Summary:ASTERISK-08549: Attended transfers using Polycom phones create 100% CPU utilization
Reporter:Anthony Rodgers (cunningpike)Labels:
Date Opened:2007-01-10 15:09:43.000-0600Date Closed:2007-01-10 18:43:13.000-0600
Versions:Frequency of
Environment:Attachments:( 0) gdb.txt
Description:We are finding 99.9% utilization when completing an attended transfer on Polycom phones (full backtrace attached).


What we notice over and over on the CLI when this happens is this message:

   -- Packet2Packet bridging SIP/4012-09c09260 and SIP/4010-09c0e5b0

These are the two _original_ parties to the call. The call progression is:

4012 calls 4010
4010 answers and hits Transfer and dials 4011
4011 answers
4010 hits transfer again
Asterisk hits 99.9% CPU utilization during the new call between 4012 and 4011
CPU utilization returns to normal when call ends

The backtrace ends prematurely because consistently at the same point, gdb stops responding and _its_ CPU utilization rises to 99.9%, requiring a ctrl-z and a restart of Asterisk.
Comments:By: Serge Vecher (serge-v) 2007-01-10 15:23:30.000-0600

I think this has been fixed recently in 1.4 branch. Please test and report back.

By: Anthony Rodgers (cunningpike) 2007-01-10 18:33:42.000-0600

Asterisk SVN-branch-1.4-r50433M fixes the issue. Not sure we want to run this in production though.........

Is there a specific patch for this?

By: Joshua C. Colp (jcolp) 2007-01-10 18:43:13.000-0600

Closing since this has already been fixed and will be in the next 1.4 release.