Summary:	ASTERISK-12883: Segmentation Fault with 1.4.21.2 in rtp.c:1131, Application chanSpy
Reporter:	Wolfgang Pichler (wuwu)	Labels:
Date Opened:	2008-10-13 12:23:01	Date Closed:	2008-12-10 12:37:40.000-0600
Priority:	Critical	Regression?	No
Status:	Closed/Complete	Components:	Core/RTP
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) backtrace_11_10_08.txt
Description:	i do have encountered a segmentation fault on a 1.4.21.2 machine while using the ChanSpy Application - i have tried to reproduce the segfault - but without success. I think the segfault is because of two different sip users where trying to start ChanSpy on the same channel - but i don't know exactly... I am running asterisk with optimized code - so the backtrace isn't as usefull as it could be - if i am able to reproduze the fault - then i will run it on a non optimized version to produce a more useable backtrace (how much does it really costs to disable the optimizations - will it be a problem on a production machine running at loadavg around 1, cpu utilization around 30 % ?)
Comments:	By: Mark Michelson (mmichelson) 2008-10-13 14:21:52 This issue appears to be similar to issue ASTERISK-12347, so I'm going to mark them related. One thing that I think will be helpful is to get valgrind output when this problem occurs. For instructions on how to use valgrind with Asterisk, see doc/valgrind.txt in the Asterisk source directory. For the load you are talking about, turning off optimizations should not make a huge difference in performance. Running valgrind, on the other hand, will cause noticeable slowdown, and so you may want to try a test run with valgrind before launching it on a production system. By: Mark Michelson (mmichelson) 2008-12-10 12:35:38.000-0600 I have become convinced that this is a duplicate of issue ASTERISK-12347. Since it has many more comments and backtraces attached, it seems like the logical place to continue trying to fix this.