Summary:	ASTERISK-10482: E&M Wink stops responding after originating about 500 calls
Reporter:	Whit Thiele (sirwhit)	Labels:
Date Opened:	2007-10-09 12:35:42	Date Closed:	2011-06-07 14:07:24
Priority:	Minor	Regression?	No
Status:	Closed/Complete	Components:	Channels/chan_zap
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:
Description:	Able to make about 500 outgoing calls via Manager Originate. Using ZapBarge, I can hear the calls being launched and DTMF digits dialing on the trunk. At approximately 500 calls, I'll hear a telco message saying that not enough digits have been dialed. Next and all future calls goes into a fast busy and the e&m trunks are useless until asterisk is restarted. Setup: 2 x E&M Wink trunks on a TE410p Digium Card Using Manager API to Originate calls to E&M Wink trunks. zaptel.conf ----------- span=1,0,0,d4,ami span=2,1,0,d4,ami e&m=1-48 loadzone=us defaultzone=us zapata.conf ----------- group = 2 signalling= em_w switchtype = national immediate=no overlapdial=yes busydetect=no callprogress=no rxgain=     0.0 txgain=     0.0 usecallerid => no context = outgoing channel =>1-23 Telco not seeing any issue on their end. Digium support logged in and saw nothing wrong with the card. No errors are reported from asterisk either in CLI or /var/log/asterisk/messages ****** ADDITIONAL INFORMATION ****** Once the E&M Wink trunks starts exhibiting problems, dialing directly through the dialplan fails as well as the Manager API Originate. The behavior is consistent everytime and have never been able to launch more then 550 calls before it fails. Thought it could be related to 9660 but Corydon76 suggested I open this as a new issue.
Comments:	By: Whit Thiele (sirwhit) 2007-10-16 13:54:25 OK, I've reverted back to Asterisk 1.4.9 and everything works fine. There must be something going on with chan_zap which is causing this problem. Please let me know what kind of debug information I can provide to help track this down. By: Joshua C. Colp (jcolp) 2008-02-14 12:17:15.000-0600 After examining the code the remembering previous conversations/issues I believe the fault that caused this has been corrected and the issue should be fixed.