Summary:	ASTERISK-05172: Intermitant delays on call setup.
Reporter:	delvar (delvar)	Labels:
Date Opened:	2005-09-28 04:55:52	Date Closed:	2011-06-07 14:10:16
Priority:	Minor	Regression?	No
Status:	Closed/Complete	Components:	Core/General
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) delay-trace.cap
Description:	We are seeing this weird problem, it seems to happen at random periods throughout the day from a few minuets to a up to an hour. the problem is seen as a gap of a second or two after answerign the phone whe reht eaudio is lost (not a delay) ****** ADDITIONAL INFORMATION ****** [Phone A] >--SIP--> [Asterisk] >--SIP--> [Phone B] Both phones are snom 360's. Asterisk is Stable 1.0.9 Pretty simple config, just a dial direct to each other like Dial(SIP/phoneA,30,t) Running Gentoo linux When we make a call during one of the problem periods, from [Phone A] to [Phone B] there is up to 2 seconds delay before A hears B. resulting in the first word or two being lost, ie [Phone B] would pickup and say &ASTERISK-7995;Hello technical support&ASTERISK-7996; and [Phone A] would hear &ASTERISK-7995;nical support&ASTERISK-7996;. Looking at a packet trace we see the SIP invites coming in and the calls being setup ok, but on the call to [Phone B] we see the RTP from [Phone B] to [Asterisk] and [Phone A] to [Asterisk] but nothing being sent from * to either of the phones, then after about a second we see a whole bunch of RTP packets being sent out from [Asterisk], all in one go to both phones. It does this 'bunching' a couple of times then settles down to normal. Id say it was some sort of timing problem or load problem, but during these times conferencing etc works ok and there is no appreciable load on the server or network. We see no problems or latency on the network either and we get the same problem with two phones and the server plugged into the same switch so they are no more than 3 feet away from each other. atached is a packet cap of both SIP and RTP. Any ideas? Thanks.
Comments:	By: Michael Jerris (mikej) 2005-09-28 08:26:36 per bug guidelines, please attach a sip debug to this bug report, also, are you able to test this on current cvs head? By: delvar (delvar) 2005-09-28 08:43:55 attached is a full packet trace including SIP and RTP, is this good enough? or do you want me to copy out all the sip packets seperatly? i have not tested with CVS head as this is a live system and switching and changeing versions is not a good thing. the trace shows both sip and rtp, the sip looks fine, the calls get setup with no problems and as i said above the problem comes with teh RTP trafic. By: Michael Jerris (mikej) 2005-09-28 08:49:44 I'm looking for the debug and verbose output from asterisk itself.  You can capture this to a file using the settings in logger.conf.  For this problem we would need verbose, sip debug and rtp debug.  Thanks By: delvar (delvar) 2005-09-28 09:30:28 hi, getting a copy of the cli at the same time as a packet trace is ging to be dificault as the problem is not reproducable, we ahve to wait for someone to report an issue then log into the server and make a few calls. it took a while for us to get this trace. i might be able to get the log files from that period if that will help. but at teh CLI we see no errors or anything that looks unusual, hence why we resorted to getting a packet capture. looking a the trace, have you seen this sort of thing happen before? what sort of things can cause it? By: dawson (dawson) 2005-09-30 11:25:38 Just a note to say that we are experiencing the same thing that is basically a bunching up of network traffic in respect to both RTP and SIP and is intermittent. Will try and get a trace. Our version of Asterisk is around 1.0.2. By: Olle Johansson (oej) 2005-10-11 23:28:20 We need a full SIP debug and configs to be able to handle this issue, since we can not reproduce it.