[asterisk-bugs] [Asterisk 0018842]: random deadlocks - SIP messages not being processed

[Asterisk Bug Tracker]

A NOTE has been added to this issue. 
====================================================================== 
https://issues.asterisk.org/view.php?id=18842 
====================================================================== 
Reported By:                usinternet
Assigned To:                
====================================================================== 
Project:                    Asterisk
Issue ID:                   18842
Category:                   PBX/General
Reproducibility:            random
Severity:                   major
Priority:                   normal
Status:                     acknowledged
Asterisk Version:           1.8.2.3 
JIRA:                       SWP-3158 
Regression:                 No 
Reviewboard Link:            
SVN Branch (only for SVN checkouts, not tarball releases): N/A 
SVN Revision (number only!):  
Request Review:              
====================================================================== 
Date Submitted:             2011-02-18 15:28 CST
Last Modified:              2011-02-22 13:13 CST
====================================================================== 
Summary:                    random deadlocks - SIP messages not being processed
Description: 
Just upgraded to Asterisk 1.8.2 for production and we're getting random
deadlocks.  When this happens, it looks like Asterisk is not able to
process new SIP messages.  Any INVITEs or REGISTERs that arrive are not
being outputted to the console or logs.

This seems to happen randomly.  I have not been able to reproduce it at
will although it does happen about 5-10 times per day.  The server is
handling about 2000 calls/day, so it's not really that busy.  There are
about 75 phones connected to the server.

Please see attached backtrace and lock output for more information.  It
shows the deadlocks but I'm not sure which is the root issue here.
====================================================================== 

---------------------------------------------------------------------- 
 (0132278) cmaj (reporter) - 2011-02-22 13:13
 https://issues.asterisk.org/view.php?id=18842#c132278 
---------------------------------------------------------------------- 
luckman212, I'm running only res_timing_dahdi for timing, and MOH works
fine. This is on 1.8.3-rc2 and a Sangoma card - couple T1's clocked from
the telco - with ulaw across the board.

As for the last core show locks, core-show-locks-11021912:05:44.txt, that
looks like Mysql lock contention, one lock being held to the Mysql backend
for both CDR insertion and realtime config. If whoever posted that could
try turning off Mysql CDR logging for a bit, then it might help them. Or
try setting "batchmode=yes" in cdr.conf so you can reduce some contention
for that lock (this would cause a post of the CDR in a separate thread from
the channel Hangup routine.) 

Issue History 
Date Modified    Username       Field                    Change               
====================================================================== 
2011-02-22 13:13 cmaj           Note Added: 0132278                          
======================================================================