[asterisk-users] zombie SIP channels after CURL cnam lookup

[Damon Estep]

Can anyone suggest a reason why these channels might end up zombies?

 

The process is;

 

Call comes in via SIP into a context that appends the caller ID name as
follows;

 

[cnam-lookup]

exten =>
_[2-9]XXXXXXXXX,1,set(CALLERID(name)=${CURL(http://cnam.provider.com/?co
mpanyId=555&password=password&number=${CALLERID(num)})})

exten => _[2-9]XXXXXXXXX,2,goto(subscriber-numbers|${EXTEN}|1)

 

the call is then sent to the context where the extension is defined.
This works well with high volume, but there are occasionally zombie
channels as a result, can not track down the cause;

 

 

Channel              Location             State   Application(Data)

SIP/1.1.251.9-b6700 5555939729 at cnam-look Ring    (None)

SIP/1.1.251.9-b6a0d 5556351876 at cnam-look Ring    (None)

SIP/1.1.251.9-b6ad0 5555874510 at cnam-look Ring    (None)

SIP/1.1.251.9-b7dcf 5559620060 at cnam-look Ring    (None)

SIP/1.1.251.9-b675f 5556690523 at cnam-look Ring    (None)

 

The channels listed above have appeared in show channels for 2 days now.

 

I assume it was either because the CURL response was not returned, since
we are still in he context cnam-lookup and the next step is a goto.

 

Is there a way to set an absolute timeout for the "set" command and
continue in the dialplan if that timeout is exceeded, without impacting
timeouts further down the line?

 

The cnam response should come within 200ms.

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20061130/6ab752d8/attachment.htm