I can re-create this crash fairly much at will. Deleting the dnscachedb
does fix it, until it recurs some time after. I have a box with a
reasonable traffic load, and it segV's within a few minutes of starting
up after the first crash. Until the first crash, it runs pretty well for
a day or so.
This is ntop svn, checked out yesterday, compiled on Debian 4.0.
Here's a backtrace of the crash from "ntop -t 6 -c -x 65536 -i eth0"
The debuggler is still attached, if anyone wants me to execute commands.
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1354769488 (LWP 24789)]
0xb755d500 in _gdbm_get_bucket () from /usr/lib/libgdbm.so.3
(gdb) bt
#0 0xb755d500 in _gdbm_get_bucket () from /usr/lib/libgdbm.so.3
#1 0xb755db79 in _gdbm_alloc () from /usr/lib/libgdbm.so.3
#2 0xb755bec5 in gdbm_store () from /usr/lib/libgdbm.so.3
#3 0xb78b29f7 in ntop_gdbm_store (g=0x80f1488, d=
{dptr = 0xaf3fa980 "1255987146", dsize = 11}, v=
{dptr = 0xaf3fa9e0 "www.ontariohockeyacademy.com", dsize = 72}, r=1,
theFile=0xb79093d2 "protocols.c", theLine=659) at leaks.c:835
#4 0xb78cd9af in processDNSPacket (srcHost=0x8a344b0, sport=53,
packetData=0xaf3fb10e "\216\217\204", length=<value optimized out>,
isRequest=0xaf3fac40, positiveReply=0xaf3fac3e) at protocols.c:659
#5 0xb78bf173 in processIpPkt (bp=0xaf3fb0f2 "E", h=0xaf3fd14c, length=198,
ether_src=0xaf3fb044 "", ether_dst=0xaf3fb03e "", actualDeviceId=0,
vlanId=65535) at pbuf.c:1754
#6 0xb78c44a3 in processPacket (_deviceId=0x0, h=0xaf3fd14c,
p=<value optimized out>) at pbuf.c:3686
#7 0xb78c903d in queuePacket (_deviceId=0x0, h=0xaf3fd14c, p=0x80cbad0 "")
at pbuf.c:2505
#8 0xb78ed3df in pcap_read_packet (handle=0x80cb8a8,
callback=0xb78c8d80 <queuePacket>, userdata=0x0) at ./pcap-linux.c:802
#9 0xb78ecee6 in pcap_read_linux (handle=0x80cb8a8, max_packets=-1,
callback=0xb78c8d80 <queuePacket>, user=0x0) at ./pcap-linux.c:469
#10 0xb78ef134 in pcap_dispatch (p=0x80cb8a8, cnt=-1,
callback=0xb78c8d80 <queuePacket>, user=0x0) at ./pcap.c:77
---Type <return> to continue, or q <return> to quit---
#11 0xb78b68cc in pcapDispatch (_i=0x0) at ntop.c:94
#12 0xb7884240 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#13 0xb77eb49e in clone () from /lib/tls/i686/cmov/libc.so.6
--
ian
Kevin Freels wrote:
> Sorry, forgot to update.
>
> I deleted the db files and made sure it was running with debug ("-t 6"),
> and it hasn't crashed in a week.
>
> Go figure.
>
> I'll leave it like this until the next problem happens. Thanks for the
> help!!!!
>
> ....k
> -=-=-=-
>
>> -----Original Message-----
>> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
>> Behalf Of Gary Gatten
>> Sent: Friday, August 15, 2008 11:17 AM
>> To: [email protected]
>> Subject: Re: [Ntop] ntop dying with SegFault at line 26
>>
>> Just for kicks delete the db files - especially the dnscache.
>> Let us know if that helps - maybe rename the old ones and
>> save them for analysis if need be!
>>
>> Gary
>>
>>
>> -----Original Message-----
>> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
>> Behalf Of Kevin Freels
>> Sent: Friday, August 15, 2008 12:53 PM
>> To: [email protected]
>> Subject: Re: [Ntop] ntop dying with SegFault at line 26
>>
>> Greetings!
