[Home]

Summary:ASTERISK-03721: chan_sip not closing channel when RTP goes idle due to a faulty SIP device.
Reporter:Kristopher Lalletti (kris2k)Labels:
Date Opened:2005-03-20 09:56:02.000-0600Date Closed:2011-06-07 14:10:15
Priority:MinorRegression?No
Status:Closed/CompleteComponents:Core/General
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:
Description:chan_sip can't terminate a call when then RTP activity is idle beyond rtptimeout=60 in sip.conf

The outbound call scenario was:
my SPA3000 -SIP-> Asterisk -IAX2-> IAX2LDPeer

Essentially, due to a buggy SIP device that never acknowledged the BYE that was sent from Asterisk, the SIP channel stayed open but my IAX2 channel closed when the conversation ended.

I have included the SIP history, and the message flooding on my console for lack of RTP activity that is beyond the 60 seconds configured in sip.conf.

The only way to come around the problem was to do a forced shutdown. I was unable to do a soft hangup, or a restart of asterisk.

Please advise if you need more, I keep a full log of my daily activities.

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

sip*CLI> sip show history 4722690e-3c585451@10.0.254.251
sip*CLI>
 * SIP Call
1. Rx              INVITE sip:01139xxxxxxxxxx@sip.tmcsolutions.net SIP/2.0 / 101 I
2. CancelDestroy  
3. TxRespRel       SIP/2.0 407 Proxy Authentication Required / 101 INVITE
4. SchedDestroy    15000 ms
5. Rx              ACK sip:01139xxxxxxxxxx@sip.tmcsolutions.net SIP/2.0 / 101 ACK
6. Rx              INVITE sip:01139xxxxxxxxxx@sip.tmcsolutions.net SIP/2.0 / 102 I
7. CancelDestroy  
8. CancelDestroy  
9. TxResp          SIP/2.0 100 Trying / 102 INVITE
10. TxResp          SIP/2.0 183 Session Progress / 102 INVITE
11. TxRespRel       SIP/2.0 200 OK / 102 INVITE
12. Rx              ACK sip:01139xxxxxxxxxx@64.26.157.238 SIP/2.0 / 102 ACK
13. Rx              BYE sip:01139xxxxxxxxxx@64.26.157.238 SIP/2.0 / 103 BYE
14. TxResp          SIP/2.0 200 OK / 103 BYE

Mar 20 10:46:14 NOTICE[6162]: chan_sip.c:8666 do_monitor: Disconnecting call 'SIP/kri-spa-200-d624' for lack of RTP activity in 686s
Mar 20 10:46:14 NOTICE[6162]: chan_sip.c:8666 do_monitor: Disconnecting call 'SIP/kri-spa-200-d624' for lack of RTP activity in 686s
Mar 20 10:46:14 NOTICE[6162]: chan_sip.c:8666 do_monitor: Disconnecting call 'SIP/kri-spa-200-d624' for lack of RTP activity in 686s
Mar 20 10:46:14 NOTICE[6162]: chan_sip.c:8666 do_monitor: Disconnecting call 'SIP/kri-spa-200-d624' for lack of RTP activity in 686s
Mar 20 10:46:14 NOTICE[6162]: chan_sip.c:8666 do_monitor: Disconnecting call 'SIP/kri-spa-200-d624' for lack of RTP activity in 686s
Mar 20 10:46:15 NOTICE[6162]: chan_sip.c:8666 do_monitor: Disconnecting call 'SIP/kri-spa-200-d624' for lack of RTP activity in 687s
Mar 20 10:46:16 NOTICE[6162]: chan_sip.c:8666 do_monitor: Disconnecting call 'SIP/kri-spa-200-d624' for lack of RTP activity in 688s
Comments:By: Olle Johansson (oej) 2005-03-20 11:15:31.000-0600

We need a SIP debug output as well, thank you! Set verbose to 4 and debug to 4 and capture *all* output.
By: Olle Johansson (oej) 2005-03-20 11:23:27.000-0600

The rtptimeout also does a softhangup - neither of you succeeded. There's something else going on there.

Can you please try this on latest CVS head? I need to know if the same situation appears there.

From the siphistory, the bye was acknowledged by a 200 OK, so I don't see the problem... Is this the history of the call that doesn't die?
By: Kristopher Lalletti (kris2k) 2005-03-20 11:30:28.000-0600

I'll upgrade this afternoon. However,  I was unable to reproduce the error, so, its difficult to provide with any supplemental traces of that level.

Yes, the siphistory was the history of the call that didn't die. The SIP channel stayed open, but my IAX channel closed when the conversation ended by the remote peer.

Anyhow, I will try to reproduce it with debug traces.
By: Mark Spencer (markster) 2005-03-20 22:28:27.000-0600

Also need to see the output of "sip show channel <foo>" where "<foo>" is the callid of this channel.  It looks as though we should be disconnecting especially since we acknowledged the bye.
By: Mark Spencer (markster) 2005-03-30 00:55:30.000-0600

bug placer appears to have lost interest.