[Home]

Summary:ASTERISK-23389: Asterisk 1.8.23.0 Crashes Signal 11/Segfault (app_queue QueueRemove Address Out of Bounds)
Reporter:Jamuel Starkey (jamuel)Labels:
Date Opened:2014-02-27 12:34:34.000-0600Date Closed:2014-03-03 17:21:23.000-0600
Priority:MajorRegression?
Status:Closed/CompleteComponents:Applications/app_queue
Versions:1.8.23.0 Frequency of
Occurrence
Related
Issues:
Environment:CentOS 5 32-bit on Citrix XenServer 5.6 SP2Attachments:( 0) backtrace-AST-23389.txt
Description:Saw this segfault on a production box--only seen it once but this box was just updated to 1.8.23.0 the night before.  Have a core with debug threads enabled.  Uploading back trace.
Comments:By: Jamuel Starkey (jamuel) 2014-02-27 12:36:53.691-0600

Backtrace optioned as gdb -se "asterisk" -ex "bt full" -ex "thread apply all bt".

By: Jamuel Starkey (jamuel) 2014-02-27 12:43:55.349-0600

Updated backtrace as thread debug was not provided in the first bt.

By: Rusty Newton (rnewton) 2014-02-28 09:02:44.943-0600

Do you happen to have an Asterisk full log from when the issue occurred?

Seeing as you said you only saw it once, I gather that you are unable to reproduce the issue?

By: Jamuel Starkey (jamuel) 2014-03-02 23:57:06.570-0600

Here's what was in the log just before the crash (nothing before it for ~ 10 seconds)
{code}
[Feb 27 09:50:05] VERBOSE[31297] app_dial.c:     -- Connected line update to SIP/HCVSBC_DR-00000a60 prevented.
[Feb 27 09:50:05] VERBOSE[31297] app_dial.c:     -- SIP/0004F237E1E8-00000a61 answered SIP/HCVSBC_DR-00000a60
[Feb 27 09:50:05] VERBOSE[31297] pbx.c:     -- Executing [s@macro-auto-blkvm:1] Set("SIP/0004F237E1E8-00000a61", "__MACRO_RESULT=") in new stack
[Feb 27 09:50:05] VERBOSE[31297] pbx.c:     -- Executing [s@macro-auto-blkvm:2] DBdel("SIP/0004F237E1E8-00000a61", "BLKVM/4242146719/SIP/HCVSBC_DR-00000a60") in new stack
[Feb 27 09:50:05] VERBOSE[31297] app_db.c:     -- DBdel: family=BLKVM, key=4242146719/SIP/HCVSBC_DR-00000a60
[Feb 27 09:50:12] VERBOSE[31297] res_musiconhold.c:     -- Started music on hold, class 'default', on SIP/HCVSBC_DR-00000a60
{code}
CRASH
{code}
[Feb 27 09:50:44] VERBOSE[31744] config.c:   == Parsing '/etc/asterisk/asterisk.conf': [Feb 27 09:50:44] VERBOSE[31744] config.c:   == Found
{code}

By: Jamuel Starkey (jamuel) 2014-03-02 23:57:42.491-0600

See previous comment for full log snippet.  

By: Jamuel Starkey (jamuel) 2014-03-02 23:58:57.255-0600

Yes, unable not able to produce and haven't seen since.  If the issue is with QueueRemove we do have users that use FOP2 to routinely change agent status (over AMI).  Not sure if that helps at all . . .

By: Rusty Newton (rnewton) 2014-03-03 17:21:23.772-0600

Unfortunately the backtrace wasn't helpful for the developers I asked to look at it. It isn't clear how things are getting into the state they are in. Plus the full log you provided didn't have the additional message types we needed to see. I'm going to close it out until we can get additional information.

Check out the [debug information on the wiki|https://wiki.asterisk.org/wiki/display/AST/Collecting+Debug+Information] and next time the issue occurs make sure to get a full log with all the message types, verbose and debug turned up to level 5 the . That should show us more about the manager action involved and maybe we'll be able to reproduce the issue.

At that point you'll want to attach your various queue configuration files and dialplan as well.

If we don't respond when you post the new information to the issue, then ask in #asterisk-bugs or #asterisk-dev for someone to reopen the issue. Thanks!