[Serusers] Problem with MTU size in SER with MIPv6.

[Andrei Pelinescu-Onciul]

On Nov 24, 2006 at 11:45, Jose Angel Calvo <jacc15dkz at yahoo.es> wrote:
> Thanks for your fast answer, Andrei. I answer to your questions:
> 
> > Could you give more details?
> Of course :-)
> > - what version of ser are you running (ser -V)?
> The exit in my screen are:
> version: ser 0.9.6 (i386/linux)
> flags: STATS: Off, USE_IPV6, USE_TCP, DISABLE_NAGLE, USE_MCAST, DNS_IP_HACK, SHM_MEM, SHM_MMAP, PKG_MALLOC, F_MALLOC, FAST_LOCK-ADAPTIVE_WAIT
> ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 16, MAX_URI_SIZE 1024, BUF_SIZE 65535
> @(#) $Id: main.c,v 1.197.2.1 2005/07/25 16:56:24 andrei Exp $
> main.c compiled on 11:38:44 Feb  1 2006 with gcc 3.3.6
> 
> > - is it vanilla ser or a modified version?
> It's a vanilla version.
> > - the packet is sent over TCP? UDP?
> Over UDP.
> > - a coredump backtrace? (for this you'll have to start ser with
> >  -w directory_where_core_can_be_written, e.g.  ser -f ser.cfg -w /tmp/)
> > After the crash, run gdb path_to_ser/ser /tmp/core, type bt
> >  and send me the output. Note: the core file might have another name
> >   (e.g. core.3456). Just look for core.* .
> I do what you say me, but thre is not a coredump backtrace. Next, i show you the syslog:

Could you try again after runing:
ulimit -c unlimited   #  you need bash for this
ser -f  ser.cfg -w /tmp/ 

(or if unlimited doesn't work,  some large value, e.g. ulimit -c 100000) ?

If it still doesn't work, could you send me a network dump with the
 INVITE causing the crash and your config, so that I can try to
 reproduce it?


Thanks,
Andrei