We've found that even non-glare scenarios are leaking. Since we just introduced a new stack, we're unsure at this point whether is our new stack or something in FreeSWITCH. However, after running valgrind, using MALLOC_CHECK_ and mtrace(), those tools report issues at switch_event.c, where one of the headers value being freed was never allocated (and therefore may be the value that was really allocated is not being freed), or is some kind of heap corruption. Many hypothesis at this point, nothing concrete.

Just thought in letting you know, it may not even be the same issue.

Moises Silva
Senior Software Engineer
Sangoma Technologies Inc. | NEW 100 Renfrew Drive, Suite 100, Markham ON L3R 9R6 Canada
t. 1 905 474 1990 x128 | e. moy@sangoma.com

On Wed, Sep 15, 2010 at 4:47 PM, Oleg Khovayko <khovayko@gmail.com> wrote:

Hi,

Just for information:

Time to time, very seldom, I see strange bug with my FS-machine.
FreeSwitch starts to use ~60% of CPU, and eat more and more memory.
When it reach ~500MB, system kill this task.

Restart does not help. Reboot also. After reboot, FS, when starts,
continues to eat CPU and memory.

When I try if from terminal, and use command "shutdown",
system not exit completely, and freeze, just after print about
destroying all connections to SIP-providers.

When I tried to debug it, looks like problem in the thread #5.

And, last time, I seen following message, after which FS starts working
normally:

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x28802250 (LWP 100075)]
sqlite3HashInsert (pH=0x0, pKey=0x297e7d78, nKey=15, data=0x28c5a080) at
./src/hash.c:347
347 xHash = hashFunction(pH->keyClass);
(gdb)

_______________________________________________
FreeSWITCH-dev mailing list
FreeSWITCH-dev@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev
http://www.freeswitch.org