Summary:ASTERISK-05414: Asterisk crashes upon call transfer finalization
Reporter:nirsim (nirsim)Labels:
Date Opened:2005-11-01 10:59:08.000-0600Date Closed:2011-06-07 14:00:38
Versions:Frequency of
Description:Ok, here is the symptom as we've seen it:

1. Caller A calls the Asterisk box.
2. Callee B1 receives the call.
3. Callee B1 transfers the call to Callee B2.
4. Callee B1 hangs up, now A and B2 are talking.
5. Caller A or Callee B2 hangup, asterisk crashes.

Please find below a trace from the core dump, after compilation of Asterisk with valgrind.


0  0x0044b7a2 in ?? () from /lib/ld-linux.so.2
1  0x0048b7d5 in raise () from /lib/tls/libc.so.6
2  0x0048d149 in abort () from /lib/tls/libc.so.6
0000003  0x004bf27a in __fsetlocking () from /lib/tls/libc.so.6
0000004  0x004c5abf in malloc_trim () from /lib/tls/libc.so.6
0000005  0x004c5e3a in free () from /lib/tls/libc.so.6
0000006  0x080a263a in ast_cdr_free (cdr=0x49) at cdr.c:438
0000007  0x0038b86a in nocdr_exec (chan=0x9fb0948, data=0xb79e07d0) at app_cdr.c:50
0000008  0x0808e693 in pbx_extension_helper (c=0x9fb0948, con=0xb79e70a0,
   context=0x9fb0a98 "macro-hangupcall", exten=0x9fb0b8c "s"
;, priority=2, label=0x0,
   callerid=0xb79e07d0 "", action=-1210824960) at pbx.c:564
0000009  0x0808ec3c in ast_spawn_extension (c=0x0, context=0x6 "", exten=0x6 "", prio
rity=6, callerid=0x6 "")
   at pbx.c:2235
0000010 0x008df59c in macro_exec (chan=0x9fb0948, data=0x811a7a4) at app_macro.c:205
0000011 0x0808e693 in pbx_extension_helper (c=0x9fb0948, con=0x0, context=0x9fb0a98 "macro-hangupca
   exten=0x9fb0b8c "s", priority=1, label=0x0, callerid=0xb79e70a0 &q
uot;hangupcall", action=0) at pbx.c:564
0000012 0x0808fbdd in __ast_pbx_run (c=0x9fb0948) at pbx.c:2235
0000013 0x0809149c in pbx_thread (data=0x0) at pbx.c:2515
0000014 0x006de341 in __pthread_initialize_minimal () from /lib/tls/libpthread.so.0
0000015 0x0052afee in clone () from /lib/tls/libc.so.6
Comments:By: nirsim (nirsim) 2005-11-01 11:02:05.000-0600

I couldn't upload the core file, so you can download it from:


By: BJ Weschke (bweschke) 2005-11-01 11:12:48.000-0600

There was a crash bug fixed very recently in relation to CDRs that was exposed frequently through AMP's hangupcall macro. Not having the CVS-HEAD date on this bug, please make sure it's current and try to reproduce this.

By: nirsim (nirsim) 2005-11-01 13:36:27.000-0600

Is beta2 updated for this issue ?

By: Serge Vecher (serge-v) 2005-11-01 13:44:14.000-0600

dimitel, AFAIK, beta2 is a snapshot of CVS HEAD as of early am on11/01/05, so the answer is _yes_. Please test and report back.

By: Clod Patry (junky) 2005-11-02 07:13:35.000-0600

dimitel: You cant upload that core, cause it's too big.
We require only the backtrace, not the whole core file.

Give a look at README.backtrace (probably in /usr/src/asterisk/doc/).


By: BJ Weschke (bweschke) 2005-11-08 10:50:19.000-0600

no addt'l info on this bug of whether or not an upgrade fixed the problem. pls re-open if you are running latest CVS-HEAD and are still having troubles.