Greger V. Teigre wrote:
Ah, of course you knew about that :-) Well, then I cannot really see why the SER children locked up. Do you have any SER logs right before locking up?
I looked at them but there was really nothing. Only the errors indicating the RTPPROXY stopped responding.
You are sure it's related to a non-responding rtpproxy?
No, but its too much of a coincidence. We monitor the RTPPROXY from a Nagios script every 5 minutes so we can confirm externally that it was not responding around the same time as the 4 children locked up.
I remember that Maxim was not able to reproduce the rtpproxy lockup, so that's one hard bug to nail down. Hopefully you will get something useful out of the debugging.
Yea I know. We left RTPROXY running in the foreground (to gather info for Maxim) for like 2 months. Millions of calls and all was perfect. We restarted it normally about 3 weeks ago and we got this problem. I am starting to think its more stable running in the foreground.
Thanks,