I wonder, since it's related to -M, if you are seeing the same host on more
than one interface??  If so, it could be that it's being freed on the 1st
device's pass so that when the 2nd device pass comes along it's already
free... --trace-level 5 will enable the NOISY messages that show the purge
details.  That might have a clue too...

It could also be another manifestation of the memory problem I posted this
weekend - Once the memory chains are corrupted, ANYTHING can happen.

I wonder, can you run under valgrind or one of the other memory monitoring
tools?  Running under Valgrind, I don't seem to get anything other than the
ultimately fatal situation, but there's got to be a root cause back in time
(usually it's a double free).  If you can tag that, all kinds of problems
just disappear.


-----Burton



-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf
Of Matthias Krebs
Sent: Wednesday, October 08, 2003 10:36 AM
To: [EMAIL PROTECTED]
Subject: RE: [Ntop-dev] ntop crashes when capturing multiple interfaces


Am Mit, 2003-10-08 um 17.18 schrieb Burton M. Strauss III:
> Does it fail when run under gdb?  If so, capture the bt full output and
post
> that.  Otherwise, sadly, the bug report is about equivalent to saying
you've
> lost you keys somewhere in New York...
>
> -----Burton
Hi Burton,

gdb output is (with set args -K --user root -M -i "eth0,eth1,eth2,eth3"
-P /usr/share/ntop --https-server 192.168.18.11:3001 --http-server 0 -L
--use-syslog=daemon --db-file-path /usr/share/ntop -m $mynetworks ):

(gdb) info stack
#0  0x42074760 in _int_free () from /lib/tls/libc.so.6
#1  0x42073786 in free () from /lib/tls/libc.so.6
#2  0x40090481 in ntop_safefree (ptr=0x42132320, file=0x400ab678
"hash.c", line=273) at leaks.c:609
#3  0x4008a630 in freeHostInfo (host=0x86321b8, actualDeviceId=1) at
hash.c:273
#4  0x4008aed8 in purgeIdleHosts (actDevice=1) at hash.c:532
#5  0x4009239d in scanIdleLoop (notUsed=0x0) at ntop.c:607
#6  0x40247332 in start_thread () from /lib/tls/libpthread.so.0

(gdb) bt full
(gdb) bt full
#0  0x42074760 in _int_free () from /lib/tls/libc.so.6
No symbol table info available.
#1  0x42073786 in free () from /lib/tls/libc.so.6
No symbol table info available.
#2  0x40090481 in ntop_safefree (ptr=0x42132320, file=0x400ab678
"hash.c", line=273) at leaks.c:609
No locals.
#3  0x4008a630 in freeHostInfo (host=0x86321b8, actualDeviceId=1) at
hash.c:273
        i = 1074474144
#4  0x4008aed8 in purgeIdleHosts (actDevice=1) at hash.c:532
        idx = 3
        numFreedBuckets = 3
        numHosts = 54
        startTime = 1065627060
        purgeTime = 1065627030
        lastPurgeTime = {1065627051, 1065627060, 1065626929, 1065626931,
0 <repeats 28 times>}
        firstRun = 0 '\0'
        theFlaggedHosts = (HostTraffic **) 0x84d9fe8
        maxHosts = 42
        scannedHosts = 98
        hiresTimeStart = {tv_sec = 1065627060, tv_usec = 781176}
        hiresTimeEnd = {tv_sec = 1065627060, tv_usec = 780691}
        el = (HostTraffic *) 0x4000
        prev = (HostTraffic *) 0x0
        next = (HostTraffic *) 0x84da090
#5  0x4009239d in scanIdleLoop (notUsed=0x0) at ntop.c:607
No locals.
#6  0x40247332 in start_thread () from /lib/tls/libpthread.so.0
No symbol table info available.

I hope this information is sufficient.

Bye,

Matthias

_______________________________________________
Ntop-dev mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev

_______________________________________________
Ntop-dev mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev

Reply via email to