[Home]

Summary:ASTERISK-14136: Millions of "We're Zap/4-2, not", than crash
Reporter:Andrey Sofronov (andrey sofronov)Labels:
Date Opened:2009-05-15 04:12:29Date Closed:2011-06-07 14:01:05
Priority:CriticalRegression?No
Status:Closed/CompleteComponents:Channels/chan_dahdi
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) gdb.txt
Description:I have an asterisk server with Digium TDM840 card. Sometimes I get millions of messages per second like "We're Zap/4-2, not" or "We're Zap/4-2, not (null)". Than asterisk crashes: asterisk[3634]: segfault at 92765b41 ip b7dd644b sp b6d4fe14 error 4 in libc-2.3.3.so[b7d5a000+142000].

Asterisk - 1.4.24
Zaptel - 1.4.11
Comments:By: Leif Madsen (lmadsen) 2009-05-15 09:55:29

You will need to provide a backtrace from the core dump with DONT_OPTIMIZE compiled into Asterisk (enabled in menuselect).
By: Leif Madsen (lmadsen) 2009-05-15 09:55:54

Additionally, I would attempt to reproduce with the latest DAHDI drivers.
By: Andrey Sofronov (andrey sofronov) 2009-06-05 05:54:14

(gdb) bt
#0  0xb7df644b in strlen () from /lib/tls/libc.so.6
#1  0xb7dc26ab in vfprintf () from /lib/tls/libc.so.6
#2  0xb7de4376 in vsnprintf () from /lib/tls/libc.so.6
#3  0x080fc0cb in ast_dynamic_str_thread_build_va (buf=0xb746da68, max_len=128, ts=0x814a4ec, append=0, fmt=0xb7b21fed "We're %s, not %s\n",
   ap=0xb746da94 "?U\e\b??\f=") at utils.c:1370
#4  0x080ac7a4 in ast_log (level=3, file=0xb7b209e2 "chan_dahdi.c", line=5042, function=0xb7b1a858 "__dahdi_exception", fmt=0xb7b21fed "We're %s, not %s\n")
   at logger.c:801
ASTERISK-1  0xb7b056b4 in __dahdi_exception (ast=Variable "ast" is not available.
) at chan_dahdi.c:5042
ASTERISK-2  0xb7b0bcef in dahdi_exception (ast=Could not find the frame base for "dahdi_exception".
) at chan_dahdi.c:5055
ASTERISK-3  0x08084ecb in __ast_read (chan=0x81cabd8, dropaudio=0) at channel.c:2183
ASTERISK-4  0x080be348 in wait_for_hangup (chan=0x81cabd8, data=Variable "data" is not available.
) at pbx.c:5384
ASTERISK-5  0x080be44b in pbx_builtin_busy (chan=0x81cabd8, data=0x811ed9a) at pbx.c:5420
ASTERISK-6 0x080c7626 in pbx_extension_helper (c=0x81cabd8, con=0x0, context=0x81cad58 "cargo_outgoing", exten=0x81cada8 "989261385132", priority=2, label=0x0,
   callerid=0x81c39f8 "2211", action=E_SPAWN) at pbx.c:537
ASTERISK-7 0x080c9381 in __ast_pbx_run (c=0x81cabd8) at pbx.c:2321
ASTERISK-8 0x080ca387 in ast_pbx_run (c=0x81cabd8) at pbx.c:2679
ASTERISK-9 0xb7b183b5 in ss_thread (data=Variable "data" is not available.
) at chan_dahdi.c:6294
ASTERISK-10 0x080fb47f in dummy_start (data=0x81bafe8) at utils.c:856
ASTERISK-11 0xb7f007d3 in start_thread () from /lib/tls/libpthread.so.0
ASTERISK-12 0xb7e63a8a in clone () from /lib/tls/libc.so.6
By: Andrey Sofronov (andrey sofronov) 2009-06-15 07:03:07

Little more messages (another  crash):
[Jun 15 14:54:38] WARNING[3418] chan_dahdi.c: We're Zap/4-1, not Zap/4-2<ZOMBIE>
[Jun 15 14:54:38] WARNING[3418] chan_dahdi.c: We're Zap/4-1, not Zap/4-2<ZOMBIE>
[Jun 15 14:54:38] WARNING[3418] chan_dahdi.c: We're Zap/4-1, not Zap/4-2<ZOMBIE>
[Jun 15 14:54:38] WARNING[3418] chan_dahdi.c: We're Zap/4-1, not Zap/4-2<ZOMBIE>
[Jun 15 14:54:38] WARNING[3418] chan_dahdi.c: We're Zap/4-1, not Zap/4-2<ZOMBIE>
[Jun 15 14:54:38] WARNING[3418] chan_dahdi.c: We're Zap/4-1, not Zap/4-2<ZOMBIE>
By: Leif Madsen (lmadsen) 2009-06-16 14:17:54

I would also provide 'bt full' and 'thread apply all bt'.

Thanks!  (please attach as text files)
By: Andrey Sofronov (andrey sofronov) 2009-06-18 02:05:17

Done.
Thanks!
By: Avinoam (avinoash) 2009-07-20 04:16:11

We have the same problem, only with:

full error =
"WARNING[24396] chan_dahdi.c: We're DAHDI/5-1, not"

messages error =
"asterisk[24396]: segfault at 80008 ip b7d93133 sp b4e728a4 error 4 in libc-2.8.so[b7d22000+134000]"

system:
-asterisk 1.4.25.1
-DHADI: dahdi-linux-2.1.0.4, dahdi-tools-2.1.0.2
-Kernel: linux-2.6.25-gentoo-r8


By: Leif Madsen (lmadsen) 2009-09-08 09:33:12

Andrey:  Which distribution and kernel version are you using? (curious)
By: Leif Madsen (lmadsen) 2009-09-08 09:36:28

I am assigning this to Tilghman to review the backtrace file. I know this is not his area of expertise. If you are not able to move this issue forward, just set back to an appropriate status. Thanks!
By: Tilghman Lesher (tilghman) 2009-09-08 11:54:25

I think this one is a candidate for valgrind debugging.  See doc/valgrind.txt.
By: Andrey Sofronov (andrey sofronov) 2009-09-10 03:52:09

I'm using asterisk 1.4.26.1, kernel 2.6.26.
Problem has disappeared after removing Busy() application.
Was:
exten => _9., 1, Dial(....)
exten => _9., n, Busy()
exten => _9., n, Hangup()

Now:
exten => _9., 1, Dial(....)
exten => _9., n, Hangup()
By: Leif Madsen (lmadsen) 2009-09-10 07:18:43

OK, can you reproduce this on a test machine, or afterhours at all? We need valgrind output in order to resolve this issue. Thanks!
By: Leif Madsen (lmadsen) 2009-09-30 10:16:51

I'm closing this issue as it has been waiting on feedback for more than 2 weeks. If the reporter has additional information that can be submitted, than they are welcome to reopen the issue. Thanks!