Summary:	ASTERISK-08079: Asterisk system crashs main server
Reporter:	peter (peter)	Labels:
Date Opened:	2006-11-07 10:52:13.000-0600	Date Closed:	2007-03-26 12:51:27
Priority:	Critical	Regression?	No
Status:	Closed/Complete	Components:	Core/General
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) bt.txt ( 1) bt2.txt
Description:	Our asterisk system consists of 4 remote asterisk servers accepting calls from a MAX tnt.  The four asterisk box's relay the call to our main server.  If we shutdown down one of the remote asterisk servers the MAX tnt will relay the call to another remote server, to relay to the main server.  But when that happens our main server crashes with a segfault. From the MAX tnt it is SIP to the remote servers, and the remote server to the main server is iax. ****** ADDITIONAL INFORMATION ****** OS: FC3 Kernel: 2.6.11-1.35_FC3 This was the last upgrade and we are still experiencing the problem. I am only able to reproduce this crash but only on our live system.
Comments:	By: Joshua C. Colp (jcolp) 2006-11-07 12:46:54.000-0600 Can you please confirm that this is an unoptimized build, and if not rebuild your Asterisk unoptimized? This is accomplished by doing: make clean make dont-optimize make install The overhead because of this is negligible and should not affect things. By: peter (peter) 2006-11-07 13:07:18.000-0600 This is a unoptimized build. By: Serge Vecher (serge-v) 2006-11-14 08:55:42.000-0600 peter: is the reason you use an svn checkout of 1.2 branch instead of 1.2.13 because the problem is also present there (in the release)? By: peter (peter) 2006-11-14 10:46:11.000-0600 We have had the problem to various degrees since version 1.2.7.  We have tried both the stable releases and the svn check outs.  Just the lastest version we have tried was an svn version. By: Joshua C. Colp (jcolp) 2006-11-15 17:04:54.000-0600 The backtrace isn't that useful :\ unfortunately. What is the box doing at the time? Is it using AGIs? can you attach the console output? can you attach another "bt" and a "bt full"? Thanks. By: peter (peter) 2006-11-15 18:00:59.000-0600 bt.txt This was a planned crash so we could get a core file with out killing any calls.  So the box was sitting with no calls.  No AGI's where running when this one happened. bt2.txt Was created about 10 minute after bt.txt.  Our agents where logging into the system, and I think there was a call in the system.  Then is when we where getting the calls routing through the correct remote asterisk system.  There might of been an AGI running here. Sorry, I couldn't find the console output. By: jmls (jmls) 2007-02-11 03:23:30.000-0600 any more news on this ? By: peter (peter) 2007-02-12 10:04:41.000-0600 No new updates. By: Joshua C. Colp (jcolp) 2007-02-15 19:49:16.000-0600 There is nothing odd in this second backtrace either. All looks routine. By: peter (peter) 2007-02-16 10:01:05.000-0600 All the backtraces look basicly the same.  I was hoping that a second set of eye would notice something odd.  And as far as I can remember none of the log file or the console show anything odd either.  It just dies. By: Serge Vecher (serge-v) 2007-03-07 13:03:59.000-0600 what happens with a clean installation from 1.2.16 tarball? By: peter (peter) 2007-03-07 13:37:07.000-0600 Now that Asterisk 1.4.1 is released we are in the process of upgrade the system to it. By: Serge Vecher (serge-v) 2007-03-26 12:51:26 if still an issue with 1.4.2, please feel free to reopen the issue with a backtrace attached.