|Summary:||ASTERISK-18197: Deadlock - Periodic Deadlock on Transferring Calls|
|Reporter:||Josh Urbanski (almagsupport)||Labels:|
|Date Opened:||2011-07-27 09:56:08||Date Closed:||2011-08-17 14:18:44|
|Environment:||Linux version 2.6.32-5-686 (Debian 2.6.32-34squeeze1) (email@example.com) (gcc version 4.3.5 (Debian 4.3.5-4) ) #1 SMP Wed May 18 07:08:50 UTC 2011||Attachments:||( 0) backtrace-threads.txt|
( 1) core-show-locks.txt
|Description:||We are currently experiencing periodic deadlocks on our phone switch when calls are being transferred. When we encounter the deadlock we need to kill and restart the process of Asterisk as no calls are able to route anywhere.|
|Comments:||By: Josh Urbanski (almagsupport) 2011-07-27 09:59:21.003-0500|
This is a back trace after experiencing the deadlock.
By: Josh Urbanski (almagsupport) 2011-07-27 10:13:45.309-0500
Attached core-show-locks file when lock has occurred.
By: Russell Bryant (russell) 2011-07-27 10:30:13.045-0500
The problem here is res_timing_timerfd. If you switch to res_timing_dahdi, this will go away. We're working on resolving the problems with the timerfd module now though.
By: Josh Urbanski (almagsupport) 2011-07-27 11:29:36.544-0500
Thanks for the very fast response. We had previously upgraded to 1.8 to help with another problem we were experiencing from 1.6. We will try the timing adjustment with 1.8.
By: Josh Urbanski (almagsupport) 2011-08-16 10:28:45.850-0500
We have not experienced another deadlock since making this change. Thank you for all your help.
By: Terry Wilson (twilson) 2011-08-17 14:18:44.463-0500
Fix committed to 1.8 branch r332320