Summary:	ASTERISK-02587: [request] "no call progress" indication timeout option.
Reporter:	mustdie (mustdie)	Labels:
Date Opened:	2004-10-12 09:59:01	Date Closed:	2011-06-07 14:10:01
Priority:	Minor	Regression?	No
Status:	Closed/Complete	Components:	Applications/app_dial
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:
Description:	Here is the deal. One of the upstream providers (SIP) is having some problems with call completion. And as side effect, they are not passing call progress back to me (When their system is down.) I need a way to detect that and switch to different provider if necessary. So far i can't acomplish this with current app_Dial. Our timeout is too general and applies on the entire call. I need some new "NO CALL PROGRESS" timeout so I can reliably switch to different service provider.
Comments:	By: Mark Spencer (markster) 2004-10-12 10:05:04 Just turn on qualify and it should accomplish the same thing on SIP, you can even set a lower qualify time. By: mustdie (mustdie) 2004-10-12 10:11:48 SIP Peer is never goes down, as matter of fact it is set as qualify=yes, and it's always alive. This problem resides on Telco side with one of the CLEC's. My problem, that I need a timeout in case if no call progress is found. By: Mark Spencer (markster) 2004-10-12 10:14:11 What does the sip debug look like for a failed call? By: mustdie (mustdie) 2004-10-12 10:18:08 Invite, then 100 trying... and then it sits there... with all appropriate acknowledgments. It is nothing wrong with SIP. By: mustdie (mustdie) 2004-10-12 10:24:27 Clarification: My asterisk Box (A) --> Provider's asterisk box (B) --> Box (c) Cisco gateway SIP. Communication between A <--> B, B <--> C is very stable. It just that Cisco gateway does not return Ringing indication, when their system is not operational. Given that: Box A, Box B, Box C is always alive and operational. All sip captures are appropriate. By: mustdie (mustdie) 2004-10-12 11:48:26 Option W([x]) would be more logical, where x would be set for default 15 seconds. IMHO, it has nothing to do with chan_sip. This option should be channel independent. By: Mark Spencer (markster) 2004-10-12 12:09:01 I would rather see us adjust the qualify code in SIP to do this, as this seems like a really ugly workaround. By: mustdie (mustdie) 2004-10-12 12:19:06 Mark, SIP peer is never going into "unreachable" state. It's much better approach to modify app_dial. Let's say I have PRI with A-Z provider, and certian destinations has very slow call setup time. I also saw situation, where "all curciuts are busy now" message plays for about 40 seconds, with no ringing indications (PRI). Then in about 40 seconds i'm getting busy message. If we would have this option W([x]), we would be able to switch to backup provider right away. And it has nothing to do with chan_sip. By: twisted (twisted) 2004-10-27 16:57:13 Where do we stand on this?   --Housekeeping By: mustdie (mustdie) 2004-10-27 20:06:54 I'm almost done writing this patch by myself (Will post it soon). By: twisted (twisted) 2004-11-14 21:29:54.000-0600 *ding* Where do we stand? --housekeeping By: Mark Spencer (markster) 2004-11-22 16:16:36.000-0600 Doesn't look like this is going anywhere, feel free to reopen when you have a patch, mustdie.