SRST trouble

Answered Question
Aug 4th, 2007

I'm setting up SRST on my MGCP gateway (2610). I've pretty much set everything up but I'm having problems with inbound calling (outbound works fine). The first issue I'm having is the tranlation rule I set up doesn't, for some reason, seem to work. When I run "test translation-rule 1 xxxxxx" I get "Translation Rule 1 is not defined" as if I don't have one created. The other issue I'm seeing is that when I run "debug voice ccapi inout" and make an inbound call, besides getting a second dial tone, the debug shows the wrong dial-peer being referenced (the MGCP dial peer.)

Anyhow, here is my command set for my translation profile and dial-peer. Please help me resolve, I rate helpful posts;

voice translation-rule 1

rule 1 /.+/ /3000/

!

!

voice translation-profile srst

translate called 1

!

!

dial-peer voice 5000 pots

translation-profile incoming srst

incoming called-number .

destination-pattern 9T

direct-inward-dial

port 1/0/0

!

!

!

-Shikamaru

I have this problem too.
0 votes
Correct Answer by Avner Izhar about 9 years 4 months ago

Do you have just fxo ports on this gateway? if so it might be better to go for h.323 instead of mgcp.

That way you can do the translation on the outbound dial-peer, so you will have two outbound dial-peers, one pointing to call manager and the other to ext 3000.

HTH, Avner.

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4.5 (4 ratings)
Loading.
Avner Izhar Sun, 08/05/2007 - 08:02

Hi,

First of all, to test the voice translation-rule, use the 'test voice translation-rule' command, the one you were using is for testing the 'old' translation rules.

As for MGCP taking control of your inbound calls, have you configured mgcp fallback to srst? if so, does it sense that the fallback mode is active? ('show ccm-manager fallback').

HTH, Avner.

shikamarunara Sun, 08/05/2007 - 08:19

Avner,

When I typed the "test voice translation-rule" command, it works fine, thank you. Now, for the second part when I check fallback mode, this is my output;

WANRouter261001#show ccm-manager fallback-mgcp

Current active Call Manager: None

MGCP Fallback mode: Enabled/ON

Last MGCP Fallback start time: 17:41:06 UTC Mar 1 1993

Last MGCP Fallback end time: None

So, it seems as though SRST is engaged at this point. Just a thought; is there any requirement for an FXO to be able to support SRST mode? My FXO is not v2, I don't know if that matters.

-Shikamaru

Avner Izhar Sun, 08/05/2007 - 08:35

Ohh, I had PRI in the back of my head, if it is FXO, is it controlled by mgcp at all?

If not then put a connection plar on the voice port.

If it is used for mgcp and for srst, then your translation rule should be /\(.*\)/ /3000/ , since there is no called number.

HTH, Avner.

shikamarunara Sun, 08/05/2007 - 08:53

Yes it is. Sorry I forgot to mention. Do I need to add something to its config?

-Shikamaru

Avner Izhar Sun, 08/05/2007 - 09:03

If it is used for mgcp and for srst, then your translation rule should be /\(.*\)/ /3000/ , since there is no called number.

HTH, Avner.

shikamarunara Sun, 08/05/2007 - 09:13

Avner,

The translation doesn't seem to work (same result).

When you say there is "no called number", do you mean that the data recieved from the analog line is different then if the call was made on a DID of a PRI?

-Shikamaru

Avner Izhar Sun, 08/05/2007 - 09:55

Yes, i don't expect the same from fxo.

Can you run 'debug voip dialpeer' and call in while in srst mode? post the results so we can try to figure out what is going on.

shikamarunara Sun, 08/05/2007 - 10:14

Avery,

Here is the output from dialin.

