|Summary:||ASTERISK-08460: Segfault on 1.4.0 with E1s|
|Date Opened:||2006-12-30 16:41:38.000-0600||Date Closed:||2007-02-26 15:38:37.000-0600|
|Environment:||Attachments:||( 0) backtrace1.txt|
( 1) backtrace2.txt
( 2) backtrace3.txt
|Description:||Hi :) We've been running 1.4.0 for a couple of days and have had a segfault |
Ultimately the crash is the result of the same thing:
#0 0x00002aaaae2c61c8 in pri_schedule_del (pri=0x108d9f0, id=-553648128) at
118 pri->pri_sched[id].callback = NULL;
Asterisk had not been compiled with DEBUG_THREADS or DONT_OPTIMIZE since I was not expecting a crash ;) I have now recompiled and will submit another backtrace if/when it crashes again.
****** ADDITIONAL INFORMATION ******
This trace was from a busy system with 2x E1 bearers, and dozens of SIP phones and queues.
|Comments:||By: gdhgdh (gdhgdh) 2006-12-30 16:43:04.000-0600|
backtrace2 was from the same system the next day when the workload was much lighter - no SIP phones, no queues.. just passing calls through to a legacy PBX connected to the other two E1 ports.
By: Brandon Kruse (bkruse) 2006-12-30 19:30:50.000-0600
Is this built without optimizations??
if its not, cd into your asterisk source dir ( probably /usr/src/asterisk )
type make menuselect go to build options, select dont optimize(by hitting space) and then hit x
type make && make install
and try to reproduce it again.
If its already built without optimization, I bet this is going to go to mattf
By: gdhgdh (gdhgdh) 2006-12-30 20:40:03.000-0600
bkruse: Please read the last paragraph of the report.
By: Brandon Kruse (bkruse) 2006-12-31 00:25:49.000-0600
well, the unoptimized build is what helps the devs :]
By: gdhgdh (gdhgdh) 2007-01-02 12:18:45.000-0600
And as if by magic, Asterisk crashed again- see backtrace3.txt - although this time the type of crash seems to have nothing to do with libpri :(
By: Serge Vecher (serge-v) 2007-02-21 14:48:58.000-0600
alright, have you been able to reproduce this with 1.4 checked out from SVN?
By: gdhgdh (gdhgdh) 2007-02-26 15:24:39.000-0600
No longer able to test. You might as well close this.