is writable, readable and executable but I 
don't understand why the file prefsCache.db make error.