I recently added a front end server to my setup and it seems to have blown up my SIP trunk.  I had a working setup with 1 enterprise front end server with mediation server on it. We had a SIP trunk set up to out Cisco Call manager that worked fine.

I added a second front end server using a hardware load balancer (a CoyotePoint).  We didn't do any DNS load balancing, instead letting the HW LB do it. I added a new IP address for the pool that is on the HW LB and then it forwards to the 2 front end servers. I forwarded every port I could think of (currently, 80,135,443,44,4443,5060,5061,5062,5063,5065,5067,5069,5070,and 8080). Everything came up fine for everything except being able to make calls from Lync to the Cisco.  We updated the Cisco SIP Trunk to point to the new IP address and if I call from the Cisco into Lync, the call will pause for about 5 seconds but eventually go through (there was no pause before), so we know the Cisco can talk to Lync.  If I call from Lync out to Cisco, I end up with a "call was not completed or ended" error.  In the logging, it ends up with a "SIP/2.0 503 Service Unavailable" (the rest is below) error.  What is odd is that it always lists the .22 address in the warning, which is the original front end server then in the source it is always Us-LyncAP105 which is the new front end server.  Is there something I missed in setting up the HW LB? Or some other port I need to forward that isn't listed anywhere?  Or any ideas at all? Thanks.

TL_INFO(TF_PROTOCOL) [0]09C8.23C8::05/07/2013-19:37:46.897.0480bba9 (SIPStack,SIPAdminLog::TraceProtocolRecord:SIPAdminLog.cpp(125))$$begin_record
Trace-Correlation-Id: 1654056051
Instance-Id: 0A6248D4
Direction: incoming
Peer: Lync.x.com:5070
Message-Type: response
Start-Line: SIP/2.0 503 Service Unavailable
From: "Weide, Steve"<sip:s.weide@x.com>;tag=217cc318a9;epid=e0db4983ba
To: <sip:+2267@x.com;user=phone>;tag=34dd2b6822;epid=1B20D58716
CSeq: 1 INVITE
Call-ID: d5a55daff842439c9170cdf1f4ffb88f
VIA: SIP/2.0/TLS 10.x.20.5:49983;branch=z9hG4bKAFF913A7.DCE3AFB3C1A268DA;branched=FALSE,SIP/2.0/TLS 10.x.4.23:52174;ms-received-port=52174;ms-received-cid=16980D00
CONTENT-LENGTH: 0
P-ASSERTED-IDENTITY: <sip:2267;phone-context=DefaultProfile@x.com;user=phone>
WARNING: 399 MPCM2 "Unable to find a device handler for the request received on port 5060 from 10.x.20.22"
SERVER: RTCC/4.0.0.0 MediationServer
ms-diagnostics: 10503;source="US-LyncAP105.x.com";reason="Gateway responded with 503 Service Unavailable";component="MediationServer";SipResponseCode="503";SipResponseText="Service Unavailable";GatewayFqdn="10.x.10.10"
ms-diagnostics-public: 10503;reason="Gateway responded with 503 Service Unavailable";component="MediationServer";SipResponseCode="503";SipResponseText="Service Unavailable"
ms-trunking-peer: 10.x.10.10
ms-endpoint-location-data: NetworkScope;ms-media-location-type=intranet
Message-Body:
$$end_record

most likely your Cisco Server does not accept connections from the new server please recheck your Cisco config on the SIP Trunk

It's not that, as the it returns the message "WARNING: 399 MPCM2 "Unable to find a device handler for the request received on port 5060 from 10.x.20.22"in the log (MPCM2 is our Cisco Call Manager).  And there is only 1 place in the Cisco to configure the SIP trunk and that is using the new load balanced address (and it obviously is correct since calls from the Cisco can get to Lync).  It seems like it is something with the load balancer, as it only happens when a call orginates in Lync. 

All traffic from the front ends is routed through the load balancer (so that everything should appear to be coming from the load balanced IP) and I think that is where the issue lies, but I'm not sure exactly what the issue is.

I have an issue close to this one,

when i added a new front end 2013 server to my pool, the integration stopped with my CUCM.

It only worked again when i stopped the mediation service on the newly created front end server

Any Ideas?

I was having and event: all gateways are down

Thanks

Hi,

Please have a test use the DNS load balance instead of the Hardware load balance then check the issue again.

Hi everybody...

Any news about this error?

Hi, 

I'm trying to add new Lync FE IP Address to CUCM SIP Trunk as below but still no luck

have anyone here try using Pool FQDN instead of IP (with check in Destination Address is an SRV) ?

Thanks

Endrik

Hi, 

you need to add your Mediation pool name into CUCM SIP Trunk, as following:

don't forget to thick "Destination Address is an SRV", 

it's work in my case.

Thanks,

Endrik

Hi 

SIP/2.0 503  means the service is not unavailable

You can check the Lync logging file to compare the failed call and the other successful call to find out what the differences.

I doubt that might be a port or route issue,would you please check if the protocols/ports especially the TCP 443,UDP3478,TCP/UDP 50,000-59,999 port ranges are opened

I would suggest you to verify firewall. Probably you could run a port query tool.

HI Steve,

Has you said you have changed the Ip Address of the Sip Trunk. I could see when you call from lync to cisco it is still looking for old Ip address may be. can you check if the new trunk ip address is updated in trunk and PSTN Gateway.also check if which ip address is marked as Default Sip trunk.

and also check if in Dial plan if it is pointing to the old SIP Gateway.

Hi Steve, 

were you able to sort  this issue. I would like to ask about the Mediation Server as we didnt discuss much on that. Do you have a dedicated Mediation server or you have a collocated one. 

Could you enable the Lync Server tracing for mediation server selected and collect  the trace  to verify which mediaion  server the call is heading to etc..