I wouldn't call it rules, just a reminder that the effective parameters include three sets, visible on the command line, @files and the stored ones from the web interface. 
 
 
My main development system is SMP FC2 and it has NPTL and doesn't hang (smp makes the problem MUCH more likely).  Understand that the cvs fix for 'hanging daemon' is really for syslog() un-thread-safe problems.  Could be that your system is upgraded from RH7/8 and may have a left over pre-NPTL glibc?
 
Here is what I would do:
 
1. Try the cvs version as it.  I would *not* expect anything different.
 
2. Copy the -DFORPRENPTL line from configureextra/LINUXredhat8.0 into the configureextra/LINUXfedora file.  ./configure, make and make install.
 
 
If #2 fixes it, then your NPTL system is exhibiting preNPTL behavior and I'd investigate the upgrade scenario, look for older glibc, etc.
 
-----Burton


From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of John Hally
Sent: Thursday, March 31, 2005 9:01 AM
To: '[email protected]'
Subject: RE: [Ntop] Ntop not collecting data

Thanks Burton.  Sorry about not following the rules, I'll make much better effort in the future. 

 

 

I did find that it only seems to happen when trying to put it in daemon mode.  Would you care to see the output of the bug report button from Ntop?  I'm going to give the CVS version a try as well.

 

Thanks!

 

 

 

 

 


From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Burton Strauss
Sent: Wednesday, March 30, 2005 5:26 PM
To: [email protected]
Subject: RE: [Ntop] Ntop not collecting data

 

Well, 60% of that is useless (repeated parms / defaults... --user vs -u, -i vs --interface and --http-server)

 

The only meaningful differences are --db and --daemon

 

So...

 

1. Does the ntop userid have read/write access to the --db directory?  Without the parameter ntop falls back to the default (which I don't even remember what that is).

2. There was some stuff re -d in the back traffic and some patches - try the cvs version.

 

 

------Burton

 


From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of John Hally
Sent: Wednesday, March 30, 2005 4:03 PM
To: '[email protected]'
Subject: RE: [Ntop] Ntop not collecting data

Ok this is really weird.

 

If I start ntop by hand :  /usr/local/bin/ntop -u ntop -i eth1  things work great and it collects data just fine 

 

BUT

 

If I start ntop like this: /usr/local/bin/ntop -u ntop -i eth1 @/etc/ntop.conf  it collects no data.  I whittled the conf file down to this:

 

 

# grep -v '#' /etc/ntop.conf

 

--user ntop

 

--daemon

 

--db-file-path /usr/local/share/ntop

 

--interface eth1

 

--http-server 3000

 

 

 

What's wrong with that?  It makes no sense to me at all.

 

??


From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of John Hally
Sent: Wednesday, March 30, 2005 2:35 PM
To: '[email protected]'
Subject: [Ntop] Ntop not collecting data

 

Hello All,

 

I have a really strange issue.

 

Running ntop 3.1 on Fedora Core2.  I just ran through a quick install (./configure, make, make install) with no problems but when I start up ntop, it's not seeing any data whatsoever.  I've double checked that there's actually data flowing to that interface (SPAN port on a cisco 2950, ran tcpdump and see plenty of packets flying around).  Ntop reports that it puts the correct interface into promiscuous mode (eth1 in this case) started up with correct user, all the pertinent directories are owned by this user, but I still get no data. 

 

I tried throwing it into debug mode (-K) but it doesn't look like it's dumping any more data to syslog

 

I'm stumped.  Any suggestions?

 

Thanks!

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

Reply via email to