Summary:	ASTERISK-07501: [patch][post 1.4] res_segfault - handle segfault with a command of choice
Reporter:	Anthony Minessale (anthm)	Labels:
Date Opened:	2006-08-09 10:55:38	Date Closed:	2006-08-13 20:14:18
Priority:	Major	Regression?	No
Status:	Closed/Complete	Components:	Core/General
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) res_segfault.c ( 1) segfault.conf.sample
Description:	This module allows the user to run a command when Asterisk crashes making it possible to alert a sysop when a production system fails. ****** ADDITIONAL INFORMATION ****** The sample config demonstrates the simple interface which is more or less: [general] exec => <shell command of choice>
Comments:	By: Serge Vecher (serge-v) 2006-08-09 11:04:06 cool. Should this be integrated somehow with the -g option? By: Russell Bryant (russell) 2006-08-09 15:20:58 It seems kind of silly to put this into its own module and a new configuration file.  I think the functionality is cool, but I vote for placing the code in asterisk.c and the option in asterisk.conf ... By: damin (damin) 2006-08-09 18:13:47 I just pissed myself laughing. By: Clod Patry (junky) 2006-08-09 20:46:22 And what about leaving this as a res and the config in asterisk.conf, maybe in a class segfault? I personnaly like it as a module, since if you dont want it, it's not in your legs and if you want it, just load it. By: Roy Sigurd Karlsbakk (rkarlsba) 2006-08-11 06:59:21 We should keep the text in segfault.conf.samlpe, though. Think about all the nice messages being sent to digium support...... or perhaps not :P Anyway, having this generate a backtrace would be fine, although it somehow seems hard...... Somehow I still find using safe_asterisk is cleaner than this, considering the what signal(2) manual is saying "According to POSIX, the behaviour of a process is undefined after it ignores a SIGFPE, SIGILL, or SIGSEGV signal that was not generated by the kill(2) or  the  raise(3) functions." roy By: Russell Bryant (russell) 2006-08-13 20:14:18 Given that the behavior of a process is undefined after a SIGSEGV, I don't think this is not something we can merge.  However, the ability to email an administrator on a crash is provided using the safe_asterisk script.  I also added an EXEC option in safe_asterisk to allow the execution of whatever command you want if Asterisk crashes.