At 11:42 06/06/2006, Cesc wrote:

Hi,

I think this issue was already risen a while back, but seen that there is no answer, i will insist :D

My SER becomes completely unresponsive (UDP and TCP) after a while. I started it 3 days ago and have a monitoring script that sends 1 REGISTER every 10 secs (plus a few other phones). This morning that SER was totally stoned! :(

I have tried sending via UDP and TCP ... same answer (that is, no answer).

But, the fifo (thus, part of SER) still works (i can do a serctl moni).

I am trying to investigate myself ... what can i do? can i force a core dump from the shell (sending some kill signal) to the already stoned SER?

Hi Cesc,

looking at WCHAN may be useful. (ps -l) generating core dump works fine too.

That's the things I would be looking at. I thought of adding some WCHAN equivalent to SER which would allow the FIFO process (if it does not hang itself) to figure out where the other processes are hanging.

-jiri

-- Jiri Kuthan http://iptel.org/~jiri/