[Asterisk-Users] Polling Asterisk for Life

[Mike Clark]

Rich Adamson wrote:

>  
>
>>>The problem isn't that asterisk isn't running, it's that asterisk is
>>>not responding.  When asterisk is in this funky state, I can still run
>>>"asterisk -vvvvr" from the command line and get access to the CLI. 
>>>While in the CLI, the only command that asterisk will respond to is
>>>"exit" which drops me back to the shell.  If I try to issue a "stop
>>>now", asterisk just immediately returns to the CLI> prompt.  It does
>>>this for every single command, except for "exit".
>>>
>>>So, simply respawning asterisk, or checking to see if it's running
>>>isn't good enough, because asterisk is indeed running.  We need to
>>>access asterisk and issue a command, and see if asterisk responds
>>>appropriately.  If not, we can assume it has died, and we can kill it
>>>off (killall -9 asterisk) and then start it back up again (or reboot
>>>the whole server if necessary).
>>>
>>>Yes, it's an odd problem, but I've noticed it so I can confirm it is a
>>>state asterisk can get into, and can confirm its symptoms.  Hopefully
>>>all that is over with now after I upgraded (also fyi, I also moved my
>>>x101p around so it'd get its own irq, so it's possible that was the
>>>problem, though I doubt it).  If it turns out I still have the
>>>problem, I'll probably whip up a script to check asterisk's condition
>>>and restart if needed.
>>>
>>>Joseph Tanner
>>>
>>> 
>>>
>>>      
>>>
>>Do you happen to have TDM400 cards in your system? I have learned to set 
>>up any machine with TDM400 cards to do a nightly auto-reboot. If we 
>>don't, they will eventually exhibit behaviour identical to what  you 
>>describe above.
>>    
>>
>
>Better RMA your tdm400 card then. Have multiple boxes with these cards
>that function for many weeks, and their uptime is only limited due to
>changes that are made about once per month or so.
>
>  
>
Well, we have 15+ systems with 2 TDM04B cards each that have pretty much 
all have exhibited this behaviour. Some of them would go for weeks but 
all eventually succumbed. We do the auto-reboot on those now  and 
switched to Sangoma A200D for all new systems.