-Shikamaru

WANRouter261001#

*Mar 1 19:39:35.852: Inside dpMatchCore:

*Mar 1 19:39:35.852: destination pattn: expanded string:

*Mar 1 19:39:35.852: Inside dpMatchCore:

*Mar 1 19:39:35.852: destination pattn: expanded string:

*Mar 1 19:39:35.852: Inside dpMatchCore:

*Mar 1 19:39:35.852: destination pattn: expanded string:

*Mar 1 19:39:35.865: %CALL_CONTROL-6-APP_NOT_FOUND: Application mgcpapp in dial-peer 999100 not fou

nd. Handing callid 31 to the alternate app default.

Avner Izhar Sun, 08/05/2007 - 10:46

hmm, looks like it match the mgcp dial-peer even when mgcp is down, can you post the mgcp dial-peer config? do you have more then 'application mgcpapp' and 'port 1/0/0' in there?

shikamarunara Sun, 08/05/2007 - 10:58

Yes, I think we are close to the issue. When I removed the mgcp dial-peer that the output was referencing, the call did go through. Now I have to figure out how the dial-peer will stop being used in an SRST scenario (weird, I've never run into this.)

Here are all of the dial peers.

dial-peer voice 999100 pots

application mgcpapp

port 1/0/0

!

dial-peer voice 999101 pots

application mgcpapp

port 1/0/1

!

dial-peer voice 5000 pots

translation-profile incoming srst

incoming called-number .

destination-pattern 9T

direct-inward-dial

port 1/0/0

-Shikamaru

Avner Izhar Sun, 08/05/2007 - 11:19

well, since there is no calling or called, we do inbound matching acording to the port command, and we prefer the first dial-peer added to the config.

Try to add back the dial-peer you removed, it should still work since it will be added after dial-peer 5000.

Just make sure that dial-peer 5000 will be the first one listed when you exec the 'show run' command.

Let me know, Avner.

shikamarunara Sun, 08/05/2007 - 12:57

Yeah, the heart of the issue is the order of the dial peers. SRST works great until I take it out of SRST mode, then the dial-peer I made for SRST remains active and its the MGCP dial peers that aren't being used.

The MGCP dial peers are placed onto the gateway by CallManager when the gateway registers and downloads its config. I don't suppose there is a way for the gateway to "lose" its MGCP dial peers if it is in SRST mode (i.e; it can't register with the CM?)

-Shikamaru

Correct Answer
Avner Izhar Sun, 08/05/2007 - 17:20

Do you have just fxo ports on this gateway? if so it might be better to go for h.323 instead of mgcp.

That way you can do the translation on the outbound dial-peer, so you will have two outbound dial-peers, one pointing to call manager and the other to ext 3000.

HTH, Avner.

shikamarunara Mon, 08/06/2007 - 17:31

Okay, I have figured out how to resolve this issue. To recap;

- One MGCP gateway with two FXO ports

- CallManager 4.2(3)

- Inbound calling doesn't work when gateway is in SRST mode. Outbound calling works fine.

The issue was that inbound calling was being handled by the MGCP dial peer which, in an SRST situation, is irrelevant. I was able to resolve this problem by configuring "connection plar opx extensionnumber" on the FXO voice-port. For some reason, this syntax has no priority when the gateway is not in SRST mode. In SRST, it trumps the dail peers and takes priority for inbound calling.

This solution is good in situations when inbound calling is forwarded to an Auto Attendant during normal operation but want calls forwarded to the receptionist extension during outages.

-Shikamaru

mdury Wed, 04/16/2008 - 13:14

Shikamaru

Can you post the config you used? I am having the same issue.

shikamarunara Wed, 04/16/2008 - 13:19

The only way I found that would resolve this issue is to set up a "plar opx" syntax on the voice port. Otherwise, the MGCP dial peer is always used as a priority no matter what. It's a weird work around, but it works.

-Shikamaru

mdury Wed, 04/16/2008 - 13:23

Shikamaru,

I tried excatly that. It still tries to match the MGCP

did you have to match your plar opx to any other peer?

What IOS are you running?

Thanks

shikamarunara Wed, 04/16/2008 - 14:38

No, the PLAR goes to the DN. I'm running advanced IP services on a 2610.

-Shikamaru

Actions

This Discussion