[nsp] VoIP Leg Problem

[Lucas Iglesias]

Hi,

We are having a problem in a very simple VoIP scheme. The net is like this:

 
phone--[PBX-A]-----[Router-A]-------wan-link--------[Router-B]-----[PBX-B]--
phone

Both routers are Cisco 2610, in location A the connection with the PBX is
through FXO voice ports, while in location B is through FXS ports. There is
a PLAR connection configured on both routers (to directly seize the line and
connect to the other PBX without dialing any number). 

This works well, but we found out that after you hang up a call made from A
to B, the router B takes 120 sec. aprox to disconnect (the Voice port is
still taken for the following 120 secs after hanging up).
The PBX-B is a small Panasonic (I don't have the exactly model here).
A "debug ccapi inout" shows that the problem could be on the leg associated
with the local port, not on the voip leg (look at the two legs ID 19 and ID
1A, and the time it takes to disconnect the second one):

 01:44:27: ccCallDisconnect (callID=0x19, cause=0x10 tag=0x0)
 01:44:27: ccCallDisconnect (callID=0x1A, cause=0x10 tag=0x0)
 01:44:27: cc_api_call_disconnect_done(vdbPtr=0x0, callID=0x19, disp=0,
tag=0x0)
 01:44:27: sess_appl: ev(13=CC_EV_CALL_DISCONNECT_DONE), cid(25), disp(0)
 01:44:27: ssaTraceSct:
cid(25)st(7)oldst(6)cfid(-1)csize(0)in(1)fDest(1)-cid2(26)st2(7)oldst2(4)
 Router#
 01:46:27: cc_api_call_disconnect_done(vdbPtr=0x8167D4B0, callID=0x1A,
disp=0, tag=0x0)
 01:46:27: sess_appl: ev(13=CC_EV_CALL_DISCONNECT_DONE), cid(26), disp(0)
 01:46:27: ssaTraceSct: cid(26)st(7)oldst(4)cfid(-1)csize(1)in(0)fDest(0)
 Router#

I've already tested with the command "timeouts call-disconnect" on the
voice-port config without success.

Does anyone have any idea of what could be happening?
Any kind of help would be very appreciated. 

Thank you,
Lucas.