[Home]

Summary:ASTERISK-15131: sip calls drop because of BYE's
Reporter:Sean Darcy (seandarcy)Labels:
Date Opened:2009-11-12 16:30:16.000-0600Date Closed:2010-01-26 08:03:44.000-0600
Priority:MajorRegression?Yes
Status:Closed/CompleteComponents:Channels/chan_sip/General
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) sip.conf-and-extensions.conf
( 1) sip.debug-clean
( 2) sip-error-sessions-resume-clean
Description:On 1.6.0.16, 1.6.0.18-rc1 and 2, SIP calls die because * sends a BYE. We have 12 internal SIP phones. The server connects them to sip, iax and dahdi over libpri. iax and dahdi work fine. The internal SIP phones work fine. But outgoing SIP calls are terminated with a BYE. We've tried sip over both Teliax and Junction. Same result. Teliax support told us they could see the BYE.

****** STEPS TO REPRODUCE ******

Dial out over SIP.

****** ADDITIONAL INFORMATION ******

I'll attach a SIP debug, but here's an example from the log:

[Nov  9 10:47:31] VERBOSE[26967] logger.c:     -- Executing [9178858521@longdistance:1] �[1;36;40mAnswer�[0;37;40m("�[1;35;40mSIP/169-00000033�[0;37;40m", "�[1;35;40m�[0;37;40m") in new stack
[Nov  9 10:47:32] VERBOSE[26967] logger.c:     -- Executing [9178858521@longdistance:2] �[1;36;40mGosub�[0;37;40m("�[1;35;40mSIP/169-00000033�[0;37;40m", "�[1;35;40mDialOut,s,1(917XXXYYYY,SIP/teliax-sip)�[0;37;40m") in new stack
[Nov  9 10:47:32] VERBOSE[26967] logger.c:     -- Executing [s@DialOut:1] �[1;36;40mNoOp�[0;37;40m("�[1;35;40mSIP/169-00000033�[0;37;40m", "�[1;35;40mContext: DialOut called with outgoing number 917XXXYYYY, and Preferred-Out=SIP/teliax-sip�[0;37;40m") in new stack
[Nov  9 10:47:32] VERBOSE[26967] logger.c:     -- Executing [s@DialOut:2] �[1;36;40mNoOp�[0;37;40m("�[1;35;40mSIP/169-00000033�[0;37;40m", "�[1;35;40m169=the local extension�[0;37;40m") in new stack
[Nov  9 10:47:32] VERBOSE[26967] logger.c:     -- Executing [s@DialOut:3] �[1;36;40mGotoIf�[0;37;40m("�[1;35;40mSIP/169-00000033�[0;37;40m", "�[1;35;40m0?dial-out�[0;37;40m") in new stack
[Nov  9 10:47:32] VERBOSE[26967] logger.c:     -- Executing [s@DialOut:4] �[1;36;40mSet�[0;37;40m("�[1;35;40mSIP/169-00000033�[0;37;40m", "�[1;35;40mCALLERID(num)=2124531169�[0;37;40m") in new stack
[Nov  9 10:47:32] VERBOSE[26967] logger.c:     -- Executing [s@DialOut:5] �[1;36;40mDial�[0;37;40m("�[1;35;40mSIP/169-00000033�[0;37;40m", "�[1;35;40mSIP/teliax-sip/9178858521�[0;37;40m") in new stack
[Nov  9 10:47:32] VERBOSE[26967] logger.c:   == Using SIP RTP TOS bits 184
[Nov  9 10:47:32] VERBOSE[26967] logger.c:   == Using SIP RTP CoS mark 5
[Nov  9 10:47:32] VERBOSE[26967] logger.c:   == Using SIP VRTP CoS mark 6
[Nov  9 10:47:32] VERBOSE[26967] logger.c:   == Using UDPTL TOS bits 184
[Nov  9 10:47:32] VERBOSE[26967] logger.c:   == Using UDPTL CoS mark 5
[Nov  9 10:47:32] VERBOSE[26967] logger.c:     -- Called teliax-sip/917XXXYYYY
[Nov  9 10:47:33] VERBOSE[26967] logger.c:     -- SIP/teliax-sip-00000034 is making progress passing it to SIP/169-00000033
[Nov  9 10:47:42] VERBOSE[26967] logger.c:     -- SIP/teliax-sip-00000034 answered SIP/169-00000033
[Nov  9 10:47:42] VERBOSE[26967] logger.c:     -- Packet2Packet bridging SIP/169-00000033 and SIP/teliax-sip-00000034
[Nov  9 10:47:48] WARNING[29749] chan_sip.c: Autodestruct on dialog '51e773197ab110b42491c1014644c8ee@aa.bb.ccc.ddd' with owner in place (Method: INVITE)
[Nov  9 10:47:48] VERBOSE[26967] logger.c:   == Spawn extension (DialOut, s, 5) exited non-zero on 'SIP/169-00000033'
Comments:By: Sean Darcy (seandarcy) 2009-11-12 16:49:29.000-0600

Attached sip.debug-clean. About line 514 you can see the BYE being sent. I don't see anything that shows why the server decided to send a BYE.

