Concurrent calls hitting the same Halef instance cause JVXML to crash. The second incoming call should get a busy signal (or placed on hold) by Asterisk. At any given point in time, a given Halef instance should only be responsible for one call. Also, in between calls, there should be a minimum of 15 seconds pause. Some of this logic is already implemented in default.conf.generator.py.
Hi Jochen,
David suggested that I turn over ownership of this ticket to you. I'd like to briefly list the changes I made on TJR1 (for more details please refer to the now closed ticket on "Full Call Recordings").
Everything in /etc/asterisk is now under version control (sync-ed to the Halef Asterisk sourceforge repo). This includes the Python script you wrote (modified by me) to automatically populate extensions.conf. However, note that the actual extension files (*.csv) are NOT under version control. So please make sure you commit your changes to the repo regularly (change the user name to your SourceForge username before committing, however, so that we know that it was your commit).
Asterisk now supports full call recordings (basically involved changing/adding 2 lines of code in the dialplans generated for each extension, by changing the Python script). A minor change was also made to features.conf.
Feel free to contact me if you have further questions related to these changes at vramanarayanan@ets.org.
Thanks!
Any progress on this, Jochen?
Assigning this to Vikram as he is currently working on the dial plan generator anyway