Have you enabled one core per process (pid)? If not, core files from other processes may be overwritten by the main process doing cleanup at shut down. Anyhow, if a worker process crashes, you should see a log message in syslog. not only the one about signal 15.

If you don't get a response, then the sip trace is not that useful for the moment.

First thing to double check -- if you enabled pike module, be sure you don't block the traffic that is trusted.

Otherwise, if you can reproduce, do following:

• when kamailio runs fine, do kamctl ps to get the list of running processes
• when kamailio stops routing sip traffic, attach with gdb to a worker process for your kind of traffic (UDP, TCP ... for example for UDP, attach to the PID listed for the process with ID 1) and get the backtrace:

gdb /path/to/kamailio PID
bt full

Also see if the CPU is at high load. That could be an indication of a dead (spin) lock. If CPU is low, maybe there is some blocking operation dns, db query, ...

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.