CUPS 8.5(1) - CUCM Presence Gateways Change? DNS-SRV

[Joshua Warcop]

Hello,

Configuring a new integration and have headed over to add the CUCM presence gateway to the configuration. I have DNS SRV records created for each CUCM node. When I go to configure the entry in Presence administration I input "cucm.domain.com" I still get the system troubleshooter message - "invalid presence gateway". If I try and input "_sip._udp.cucm.domain.com" as the presence gateway I get a popup saying I need to input a valid hostname or ip address. I'm guessing something has changed with what is supported here? Checking on assistance here before TAC.

SRV Record entries

_sip._tcp.cucm.domain.com = pub.domain.com

_sip._tcp.cucm.domain.com = sub.domain.com

_sip._udp.cucm.domain.com =pub.domain.com

_sip._udp.cucm.domain.com = sub.domain.com

SIP Proxy Service Settings:

Proxy domain: domain.com

SRV Cluster Name: cups.domain.com

Address Resolution Type: SRV

CUCM Enterprise Params:

TLD: domain.com

Cluster FWDN: cucm.domain.com

CUCM Trunk:

Destination is SRV: Checked (checking this box sets the port to 0/changing the port unchecks the box)

Destination address: cups.domain.com

[Christos Georgiadis · Accepted solution]

Hi,

Just a thought here that might help.

Did you try to ignore the troubleshooter and move forward with the rest of the configuration? I am saying this as I suspect that the troubleshooter tries to resolve the cucm.domain.com via the DNS and of course it is not successful as this is not a 'A record'

You should not put the_sip._tcp part.

I also wanted to draw your attention to the following defect

CSCth25928

as it will probably apply in your case.

HTH,

Christos

[Joshua Warcop]

I haven't moved forward yet, but not purely because of this but there are multiple things going at once. Yes I need to go ahead and check because I found other notes showing that it is just the troubleshooter may be having issues. I need to find the "next steps" document to test some of this out.

Yeah I read that bug ID earlier today when browing the tool and not sure it will present an issue. I'll add notifications to the bug.

Yes - both CUCM and CUPS are configured with addresses only with the "host.domain.com" portion and nothing pertaining to the server/protocol.

[Christos Georgiadis]

You may hit that defect for example when one of your CUP server goes down. Then when some users are on the phone the status might not change on their client from available to 'on the phone' because callmanager will ignore the second srv record presented and will continue to send the sip publish message that contains the 'on the phone' status to the node that is down.

That's just one scenario.

For the next steps you might want to check the CUP deployment guide

http://www.cisco.com/en/US/docs/voice_ip_comm/cups/8_0/english/install_upgrade/deployment/guide/dgcupc.html

HTH,

Christos

[Joshua Warcop]

The very thing (DNS SRV) to allow that seemless failover is the one thing that is broken.. Ahh gotta love bugs. Seems like this one should be fixed ASAP even though not very many people are using SRV records. Still everyone setting up with two trunks then? Or multiple destinations on one trunk - I think I see that option now on the CUCM trunk configuration page. CUCM 8.5(1).

[Christos Georgiadis]

yeah I remember spending quite some time on this before I find this bug. I never forgot it after that

Yes in CUCM 8.5 I think you can use multiple destinations on one truck and it should work.

Another trick although it requires a bit more time configuring stuff is this:

Use two A records in DNS resolving in both nodes ip addresses.

The caveat is that this dummy fqdn that you will create for those A records you need to put it under the sip proxy service parameters in CUPS. Here is the parameter

Server Name (supplemental)

otehrwise you end up in a loop since CUP doesn't know that this dummy fqdn is actully itself.

That's what I did for the phone status problem

Regards,

Christos