Hi Ntop developers,
I took over maintenance of the Ntop Debian package along with Jordan
Metzmeier.
We have uploaded version 4.0.3 to unstable and now we are trying to get
reported bugs fixed.
First of all, I have a question: which one is the official stable release?
Ntop homepage says 4.0.3, and for this reason we packaged 4.0.3.
However sourceforge offers you to download 4.0.1, and ntop 4.0.3 itself
complains with:
CHKVER: Checking current ntop version at version.ntop.org/version.xml
CHKVER: Version file is from 'version.ntop.org'
CHKVER: as of date is '2010-10-06T01:12:25'
CHKVER: This version of ntop is the current DEVELOPMENT version - Expect
the unexpected!
Second: is there any bug tracker? (I could not find one...)
Or the preferred way for reporting bugs is to this ML?
Then, I would like to report a couple of issues I am having on a machine
I am trying to run ntop.
1)
I often get my syslog flooded with the message:
**ERROR** Buffer too short @ report.c:4013 (increase to at least 1742)
Yes, I changed the source code and the problem went away, but asking a
user to change the source code and recompile it is too user-unfriendly,
especially if the user got a pre-compiled package
2)
I am getting consistently corruped HTML like
<img class=tooltip alt="OS: Linux" title="OS: Linux" align="middle"
src="/statsicons/os/linux.gif"> <img class=tooltip src="/bulb.gif"
border="0" alt="DHCP Client" title="DHCP Client"> <img
class=tooltip src="/multihomed.png" border="0" alt="Multihomed"
title="Multihomed"><img class=tooltip src="/card.gif" border="0"
alt="Network Card" title="Network Card"> <img class=tooltip
src="/router.gif" border="0" alt="Router" title="Router"> <img
class=tooltip src="/dns.gif" border="0" alt="DNS" title="DNS"> <img
class=tooltip src="/web.gif" border="0" alt="HTTP Server" title="HTTP
Server"> <img class=tooltip src="/clock.gif" border="0" alt="NTP
Server" title="NTP Server"> <img class=tooltip src="/Risk_medium.gif"
border="0" alt="Medium Risk" title="Medium Risk"> <img class=tooltip sr
followed by 3 non printable characters (<80>`<82>). Then, the next HTML
cells starts. This happens at least in the network throughput report
page. Being it the longest host information string I am seeing, I
thought it was related to issue 1 (see above), but it did not actually
go away after changing the buffer size, and I am not seeing other
"buffer too short" messages.
3)
if I do not use the -L (log to syslog) options, still some messages end
up in the syslog, like the above "Buffer too short" message and
THREADMGMT[t140737354008320]: ntop RUNSTATE: PREINIT(1)
4)
I am running ntop on a machine with multiple interfaces for a total
thoughput of about 40 Mbps, and I get a ntop crash (segfault,
glibc-detected double-free, silent exit) every couple of hours. I am
trying to get a useful report about this. Is there any private email
address I can send a link to a core dump? (I do not want to publicly
post it).
Btw, looks like that running with "-n 0" prevents the crashes.
Still, if I shut it down form the web interface, I get a segfault:
Mon 24 Jan 2011 02:33:48 PM PST WEB: shutdown.html - request has been
received - processing
Mon 24 Jan 2011 02:33:48 PM PST WEB: Beginning actual shutdown sequence
Mon 24 Jan 2011 02:33:48 PM PST THREADMGMT[t140737106753280]: ntop
RUNSTATE: SHUTDOWNREQ(6)
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fffe2ffd700 (LWP 20288)]
_int_malloc (av=0x7fffe4000020, bytes=126) at malloc.c:4438
4438 malloc.c: No such file or directory.
in malloc.c
(gdb) bt
#0 _int_malloc (av=0x7fffe4000020, bytes=126) at malloc.c:4438
#1 0x00007ffff71ac58e in *__GI___libc_malloc (bytes=126) at malloc.c:3660
#2 0x00007ffff7702e97 in ntop_safestrdup (ptr=0x7fffe2ffb810 "Mon 24
Jan 2011 02:33:48 PM PST THREADMGMT[t140737001805568]:
NPS(eth2(unfiltered)): pcapDispatch thread terminated [p20282]",
file=<value optimized out>,
line=<value optimized out>) at leaks.c:203
#3 0x00007ffff7728510 in traceEvent (eventTraceLevel=3, file=<value
optimized out>, line=148, format=<value optimized out>) at util.c:2812
#4 0x00007ffff77040a0 in pcapDispatch (_i=0x1) at ntop.c:148
#5 0x00007ffff68c79ca in start_thread (arg=<value optimized out>) at
pthread_create.c:300
#6 0x00007ffff721670d in clone () at
../sysdeps/unix/sysv/linux/x86_64/clone.S:112
#7 0x0000000000000000 in ?? ()
or
Mon 24 Jan 2011 02:35:59 PM PST WEB: shutdown.html - request has been
received - processing
Mon 24 Jan 2011 02:35:59 PM PST WEB: Beginning actual shutdown sequence
Mon 24 Jan 2011 02:35:59 PM PST THREADMGMT[t140737185306368]: ntop
RUNSTATE: SHUTDOWNREQ(6)
Mon 24 Jan 2011 02:35:59 PM PST THREADMGMT[t140737157678848]:
NPS(eth2(unfiltered)): pcapDispatch thread terminated [p22959]
*** glibc detected *** /usr/sbin/ntop: malloc(): memory corruption:
0x00007ffff7ff4010 ***
[Thread 0x7fffec4a4700 (LWP 22966) exited]
======= Backtrace: =========
/lib/libc.so.6(+0x775b6)[0x7ffff71a75b6]
/lib/libc.so.6(+0x7b6d8)[0x7ffff71ab6d8]
/lib/libc.so.6(__libc_malloc+0x6e)[0x7ffff71ac58e]
/usr/lib/ntop/libntop-4.0.3.so(ntop_safestrdup+0x37)[0x7ffff7702e97]
/usr/lib/ntop/libntop-4.0.3.so(traceEvent+0x370)[0x7ffff7728510]
/usr/lib/ntop/libntopreport-4.0.3.so(handleWebConnections+0x4cf)[0x7ffff7ba1f2f]
/lib/libpthread.so.0(+0x69ca)[0x7ffff68c79ca]
/lib/libc.so.6(clone+0x6d)[0x7ffff721670d]
I tried to run it with valgrind: this time it did not crash, but it
reported invalid read and *writes* in the logging code. Maybe this is
what is corrupting the heap.
Here is the valgrind output:
http://www.cs.ucsb.edu/~cavedon/ntop-valgrind.log
(I am running on 64 bit Ubuntu 10.04).
Thanks for your help!
Cheers,
Ludovico
_______________________________________________
Ntop-dev mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev
