asterisk users - Jul 2004 - Asterisk crashing with no indication why.

TC wrote:
>>I'm hoping someone might have seen this before because I'm just
about
>>at a loss of what to do.  I have an asterisk system setup in a call
>>center environment with multiple queues. After a random uptime asterisk
>>will suddenly come to a partial halt where I can connect to the cli but
>>issuing a command such as show channels gives no response, and calls
>>cannot be made in or out. Calls in progress usually drop as well, but
>>if they don't right away, after a minute or so they will. To remedy
the
>>problem I have to do a restart on asterisk, which of course makes all
>>the agents have to login again and is just a big mess.
>>
>>I have agents being dynamically added to the queues via an AGI script,
>>also the agents are added to all queues so that they can take calls
>>from any of them. I'm not sure if this is important but since I use
the
>>AgentCallbackLogin function I have all the agents inside their own
>>context so that I can use a macro to determine if they are on an
>>outgoing call (using app_checkgroup) before ringing them to prevent
>>call waiting tones.
>>
>>I've thoroughly searched the messages log, in which I have both
verbose
>>and debug logging enabled. I've never found anything to indicate a
>>problem, it simply looks like calls just slow down and stop. One other
>>thing that may be important, I have a daemon running which stays
>>connected to the manager api listening for events and sending off two
>>commands every 10 seconds, one to get the status of the queues, and one
>>to get the status of agents.
> 
> say hello to app_noQ and chan_deadlock :)
> http://www.voip-info.org/tiki-index.php?page=Asterisk%20debugging
> 
> I might be wrong about chan_deadlock but i am pretty sure yiou have
> dead lock situation might be cause by manager blocking all else...
> try the dead lock debug to see if tou can id when threads are competing for
> the same
> locks
> 

I'm with TC.  I'm sure you have a deadlock.  Same thing happened to me 
with IAX and fax detection.  I turned off fax detection and my problems 
were solved.

TL

I'm hoping someone might have seen this before because I'm just about 
at a loss of what to do.  I have an asterisk system setup in a call 
center environment with multiple queues. After a random uptime asterisk 
will suddenly come to a partial halt where I can connect to the cli but 
issuing a command such as show channels gives no response, and calls 
cannot be made in or out. Calls in progress usually drop as well, but 
if they don't right away, after a minute or so they will. To remedy the 
problem I have to do a restart on asterisk, which of course makes all 
the agents have to login again and is just a big mess.

I have agents being dynamically added to the queues via an AGI script, 
also the agents are added to all queues so that they can take calls 
from any of them. I'm not sure if this is important but since I use the 
AgentCallbackLogin function I have all the agents inside their own 
context so that I can use a macro to determine if they are on an 
outgoing call (using app_checkgroup) before ringing them to prevent 
call waiting tones.

I've thoroughly searched the messages log, in which I have both verbose 
and debug logging enabled. I've never found anything to indicate a 
problem, it simply looks like calls just slow down and stop. One other 
thing that may be important, I have a daemon running which stays 
connected to the manager api listening for events and sending off two 
commands every 10 seconds, one to get the status of the queues, and one 
to get the status of agents. My cvs version is 
CVS-HEAD-06/24/04-06:49:37. I've looked through all the latest cvs 
updates and bug reports and don't see anything that would be related. 
Has anyone seen this before, can anyone suggest anything I might try? 
With both being unable to reproduce this at will and the lack of 
messages or log entries pointing to the problem I'm pretty much up 
against a wall.

Thank you for any help anyone can offer,

--Daniel Daley--
dan@chameleoncode.jp

> I'm hoping someone might have seen this before because I'm just
about
> at a loss of what to do.  I have an asterisk system setup in a call
> center environment with multiple queues. After a random uptime asterisk
> will suddenly come to a partial halt where I can connect to the cli but
> issuing a command such as show channels gives no response, and calls
> cannot be made in or out. Calls in progress usually drop as well, but
> if they don't right away, after a minute or so they will. To remedy the
> problem I have to do a restart on asterisk, which of course makes all
> the agents have to login again and is just a big mess.
>
> I have agents being dynamically added to the queues via an AGI script,
> also the agents are added to all queues so that they can take calls
> from any of them. I'm not sure if this is important but since I use the
> AgentCallbackLogin function I have all the agents inside their own
> context so that I can use a macro to determine if they are on an
> outgoing call (using app_checkgroup) before ringing them to prevent
> call waiting tones.
>
> I've thoroughly searched the messages log, in which I have both verbose
> and debug logging enabled. I've never found anything to indicate a
> problem, it simply looks like calls just slow down and stop. One other
> thing that may be important, I have a daemon running which stays
> connected to the manager api listening for events and sending off two
> commands every 10 seconds, one to get the status of the queues, and one
> to get the status of agents.
say hello to app_noQ and chan_deadlock :)
http://www.voip-info.org/tiki-index.php?page=Asterisk%20debugging

I might be wrong about chan_deadlock but i am pretty sure yiou have
dead lock situation might be cause by manager blocking all else...
try the dead lock debug to see if tou can id when threads are competing for
the same
locks

Update .. do make update

bkw
> -----Original Message-----
> From: asterisk-users-admin@lists.digium.com [mailto:asterisk-users-
> admin@lists.digium.com] On Behalf Of Daniel Daley
> Sent: Monday, July 12, 2004 2:06 PM
> To: asterisk-dev@lists.digium.com; asterisk-users@lists.digium.com
> Subject: [Asterisk-Users] Asterisk crashing with no indication why.
>
> I'm hoping someone might have seen this before because I'm just
about
> at a loss of what to do.  I have an asterisk system setup in a call
> center environment with multiple queues. After a random uptime asterisk
> will suddenly come to a partial halt where I can connect to the cli but
> issuing a command such as show channels gives no response, and calls
> cannot be made in or out. Calls in progress usually drop as well, but
> if they don't right away, after a minute or so they will. To remedy the
> problem I have to do a restart on asterisk, which of course makes all
> the agents have to login again and is just a big mess.
>
> I have agents being dynamically added to the queues via an AGI script,
> also the agents are added to all queues so that they can take calls
> from any of them. I'm not sure if this is important but since I use the
> AgentCallbackLogin function I have all the agents inside their own
> context so that I can use a macro to determine if they are on an
> outgoing call (using app_checkgroup) before ringing them to prevent
> call waiting tones.
>
> I've thoroughly searched the messages log, in which I have both verbose
> and debug logging enabled. I've never found anything to indicate a
> problem, it simply looks like calls just slow down and stop. One other
> thing that may be important, I have a daemon running which stays
> connected to the manager api listening for events and sending off two
> commands every 10 seconds, one to get the status of the queues, and one
> to get the status of agents. My cvs version is
> CVS-HEAD-06/24/04-06:49:37. I've looked through all the latest cvs
> updates and bug reports and don't see anything that would be related.
> Has anyone seen this before, can anyone suggest anything I might try?
> With both being unable to reproduce this at will and the lack of
> messages or log entries pointing to the problem I'm pretty much up
> against a wall.
>
> Thank you for any help anyone can offer,
>
> --Daniel Daley--
> dan@chameleoncode.jp
>
> _______________________________________________
> Asterisk-Users mailing list
> Asterisk-Users@lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-users
> To UNSUBSCRIBE or update options visit:
>    http://lists.digium.com/mailman/listinfo/asterisk-users

How many calls at one time? Specs of server, esp CPU, RAM and disk? SIP
calls or analog? Channel bank?

 

-----Original Message-----
From: asterisk-users-admin@lists.digium.com
[mailto:asterisk-users-admin@lists.digium.com] On Behalf Of Daniel Daley
Sent: Monday, July 12, 2004 12:06 PM
To: asterisk-dev@lists.digium.com; asterisk-users@lists.digium.com
Subject: [Asterisk-Users] Asterisk crashing with no indication why.

I'm hoping someone might have seen this before because I'm just about at
a
loss of what to do.  I have an asterisk system setup in a call center
environment with multiple queues. After a random uptime asterisk will
suddenly come to a partial halt where I can connect to the cli but issuing a
command such as show channels gives no response, and calls cannot be made in
or out. Calls in progress usually drop as well, but if they don't right
away, after a minute or so they will. To remedy the problem I have to do a
restart on asterisk, which of course makes all the agents have to login
again and is just a big mess.

I have agents being dynamically added to the queues via an AGI script, also
the agents are added to all queues so that they can take calls from any of
them. I'm not sure if this is important but since I use the
AgentCallbackLogin function I have all the agents inside their own context
so that I can use a macro to determine if they are on an outgoing call
(using app_checkgroup) before ringing them to prevent call waiting tones.

I've thoroughly searched the messages log, in which I have both verbose and
debug logging enabled. I've never found anything to indicate a problem, it
simply looks like calls just slow down and stop. One other thing that may be
important, I have a daemon running which stays connected to the manager api
listening for events and sending off two commands every 10 seconds, one to
get the status of the queues, and one to get the status of agents. My cvs
version is CVS-HEAD-06/24/04-06:49:37. I've looked through all the latest
cvs updates and bug reports and don't see anything that would be related. 
Has anyone seen this before, can anyone suggest anything I might try? 
With both being unable to reproduce this at will and the lack of messages or
log entries pointing to the problem I'm pretty much up against a wall.

Thank you for any help anyone can offer,

--Daniel Daley--
dan@chameleoncode.jp

_______________________________________________
Asterisk-Users mailing list
Asterisk-Users@lists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-users
To UNSUBSCRIBE or update options visit:
   http://lists.digium.com/mailman/listinfo/asterisk-users

Daniel Daley wrote:
> I'm hoping someone might have seen this before because I'm just
about at
> a loss of what to do.  I have an asterisk system setup in a call center 
> environment with multiple queues. After a random uptime asterisk will 
> suddenly come to a partial halt where I can connect to the cli but 
> issuing a command such as show channels gives no response, and calls 
> cannot be made in or out. Calls in progress usually drop as well, but if 
> they don't right away, after a minute or so they will. To remedy the 
> problem I have to do a restart on asterisk, which of course makes all 
> the agents have to login again and is just a big mess.
> 
> I have agents being dynamically added to the queues via an AGI script, 
> also the agents are added to all queues so that they can take calls from 
> any of them. I'm not sure if this is important but since I use the 
> AgentCallbackLogin function I have all the agents inside their own 
> context so that I can use a macro to determine if they are on an 
> outgoing call (using app_checkgroup) before ringing them to prevent call 
> waiting tones.
> 
> I've thoroughly searched the messages log, in which I have both verbose
> and debug logging enabled. I've never found anything to indicate a 
> problem, it simply looks like calls just slow down and stop. One other 
> thing that may be important, I have a daemon running which stays 
> connected to the manager api listening for events and sending off two 
> commands every 10 seconds, one to get the status of the queues, and one 
> to get the status of agents. My cvs version is 
> CVS-HEAD-06/24/04-06:49:37. I've looked through all the latest cvs 
> updates and bug reports and don't see anything that would be related. 
> Has anyone seen this before, can anyone suggest anything I might try? 
> With both being unable to reproduce this at will and the lack of 
> messages or log entries pointing to the problem I'm pretty much up 
> against a wall.
> 
> Thank you for any help anyone can offer,
> 
> --Daniel Daley--
> dan@chameleoncode.jp
> 
Depending on your dev skills, you could run asterisk in gdb and then 
look at the status of each thread when the problem occurs. Other than 
that, try an older version of asterisk

PS Please don't post in both lists, it isn't a dev question

-Adam

Hi!
> problem, it simply looks like calls just slow down and stop. One other 
> thing that may be important, I have a daemon running which stays 
> connected to the manager api listening for events and sending off two 
> commands every 10 seconds, one to get the status of the queues, and one 
> to get the status of agents.
That can very well be the cause - there is a known issue, first stated by 
"mattf" on this list, about trouble when a manager API client looses
the
connection and then another API client (maybe the same using a different 
connection) comes into play.

So: Stop using that deamon and see if Asterisk becomes stable. 
You could also introduce regular - but not too many! - "reloads", e.g.
1-
4 times a day plus one "restart" during night time with the help of a 
CRON job.

Also: Check /tmp for a core file and run gdb and a backtrace if you find 
one. Search for "debugging" on the Wiki for more detailed
instructions.

Cheers, Philipp