Also attached is sip.conf and extensions.conf
By: Leif Madsen (lmadsen) 2009-11-13 07:55:49.000-0600

Can you try adding

session-timers=refuse

to your sip.conf and see if that helps?
By: Sean Darcy (seandarcy) 2009-11-13 12:36:12.000-0600

Nope. The call drops when it's picked up. Here's CLI and sip.conf context. I'll upload sip debug.

   -- Executing [1XXXYYY4299@internal:1] Answer("DAHDI/1-1", "") in new stack
   -- Executing [1XXXYYY4299@internal:2] NoOp("DAHDI/1-1", "Context: outbound-long-distance") in new stack
   -- Executing [1XXXYYY4299@internal:3] Dial("DAHDI/1-1", "SIP/teliax-sip/XXXYYY4299") in new stack
 == Using SIP RTP TOS bits 184
 == Using SIP RTP CoS mark 5
 == Using SIP VRTP TOS bits 136
 == Using SIP VRTP CoS mark 4
 == Using UDPTL TOS bits 184
 == Using UDPTL CoS mark 5
   -- Called teliax-sip/XXXYYY4299
   -- SIP/teliax-sip-00000003 is making progress passing it to DAHDI/1-1
   -- SIP/teliax-sip-00000003 answered DAHDI/1-1
   -- fixed jitterbuffer created on channel DAHDI/1-1
WARNING[8250]: chan_sip.c:3587 __sip_autodestruct: Autodestruct on dialog '1ea0b25d06cdf9d36417a52e65d52f1a@76.248.146.19' with owner in place (Method: INVITE)
 == Spawn extension (internal, 1XXXYYY4299, 3) exited non-zero on 'DAHDI/1-1'
   -- Hungup 'DAHDI/1-1'
   -- fixed jitterbuffer destroyed on channel DAHDI/1-1


[teliax-sip]
type=peer
nat=yes
canreinvite=no
disallow=all
allow=ulaw
allow=gsm
dtmfmode=rfc2833
insecure=port,invite
context=from-teliax-sip ; no such context. There should be no calls from teliax.
defaultuser=USERNAME(Sip)
secret=PASSWORD
host=nyc.teliax.net
qualify=yes
session-timers=refuse
By: Bereterbide Marcelo (marhbere) 2009-11-16 10:32:01.000-0600

I see It´s related to 0016185
By: Leif Madsen (lmadsen) 2009-11-16 10:58:19.000-0600

Can you elaborate as to what this is related? (I'm presuming commit there is what is causing your issue here)
By: Michael L. Young (elguero) 2009-11-16 11:20:17.000-0600

It appears to be related.  The other issue is not a good description.  There really is no crash from what I can discern.  What is happening is that the call is hanging up as soon as the called sip channel is answered.

I believe this post on the dev list nails it on the head:  http://lists.digium.com/pipermail/asterisk-dev/2009-November/040582.html

Olle's response:
http://lists.digium.com/pipermail/asterisk-dev/2009-November/040609.html

This only affects outbound calls.  Inbound calls don't appear to be affected.  I have tried turning on ignoresdpversion and the outbound calls work again as mentioned on the mailing list.

Hope this information helps to narrow the issue down.

(I don't think that the commit mentioned in the other bug report is the culprit but the issues do seem related based on the logs that have been posted)


By: Alex Balashov (abalashov) 2009-11-25 19:41:25.000-0600

I am getting this on Asterisk 1.6.2.0rc6-1 as well (Debian unstable package).
By: Elazar Broad (ebroad) 2009-11-25 22:07:52.000-0600

I have been noticing this as well on SVN trunk(recent, within 24 hours). Asterisk sends an immediate bye after the callee picks up, nothing really in the logs.

Edit: ignoresdpversion=yes in the peer definition(in this case Viatalk) resolves the issue, as stated above.


By: jiddings (jiddings) 2009-11-26 12:01:27.000-0600

Using 1.6.2.0-rc6 here, having the same issue with outbound calls.

