[Home]

Summary:ASTERISK-03918: Music On Hold with Re-Invite
Reporter:dsandras (dsandras)Labels:
Date Opened:2005-04-12 09:50:42Date Closed:2005-08-12 09:14:17
Priority:MinorRegression?No
Status:Closed/CompleteComponents:Channels/chan_sip/Interoperability
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) 1.0.5_ok.txt
( 1) 1.0.9_broken.txt
Description:When using Re-Invite between 2 SIP Phones, Asterisk detects that there should be Music On Hold (Hold was pressed), but it doesn't issue a Re-Invite to make the phones switch back from their direct RTP connection to a connection with the Asterisk server. It works with canreinvite=no, the problem is simply that the re-Invite is not sent to tell the IP Phone to connect to Asterisk.

Reverting chan_sip.c from the 1.0.7 version to the 1.0.5 version fixes the problem.
Comments:By: Brian West (bkw918) 2005-04-12 09:53:12

This is related to 0003974

/b

By: dsandras (dsandras) 2005-04-12 10:06:49

I wondered if it was related or not, in 0003974, music on hold is correctly played, so the first reinvite to switch back from IP Phone B to Asterisk is correctly sent. Not in my case. But that is highly similar indeed and I can confirm that chan_sip.c from 1.0.5 didn't have that problem.

By: Kevin P. Fleming (kpfleming) 2005-04-29 13:19:17

What's the status of this? Is it still a problem with CVS STABLE?

By: Clod Patry (junky) 2005-05-09 23:52:38

Reporter seems to have lost interest.

By: dsandras (dsandras) 2005-08-01 08:20:53

It is still the case with Asterisk 1.0.9. I've not lost interest, but I'm astonished I'm the only one to have that problem. Having music on hold with Re-INVITE enabled doesn't seem so uncommon.

Tested with several IP Phones : SNOM, SwissVoice, GrandStream, GnomeMeeting.

An ethereal trace on the Asterisk machine shows Asterisk receives the INVITE from the phone setting the hold, but doesn't send the reINVITE to the other phone to make the RTP stream switch to Asterisk.

By: Olle Johansson (oej) 2005-08-01 10:41:04

According to the bug guidelines, you always have to provide a complete SIP DEBUG. Set verbose to 4, debug to 4 and upload a complete SIP debug from the version where it worked and from the latest release version - 1.0.9. Thank you!

By: dsandras (dsandras) 2005-08-05 09:13:41

Here it is. With 1.0.9, the re-INVITE to make the phone switch the stream from the remote phone to the asterisk server is *sometimes* executed, and the re-INVITE to make the phone switch the stream back from the asterisk server to the remote phone is *never* executed. I could only capture the case where it is not executed.

With 1.0.5, no problem.

However, not sure the trace will help you more than my description, but here it is.

Notice it could be related to 0003974, however I'm not sure as the first re-INVITE is sometimes executed in my case.

By: Mark Spencer (markster) 2005-08-07 22:30:12

Does this bug affect CVS head?

By: dsandras (dsandras) 2005-08-09 04:32:49

I can unfortunately not test with Asterisk CVS. We are only working with the stable branch, and I can confirm the bug is present in 1.0.6, 1.0.7, 1.0.8, 1.0.9.

So except if you have fixed it, it should still be there.

Thank you,

By: Mark Spencer (markster) 2005-08-09 16:11:17

I need to know if it affects head or not in order to properly locate the problem.  Please install head, even if only temporarily, and let me know the result.  Thanks.

By: dsandras (dsandras) 2005-08-09 16:14:17

HEAD of V1-0 or HEAD of the future 1.2?

I'll try to do that tomorrow.

By: Michael Jerris (mikej) 2005-08-09 19:37:48

cvs head, which is what you get when you do a cvs co asterisk with no -r.  The curent stable is v1-0 tag (cvs co -r v1-0 asterisk).

By: dsandras (dsandras) 2005-08-10 03:30:43

Yes the problem is the same with CVS HEAD.
6005 calls 6001. It sends an INVITE. Then Asterisk sends an INVITE to 6001 and 6005 to do a re-INVITE. Then 6001 puts the call on hold; and Asterisk forgets to send an INVITE to 6001 to make the RTP stream switch back from 6001 to Asterisk.

By: Olle Johansson (oej) 2005-08-11 12:11:22

Please test the patch in ASTERISK-3882. Thank you!

By: dsandras (dsandras) 2005-08-12 03:44:58

Yes, that WORKS!

Could you make such a patch available for 1.0.9?
I would like to test it too on the stable branch.

Thanks a lot!

By: Olle Johansson (oej) 2005-08-12 06:51:38

THanks for the fast feedback. I'll see what I can do for stable.

By: Michael Jerris (mikej) 2005-08-12 09:14:17

duplicate of ASTERISK-3882 at this point.  Rolling this into that bug report.