[Kamailio-Users] crashing problem

[Jonathan K. Creasy]

Thanks for the assistance. So far I have not been able to re-create the problem in the lab but have had the crash with identical back traces (same location) on 7 different border servers.

More RAM and less children seems to have led to stability but I will still be trying to recreate the problem and fix the root cause.

-Jonathan

-----Original Message-----
From: Daniel-Constantin Mierla [mailto:miconda at gmail.com]
Sent: Monday, August 25, 2008 3:56 AM
To: Jonathan K. Creasy
Cc: Henning Westerholt; users at lists.kamailio.org
Subject: Re: [Kamailio-Users] crashing problem

Hello,

as it seems to happen in a place that look safe from this point of view,
might be the case of memory corruption (overwrite, double free, ...).

Needs to compile in memory debug mode, guidelines at:
http://www.kamailio.org/dokuwiki/doku.php/troubleshooting:memory

When run in debug mode, a double free or overwrite is reported in the log.

Cheers,
Daniel


On 08/25/08 11:42, Henning Westerholt wrote:
> On Saturday 23 August 2008, Jonathan K. Creasy wrote:
>
>> We're using Carrier route with about 53,000 routes stored in a mysql
>> databse. We're getting this crash in production that we never had in the
>> lab.
>>
>> The bt is below. Does anyone have any suggestions?
>>
>> nline static void free_to_params(struct to_body* tb)
>> {
>>         struct to_param *tp=tb->param_lst;
>>         struct to_param *foo;
>>         while (tp){
>>                 foo = tp->next;   <-- ****  Line 75 *****
>>                 pkg_free(tp);
>>                 tp=foo;
>>         }
>> }
>>
>> #0  free_to (tb=0x817cf10) at parser/parse_to.c:75
>> #1  0x080daf70 in clean_hdr_field (hf=0x8a8ee18) at parser/hf.c:182
>> #2  0x080db05b in free_hdr_field_lst (hf=0x8a905b8) at parser/hf.c:209
>> #3  0x080db12d in free_sip_msg (msg=0x8193030) at parser/msg_parser.c:660
>>
>
> Hi Jonathan,
>
> this seems not related to carrierroute, as you can see from the backtrace the
> crash is in the parser code. Would be interesting if this is related to some
> invalid message, or an unrelated error. Have you tried to replay the message
> in your lab that causes the crash, after isolating it from the backtrace?
> You should also take a look into the "tp" pointer with the debugger, try to
> investigate why it has an apparently invalid value here.
>
> Cheers,
>
> Henning
>
> _______________________________________________
> Users mailing list
> Users at lists.kamailio.org
> http://lists.kamailio.org/cgi-bin/mailman/listinfo/users
>
>

--
Daniel-Constantin Mierla
http://www.asipto.com