[asterisk-bugs] [Asterisk 0011999]: memory leak

[noreply at bugs.digium.com]

A NOTE has been added to this issue. 
====================================================================== 
http://bugs.digium.com/view.php?id=11999 
====================================================================== 
Reported By:                destiny6628
Assigned To:                
====================================================================== 
Project:                    Asterisk
Issue ID:                   11999
Category:                   Core-General
Reproducibility:            always
Severity:                   major
Priority:                   normal
Status:                     feedback
Asterisk Version:           1.4.18 
SVN Branch (only for SVN checkouts, not tarball releases): N/A 
SVN Revision (number only!):  
Disclaimer on File?:        N/A 
Request Review:              
====================================================================== 
Date Submitted:             02-15-2008 01:20 CST
Last Modified:              03-07-2008 10:13 CST
====================================================================== 
Summary:                    memory leak
Description: 
hi we are using asterisk 1.4.18 , zaptel-1.4.8 and libpri 1.4.3 with 4 e1
card on ibm x3200 server and 90 active calls running on asterisk cli and
after some hours of dialing , swap memory starts getting used and asterisk
gives core dump .

Asterisk is running with safe_asterisk.

vmstat ouput is as follows 

[root at localhost cron]# vmstat
procs -----------memory---------- ---swap-- -----io---- --system--
-----cpu------
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id
wa st
 1  0    568  57308  44708 1848488    0    0    10   119 1293  509  2  2
95  0  0

Free -m is as follows 

[root at localhost cron]# free -m
             total       used       free     shared    buffers     cached
Mem:          2025       1974         51          0         43       1806
-/+ buffers/cache:        124       1901
Swap:         1983          0       1983


OPERATING SYSTEM is centos 5 and kernel version is [root at localhost cron]#
uname -a
Linux localhost.localdomain 2.6.18-8.el5 http://bugs.digium.com/view.php?id=1
SMP Thu Mar 15 19:57:35 EDT
2007 i686 i686 i386 GNU/Linux

Cannot do valgrind test as it put heavy load on the server and its a
production server .

RAM On the server is 2GB .

Only Asterisk application is running on that .



====================================================================== 

---------------------------------------------------------------------- 
 putnopvut - 03-07-08 10:13  
---------------------------------------------------------------------- 
Update: I have not found the cause of the crashes you are experiencing.
What makes things difficult is that it appears that your crashes are
happening all over the place, and it makes it even more difficult to
pinpoint the crash cause. Here are some of the observations I've made:

1. The most common causes of crashes seem to come from frame list
corruption, either on the channel's readq or another frame list, although
some of the crashes you have presented are happening in dsp code or other
channel-related code, indicating that perhaps the channel itself is getting
corrupted.

2. The cored.txt upload shows for once that there is definitely corruption
in the frame list. If you look at the bt full from that upload, the value
of cur in frame 0 is 0xdeadbeef, which is a magic value used to separate
allocations and should absolutely NEVER show up as an actual value for any
variable.

The console output that you uploaded from March 6 is odd because it
doesn't show a crash. It actually shows Asterisk exiting cleanly.

The one thing you could do that would be most helpful in resolving this
situation is to run just one of your servers under valgrind. We're working
very hard trying to find an answer, but when my local tests don't crash,
valgrind output from the machine where crashes are occurring is the most
helpful thing we can have.

Also, to echo jmls, are you using ODBC at all, and what are you doing in
your AGI script(s)? 

Issue History 
Date Modified   Username       Field                    Change               
====================================================================== 
03-07-08 10:13  putnopvut      Note Added: 0083598                          
======================================================================