Michael Maier
2016-Apr-25 02:26 UTC
[asterisk-users] Second invite after 100ms (with default t1min=100) --> canceled call problem!
Hello! I encounter the following problem (asterisk 11 and 13) with Teconisy as trunk provider with enabled qualify and default t1min (100ms): Teconisy most often doesn't answer the first invite before asterisk default t1min ended. Therefore asterisk sends one more invite. This second invite is answered by Teconisy with status 486 - Request terminated - Channel limit exceeded. (The second invite obviously is interpreted as second, new call). Asterisk therefore cancels the first(!!) call - but Teconisy proceeds with exactly this first call (which now can't be handled by asterisk any more). For me, there are two problems in asterisk at this point: 1. The VoIP standard defines 500ms for t1 - using this standard value for t1min works fine with Teconisy, too. t1min should be always 500ms - it prevents a completely blocked line! 2. Why does asterisk stop the call completely after the second invite, which is canceled by Teconisy? It should be ignored because it means, that the first invite is already processed by Teconisy. Btw: I can see the same behavior with German Telekom. Asterisk most of the time sends two invites, too, but here I don't get an error back from Telekom, because I have two channels and therefore the channel limit isn't exceeded most of the time. Setting t1min to the VoIP standard of 500ms prevents these unnecessary double invites with German Telekom, too. Therefore: Please use the standards compliant t1min=500 instead of the not compliant value of 100, which can cause much trouble and which creates totally unnecessary network overhead. Or is there another solution I overlooked? Thanks, Michael Maier
Joshua Colp
2016-Apr-25 10:35 UTC
[asterisk-users] Second invite after 100ms (with default t1min=100) --> canceled call problem!
Michael Maier wrote:> Hello! > > I encounter the following problem (asterisk 11 and 13) with Teconisy as > trunk provider with enabled qualify and default t1min (100ms): > > Teconisy most often doesn't answer the first invite before asterisk > default t1min ended. Therefore asterisk sends one more invite. This > second invite is answered by Teconisy with > > status 486 - Request terminated - Channel limit exceeded. > > (The second invite obviously is interpreted as second, new call). > > Asterisk therefore cancels the first(!!) call - but Teconisy proceeds > with exactly this first call (which now can't be handled by asterisk any > more). > > For me, there are two problems in asterisk at this point: > > 1. The VoIP standard defines 500ms for t1 - using this standard value > for t1min works fine with Teconisy, too. t1min should be always > 500ms - it prevents a completely blocked line!The standard actually allows you to ignore this and use the estimated round trip time, which chan_sip will derive from the time it takes an OPTIONS packet to go out and get a response. The t1min just enforces a minimum.> 2. Why does asterisk stop the call completely after the second invite, > which is canceled by Teconisy? It should be ignored because it > means, that the first invite is already processed by Teconisy.The 486 response code is actually for indicating busy. The 4XX series is also for client failure, which tear down the session. It can't be ignored. This would break things. The retransmission of an INVITE shouldn't result in multiple sessions being set up, the other side should treat it as a retransmission. This is a bug on their side. Your adjustment of the T1 just makes it so you don't allow it to happen. I'd say the problem isn't with Asterisk but with the remote side. Since you can configure things to work that's great but I don't see any code changes we can do. Cheers, -- Joshua Colp Digium, Inc. | Senior Software Developer 445 Jan Davis Drive NW - Huntsville, AL 35806 - US Check us out at: www.digium.com & www.asterisk.org
Michael Maier
2016-Apr-25 15:43 UTC
[asterisk-users] Second invite after 100ms (with default t1min=100) --> canceled call problem!
Hello Joshua, On 04/25/2016 at 12:35 PM, Joshua Colp wrote:> Michael Maier wrote: >> Hello! >> >> I encounter the following problem (asterisk 11 and 13) with Teconisy as >> trunk provider with enabled qualify and default t1min (100ms): >> >> Teconisy most often doesn't answer the first invite before asterisk >> default t1min ended. Therefore asterisk sends one more invite. This >> second invite is answered by Teconisy with >> >> status 486 - Request terminated - Channel limit exceeded. >> >> (The second invite obviously is interpreted as second, new call). >> >> Asterisk therefore cancels the first(!!) call - but Teconisy proceeds >> with exactly this first call (which now can't be handled by asterisk any >> more). >> >> For me, there are two problems in asterisk at this point: >> >> 1. The VoIP standard defines 500ms for t1 - using this standard value >> for t1min works fine with Teconisy, too. t1min should be always >> 500ms - it prevents a completely blocked line! > > The standard actually allows you to ignore this and use the estimated > round trip time, which chan_sip will derive from the time it takes an > OPTIONS packet to go out and get a response. The t1min just enforces a > minimum. > >> 2. Why does asterisk stop the call completely after the second invite, >> which is canceled by Teconisy? It should be ignored because it >> means, that the first invite is already processed by Teconisy. > > The 486 response code is actually for indicating busy. The 4XX series is > also for client failure, which tear down the session. It can't be > ignored. This would break things. > > The retransmission of an INVITE shouldn't result in multiple sessions > being set up, the other side should treat it as a retransmission. This > is a bug on their side. Your adjustment of the T1 just makes it so you > don't allow it to happen. > > I'd say the problem isn't with Asterisk but with the remote side. Since > you can configure things to work that's great but I don't see any code > changes we can do.thanks for clarification! Besides that - I'm really wondering why others don't face this problem - I hardly can believe that I'm the only one affected. Regards, Michael