5 Jun
2009
5 Jun
'09
11:12 a.m.
On Jun 05, 2009 at 11:44, Juha Heinanen <jh(a)tutpro.com> wrote:
when i read the "last words" debug more
closely, it appears that the
proxy handling three requests at the time when it goes into coma: it is
executing save function from one register request, getting another
register request, and getting a notify from presence server.
any pointers on how to find out why sr becomes unresponsive are greatly
appreciated. how can it happen that the proxy stops doing anything even
though its processes (or at least some of them) are still alive?
Could you attach with gdb to some processes and send us a backtrace?
(gdb ser <pid_of_process> ; bt )
You could start sr with a small number of processes (e.g. -l eth0 -n 2)
and then dump the backtrace from all of them using:
for r in `pidof ser`; do echo PID $r >>/tmp/bt.log;
gdb -batch -ex bt ser $r >>/tmp/bt.log ; done
Please send also the new log (with maximum debugging) so that we can try to
correlate and the config (you could send only to me and Jan if) or at
least the list of the modules loaded.
Thanks,
Andrei