Have a customer with cisco 2800 with fxo. Gateway is MGCP go CM back at home office. All ports registered but does not work yet if he plugs in his phone it does work. One interesting thing is that this is an ISDN setup which provides his internet service and phone. Does anyone have any experience with this?
Most analog FXO ports cannot detect short disconnect signals. It happened in UK only where the disconnect signal has length
between 90-130 ms which far' less than 350 ms cuurently used by IOS.
Excuse me, but what wold be the relationship between "disconnect signal" (what is that anyway ?) and an FXO port not working on a router ?
What is pertinent instead, is to know the configuration and how the ports have been tested with the router.
Set up originally as MGCP trace of mgcp packet shows Oct 3 13:23:55: htsp_process_event: [0/0/2, FXOLS_WAIT_DIAL_TONE, E_DSP_SIG_1100]fxols_power_denial_proc
*Oct 3 13:23:55: htsp_timer2 - 1000 msec
*Oct 3 13:23:55: htsp_timer_stop
*Oct 3 13:23:56: htsp_process_event: [0/0/2, FXOLS_WAIT_DIAL_TONE, E_HTSP_EVENT_TIMER2]fxols_timer2_proc
*Oct 3 13:23:56: htsp_timer_stop
*Oct 3 13:23:56: htsp_timer_stop2
*Oct 3 13:23:56: [0/0/2] set signal state = 0x4 timestamp = 0
*Oct 3 13:23:56: MGCP Packet sent to 172.16.25.11:2427--->
403 508603 Setup failed 0:(11:35):0:34
set up as h323 with capi trace
Oct 8 16:31:20: //71/8079525D0100/CCAPI/ccSaveDialpeerTag:
*Oct 8 16:31:20: htsp_timer_stop3 htsp_setup_req
*Oct 8 16:31:20: htsp_process_event: [0/0/0, FXOLS_ONHOOK, E_HTSP_SETUP_REQ]fxols_onhook_setup
*Oct 8 16:31:20: [0/0/0] set signal state = 0xC timestamp = 0
*Oct 8 16:31:20: htsp_timer - 1300 msec
*Oct 8 16:31:20: //72/8079525D0100/CCAPI/cc_api_call_proceeding:
Interface=0x475B3A58, Progress Indication=NULL(0)
*Oct 8 16:31:20: htsp_process_event: [0/0/0, FXOLS_WAIT_DIAL_TONE, E_DSP_SIG_1100]fxols_power_denial_proc
*Oct 8 16:31:20: htsp_timer2 - 1000 msec
*Oct 8 16:31:20: htsp_timer_stop
*Oct 8 16:31:21: htsp_process_event: [0/0/0, FXOLS_WAIT_DIAL_TONE, E_HTSP_EVENT_TIMER2]fxols_timer2_proc
*Oct 8 16:31:21: htsp_timer_stop
*Oct 8 16:31:21: htsp_timer_stop2
*Oct 8 16:31:21: [0/0/0] set signal state = 0x4 timestamp = 0
in addtion, a call in does not even show the port being seized.
I assume you have the correct cptone and impedance configuration.
It seems that when router goes off-hook, it receives a power denial from the switch - this is a bit strange.
Try configuring "no supervisory disconnect signal" and check again.
For incoming, I would again need to see the trace - to see if at least ring is detected.
Thanks for your input, I am pulling my hair out. I have cptone gb I have nothing for impedance, if you have ideas on that I would welcome them. I also have a TAC case opened but the kind of questions they are asking me tell me that I am at the low end of the technical side. As for incoming, if you know of a trace other than ccapi inout let me know because there is absolutely nothing coming in that I can see repeated sh voice call sum doesn't even indicate that the fxo sees the call, yet he merely plugs his phone back in an viola can make and receive calls no problem.
Are you certain that the correct supported PVDM's are installed for this GW?, wouldn't mind taking a look at show diag if you can please.
On top of that - have you seen CSCsc11833
Voice ports are no present in configuration when not enough DSP resources, so that should not be an issue.
To check if you are running into any type of "stuck port" bug, check if reloading the router helps.
However, it is recommended that you run a recent release. Like in most telephony applications, I would recommend 12.4(11)XJ4.
If you want a permanent solution with more features, get a BRI VIC card for the router and connect to the S/T bus.
Thanks for letting us know.
Indeed the power denial after seize was revealing, one way or another the exchange could not supply current for the loop and voltage dropped.