>>
>> The crashes are starting to happen again. It ran fine for a
>> few weeks, and then suddenly started dying about three weeks
>> ago. Nothing has been changed on the system it's running on
>> (FC7). Here's the flags for the command line:
>>
>> NTOPCfg=" -w 3001 -i eth0,eth1 -M -K -z -t 6 -x 4096"
>>
>> Attached is the stdout from the start up until the crash, it
>> ran about 50 minutes.
>>
>> ....k
>> -=-=-=-
>>
>>> -----Original Message-----
>>> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
>> On Behalf
>>> Of Kevin Freels
>>> Sent: Monday, June 02, 2008 1:42 PM
>>> To: [email protected]
>>> Subject: Re: [Ntop] ntop dying with SegFault at line 26
>>>
>>> Luca,
>>>
>>> I have started ntop with the CLI you prescribed. Here is
>> the history
>>> of the commands I had set up in the init.d script (commented-out
>>> configs are ones I tried in succession):
>>>
>>> NTOPBin=/usr/local/bin/ntop
>>> #NTOPCfg=" -w 3001 -i eth0,eth1 -M -c -d -t 5"
>>> #NTOPCfg=" -w 3001 -i eth0,eth1 -M -c -t 5"
>>> #NTOPCfg=" -w 3001 -i eth0 -c -t 5"
>>> #NTOPCfg=" -w 3001 -i eth0 -c -t 5 -x 4096"
>>> #NTOPCfg=" -w 3001 -i eth0,eth1 -M -z -t 5 -x 4096"
>>> NTOPCfg=" -w 3001 -i eth0,eth1 -M -K -z -t 6 -x 4096"
>>>
>>> Over this last weekend, I limited the hosts hash and --sticky-hosts
>>> with "-x 4096" and dropped the "-c". It ran all weekend
>> long until the
>>> disk filled up (but not from ntop, another issue). So perhaps the
>>> --sticky-hosts and the host hash limits are the trick....? In any
>>> event, it seems to be working now.
>>>
>>> I'll keep an eye on it for the week (or until it crashes).
>> If it keeps
>>> going, maybe I'll fool around with the host hash limit,
>> incrementally
>>> raising it. If anything happens to cause it to die, I'll forward on
>>> the gruesome details.
>>>
>>> Thanks for the reply!!!
>>>
>>> ....k
>>> -=-=-=-
>>>
>>>> -----Original Message-----
>>>> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
>>> On Behalf
>>>> Of Luca Deri
>>>> Sent: Sunday, June 01, 2008 12:30 AM
>>>> To: [email protected]
>>>> Subject: Re: [Ntop] ntop dying with SegFault at line 26
>>>>
>>>> Kevin
>>>> in what setup are you running ntop? WAN perhaps? What is
>>> the exact CLI
>>>> command you're using? Can you please start it from shell
>>> and add -K -t
>>>> 6 and send the log.
>>>>
>>>> Thanks Luca
>>> _______________________________________________
>>> Ntop mailing list
>>> [email protected]
>>> http://listgateway.unipi.it/mailman/listinfo/ntop
>>>
>>
>>
>>
>>
>> <font size="1">
>> <div style='border:none;border-bottom:double windowtext
>> 2.25pt;padding:0in 0in 1.0pt 0in'> </div> "This email is
>> intended to be reviewed by only the intended recipient and
>> may contain information that is privileged and/or confidential.
>> If you are not the intended recipient, you are hereby
>> notified that any review, use, dissemination, disclosure or
>> copying of this email and its attachments, if any, is
>> strictly prohibited. If you have received this email in
>> error, please immediately notify the sender by return email
>> and delete this email from your system."
>> </font>
>>
>> _______________________________________________
>> Ntop mailing list
>> [email protected]
>> http://listgateway.unipi.it/mailman/listinfo/ntop
>>
> _______________________________________________
> Ntop mailing list
> [email protected]
> http://listgateway.unipi.it/mailman/listinfo/ntop
>
>
_______________________________________________
Ntop mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop
