Cisco fxs goes on hold, sends 'sendonly' but no RTP?

[tod.higman]

Hello,

I've got some expierience with the CUBE, but now I'm setting up a proof of concept where it's SIP to FXS.  I'm new to FXS.

Nothing fancy, a SIP call comes into the 2921 with FXS, the dial peer routes to the FXS voice card port, the analog phone rings, answers, and RTP flows both ways normally.

When the analog phone performs a hookflash, I get the dialtone at the alanog phone and a new INVITE is generated by the 2921 back up to to the original SIP UA. That INVITE includes "Media Attribute (a): sendonly", but the RTP stops from the 2921.  This is causing the far side to time out with the call because it's not getting the RTP it's looking for.

This doesn't look like normal behaviour to me, but before I start posting configs, I wanted to ask if this is normal and/or is there a simple setting to correct my problem?

Thanks in advance.

[Ayodeji Okanlawon · Accepted solution]

Todd, this is normal, the far end should send media attribute a=receive only, upon receipt of a send only. In Sip when a call is on hold, it's not a two way traffic, hence why you see send only

[tod.higman]

Thanks.  Is there a way to get the 2921 to send a "a=inactive" instead?

[Ayodeji Okanlawon]

I guess you can use sip profiles to modify the attributes, but that may break something else..

[tod.higman]

Okay, thanks for the heads up.