Adding 'ignoresdpversion=yes' it solves the issue for me as well (using Teliax).
By: Leif Madsen (lmadsen) 2009-11-30 13:45:50.000-0600

Any chance issue ASTERISK-15159 resolved this?
By: Leif Madsen (lmadsen) 2009-11-30 13:48:15.000-0600

It did not, per seanbright on IRC who tested.
By: Digium Subversion (svnbot) 2009-11-30 14:51:58.000-0600

Repository: asterisk
Revision: 231602

U   trunk/channels/chan_sip.c

------------------------------------------------------------------------
r231602 | file | 2009-11-30 14:51:57 -0600 (Mon, 30 Nov 2009) | 5 lines

When receiving SDP that matches the version of the last one do not treat it as a fatal error.

(closes issue ASTERISK-15131)
Reported by: seandarcy

------------------------------------------------------------------------

http://svn.digium.com/view/asterisk?view=rev&revision=231602
By: Digium Subversion (svnbot) 2009-11-30 14:53:02.000-0600

Repository: asterisk
Revision: 231603

_U  branches/1.6.0/
U   branches/1.6.0/channels/chan_sip.c

------------------------------------------------------------------------
r231603 | file | 2009-11-30 14:53:02 -0600 (Mon, 30 Nov 2009) | 12 lines

Merged revisions 231602 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk

........
 r231602 | file | 2009-11-30 16:44:30 -0400 (Mon, 30 Nov 2009) | 5 lines
 
 When receiving SDP that matches the version of the last one do not treat it as a fatal error.
 
 (closes issue ASTERISK-15131)
 Reported by: seandarcy
........

------------------------------------------------------------------------

http://svn.digium.com/view/asterisk?view=rev&revision=231603
By: Digium Subversion (svnbot) 2009-11-30 14:54:32.000-0600

Repository: asterisk
Revision: 231604

_U  branches/1.6.1/
U   branches/1.6.1/channels/chan_sip.c

------------------------------------------------------------------------
r231604 | file | 2009-11-30 14:54:32 -0600 (Mon, 30 Nov 2009) | 12 lines

Merged revisions 231602 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk

........
 r231602 | file | 2009-11-30 16:44:30 -0400 (Mon, 30 Nov 2009) | 5 lines
 
 When receiving SDP that matches the version of the last one do not treat it as a fatal error.
 
 (closes issue ASTERISK-15131)
 Reported by: seandarcy
........

------------------------------------------------------------------------

http://svn.digium.com/view/asterisk?view=rev&revision=231604
By: Digium Subversion (svnbot) 2009-11-30 14:55:21.000-0600

Repository: asterisk
Revision: 231605

_U  branches/1.6.2/
U   branches/1.6.2/channels/chan_sip.c

------------------------------------------------------------------------
r231605 | file | 2009-11-30 14:55:20 -0600 (Mon, 30 Nov 2009) | 12 lines

Merged revisions 231602 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk

........
 r231602 | file | 2009-11-30 16:44:30 -0400 (Mon, 30 Nov 2009) | 5 lines
 
 When receiving SDP that matches the version of the last one do not treat it as a fatal error.
 
 (closes issue ASTERISK-15131)
 Reported by: seandarcy
........

------------------------------------------------------------------------

http://svn.digium.com/view/asterisk?view=rev&revision=231605
By: Digium Subversion (svnbot) 2009-11-30 14:58:39.000-0600

Repository: asterisk
Revision: 231606

U   tags/1.6.0.19/channels/chan_sip.c

------------------------------------------------------------------------
r231606 | file | 2009-11-30 14:58:38 -0600 (Mon, 30 Nov 2009) | 5 lines

When receiving SDP that matches the version of the last one do not treat it as a fatal error.

(closes issue ASTERISK-15131)
Reported by: seandarcy

------------------------------------------------------------------------

http://svn.digium.com/view/asterisk?view=rev&revision=231606
By: Digium Subversion (svnbot) 2009-11-30 14:58:56.000-0600

Repository: asterisk
Revision: 231607

U   tags/1.6.1.11/channels/chan_sip.c

------------------------------------------------------------------------
r231607 | file | 2009-11-30 14:58:55 -0600 (Mon, 30 Nov 2009) | 5 lines

When receiving SDP that matches the version of the last one do not treat it as a fatal error.

(closes issue ASTERISK-15131)
Reported by: seandarcy

------------------------------------------------------------------------

http://svn.digium.com/view/asterisk?view=rev&revision=231607