Summary:ASTERISK-06264: SIP Hold on Avaya 4610 kills iax leg of call
Reporter:caspy (caspy)Labels:
Date Opened:2006-02-07 04:14:56.000-0600Date Closed:2006-03-02 15:52:40.000-0600
Versions:Frequency of
Environment:Attachments:( 0) sip-iax-hold.1.log
( 1) sip-iax-hold.log
Description:Hold/Unhold procedure on SIP Avaya phone kills audio in direction to IAX2 client. After hang-up of SIP side, IAX2 side stays on-line (untill timeout comes) doing nothing.

This issue is not always reproduceable from the first attempt,
but after several tryes - always.
Things such debugging on, makes chances lower.
Disabling MOH also make chances of bug apperance lower.
But after no more that 8-10 attemts, this bug happens.

Other combinations: sip to sip, iax to iax,
AND sip to iax with other sip client (tested on ata-186, sjphone) -- work ok.


Call is made between SIP exten: Avaya 4610SW, SIP R2.2,
and IAX2 exten: idefisk, 1.32.

To reproduce bug we need: press/unpress Hold, transfer a call, etc...
Thus, anything, that puts iax side on hold.

OS: Debian, kernel, SMP
Comments:By: twisted (twisted) 2006-02-08 16:37:50.000-0600

can you provide a sip trace of the call?

By: caspy (caspy) 2006-02-09 02:14:45.000-0600

Here it is: sip-iax-hold.log

Asterisk was started as 'asterisk -vvvvvvcgdddddd',
'sip debug', 'iax2 debug' was entered at console.

Logging was made in such way:
$ grep ^debug /etc/asterisk/logger.conf
debug => debug

sip leg - 1006, iax leg - 2006.

Hold/Unhold was made twice. After first attempt all was ok, and after second - sound from sip to iax disappeared. After hanging up sip side, iax released line after 30-40 sec (it is logged too).

By: twisted (twisted) 2006-02-09 16:10:22.000-0600

I meant more along the lines of simply the output of the 'sip debug' command for that call, as the debug in this situation is kinda difficult to read.  

i want to compare the invites for hold from the avaya to that of another phone and see if a difference can be isolated that could be causing the problem.

Can you provide this filtered debug output?


By: caspy (caspy) 2006-02-10 03:18:28.000-0600

the same, but without iax debugging info

By: twisted (twisted) 2006-02-10 09:25:12.000-0600

this still reformats the sip messages in an odd way, but it's a little cleaner.  

What kind of IAX device are you using?  

Is this reproducable on sip to sip calls, or sip to zap calls?

By: caspy (caspy) 2006-02-10 09:38:40.000-0600

iax device - is a PC laptop with WinXP and Idefisk 1.32.
Two previous versions of Idefisk and last version of DIAX were also tried, - the same thing.

This bug is not reproducable on sip-to-sip or sip-to-zap, even with this Avaya device.

By: Mark Spencer (markster) 2006-02-11 10:39:39.000-0600

Does your server continue to accept IAX calls after this takes place?  Is there any reason you suspect this is an asterisk bug vs. a bug in your IAX client?

By: caspy (caspy) 2006-02-14 06:39:01.000-0600

We took an another asterisk, register it to the same extention, as idefisk was.
In such configuration no sound loss or other troubles was mentioned.
So, may be, you are quite right.

If we will found any more interesting facts - i'll post details.

By: Olle Johansson (oej) 2006-03-02 15:52:26.000-0600

Client issue. Please re-open or find a bug marshal if you find out any more information about a bug on the Asterisk side.