SIP/2.0 503 Service Unavailable - In incoming calls when usign backup GW

Unanswered Question

Dear all,


I have a CUCM 6.1(2) with 4 servers that are divided in 2 sites: CUCM1 and CUCM2 in site1 and CUCM3 and CUCM4 in site2.

In each site I have a GW that is used to connect with SP (Service Provider) by a SIP trunk. So I have Site1-GW and Site2-GW.Both sites are interconnected through a VPLS network.

When all the machines are working OK if a call for site1 is received by the SP it sends it to site1-gw and this negociates with a CCM in site1 the SIP session:

GW CCM1/CCM2

--> INVITE

<-- TRYING

<-- RINGING

--> PRACK

<-- OK

<-- OK with Session Description

and the call works properly.


Simultaneously the same happens for site2. When the SP receives a call for site2 it sends call to site2-gw and this negociates with CCM3 and CCM4 in site2. The calls are established correctly.


The problem starts for incoming calls when one of the gateways is down.

1---

If site1-gw is down the SP realises about it and sends the call to site2-gw. This gateway tries to negociate SIP with CCMs in site1. Firstly, it tries against CCM1

GW CCM1

--> INVITE

<-- TRYING

<-- 503 Service Unavailable (Warning 399 "Routing Failed:ccbid=154281 tcpindex=998 socket=IP_SITE2_GW:5060")


So then the GW tries against CCM2 and it works.

GW CCM2

--> INVITE

<-- TRYING

<-- RINGING

--> PRACK

<-- OK

<-- OK with Session Description

and the call works properly.



2---

If site2-gw is down the SP realises about it and sends the call to site1-gw. This gateway tries to negociate SIP with CCMs in site2. Firstly, it tries against CCM3

GW CCM3

--> INVITE

<-- TRYING

<-- 503 Service Unavailable (Warning 399 "Routing failed: ccbid=7933 tcpindex=563 socket=IP_SITE1_GW:5060')


So then the GW tries against CCM4 and but the same error happens, so the call fails.

GW CCM4

--> INVITE

<-- TRYING

<-- 503 Service Unavailable (Warning 399 "Routing failed: ccbid=7317 tcpindex=624 socket=IP_SITE1_GW:5060')



The outgoing calls always work properly.

We force to use MTPs in the trunks.

The MTPs used are configured in the gateways.

The CSS, Partitions, Regions and locations seem to be OK.


What could be the cause of this behaviour?


Best regards,


Amaia Lesta

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Nicholas Matthews Wed, 08/12/2009 - 14:03

Hi Amaia,


This error is because the IP address that the gateway is sending from isn't recognized by the CUCM.


The source IP address does not match what CUCM has defined for the trunk.


The source IP will be in this order:

1) What is in this configuration:


voice service voip

sip

bind signal interface


2) The IP address associated to the outgoing interface showing in "show ip route "



Either add the bind command to match, or change the IP in CUCM. If you're using DNS, make sure the DNS entry matches the IP address you've either bound to or are routing with.



hth,

nick

Dear all,


I solved the problem.

This is the summary of what I have found:


- The site1-GW1 = site1-trunk1 has associated a Device Pool(DP) which has associated a CCM Group called 'Default'. This CCM group consists of CCM1 and CCM2, which are the CCMs in site1, it has not associated any CCM of site2.


In our tests when a call incoming through site1-GW1 that negociates SIP against CCM3 and CCM4 receives error 503 and isn't established. These gateways don't belong to the CCM-group used by the trunk.


We changed this, we included CCM3 in the CCM-Group and we forced the GW to negociate SIP with CCM1, CCM2 and CCM3 ( a CCM-Group can contain up to 3 CCMs).

And it works. No error 503 now.


- The site2-GW1 = site2-trunk1 has associated a Device Pool(DP) which has a CCM Group called 'CCM-G3'. This CCM group consists of CCM3, CCM4 and CCM2, which are the CCMs in site1 and in site2.


In our tests when a call incoming through site2-GW1 that negociates SIP against CCM1 receives error 503 and isn't established. This gateway doesn't belong to the CCM-group used by the trunk.


Now we have changed this and the gateways negociates SIP against CCM2 and it works.



From the administration guide: http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/admin/6_1_1/ccmcfg/b02cmgrp.html



When a device registers, it attempts to connect to the primary (first) Cisco Unified Communications Manager in the group that is assigned to its device pool. If the primary Cisco Unified Communications Manager is not available, the device tries to connect to the next Cisco Unified Communications Manager that is listed in the group, and so on.


I asked TAC about the sense of the CCM group for a SIP trunk, since they don't register themselves. This was the answer:

For INCOMING calls, the CCM group in the sip trunks to determin through which cucm Server the call will be routed/proccessed first (determining the cucm servers priority to proccess the call)



I hope this could help any of you.


Best regards,

Amaia Lesta

Actions

This Discussion