Summary:	ASTERISK-02814: Crash when terminating several calls at the same time
Reporter:	dsandras (dsandras)	Labels:
Date Opened:	2004-11-15 05:06:51.000-0600	Date Closed:	2011-06-07 14:05:06
Priority:	Major	Regression?	No
Status:	Closed/Complete	Components:	Core/General
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) bt.txt ( 1) core_result ( 2) core_result2 ( 3) Levent_patch_part_from_asterisk_patch_of_bristuff ( 4) Levent_patch_part_from_asterisk_patch_of_bristuff
Description:	We have a packet generator that will generate randomly several calls at the same time. The calls are started randomly and end after 30 to 45 seconds. If we stop them all at once, then we get a crash. ****** ADDITIONAL INFORMATION ****** bt full, in attachment I have not been able to fix this myself, any suggestion welcome.
Comments:	By: dsandras (dsandras) 2004-11-15 05:07:53.000-0600 Ah, no idea why it put this in Database Support, and no idea how to move it in another category. By: Mark Spencer (markster) 2004-11-15 08:51:58.000-0600 To confirm, the problem only occurs when you disconnect all the calls at once, is that correct? By: dsandras (dsandras) 2004-11-15 09:03:16.000-0600 That is absolutely correct. If calls are started/stopped randomly, then it doesn't occur. It only happens when all calls are stopped at once by the generator. By: Pavel Litvinenko (pavell) 2004-11-15 09:46:44.000-0600 I have the similar problem I have ~20-50 online iax "friends" registered on the * and almost all use iaxcomm By: Mark Spencer (markster) 2004-11-15 10:57:07.000-0600 Can you rebuild with "make clean ; make valgrind" ? By: litw (litw) 2004-11-15 11:46:49.000-0600 should I uncomment  #-DMAKE_VALGRIND_HAPPY ? By: Mark Spencer (markster) 2004-11-15 13:08:16.000-0600 You can, but it's likely not necessary anymore. By: litw (litw) 2004-11-15 13:13:36.000-0600 forgot to say I run asterisk version 1.0.2 By: triona (triona) 2004-11-15 18:05:50.000-0600 category changed to general. By: Pavel Litvinenko (pavell) 2004-11-16 08:04:45.000-0600 core_result2 from * installed by make valgrind By: dsandras (dsandras) 2004-11-16 11:03:53.000-0600 I apologize, but with fewer calls, I can not reproduce the problem and with valgrind, it is slowed down, so I can't reach so many calls. By: Brian West (bkw918) 2004-11-18 15:20:00.000-0600 I would be very intrested to know what CDR logging you might be doing? mysql, odbc or csv? bkw By: dsandras (dsandras) 2004-11-18 15:41:25.000-0600 MySQL, but IAX is the only channel where that behavior is reproducible. So that doesn't seem to be CDR-related. By: Mark Spencer (markster) 2004-11-18 16:07:38.000-0600 This isn't latest CVS head (at least the line numbers don't line up to what I see here).  Either update to latest CVS head and give me a new backtrace or find me on IRC so I can login and look at this problem on your system with your existing core/sources. By: Mark Spencer (markster) 2004-11-18 16:08:27.000-0600 Is this code patched in any way? By: Pavel Litvinenko (pavell) 2004-11-18 23:03:40.000-0600 You are right, kapejod's bristuff touched chan_iax2.c - see uploaded file ... By: Mark Spencer (markster) 2004-11-19 00:02:48.000-0600 What a surprise, it's crashing on the line that is added by the patch! SCREAM!!!!!!!! Ditch the patch and see if the problem goes away. By: dsandras (dsandras) 2004-11-19 06:05:43.000-0600 Pfiuu, you are absolutely right again! If I ditch the IAX part of the patch, Asterisk becomes very stable again. I guess the bug can be closed, I will mail kapejod :D Thanks for your time, next time, I will test pure Asterisk before reporting! By: triona (triona) 2004-11-19 07:27:47.000-0600 Reviewed bugnotes. Bug doesn't occur on unpatched systems. Not an Asterisk bug. By: triona (triona) 2004-11-19 07:29:18.000-0600 Only occurs with third party patch.