Proper Configuration for Call Overflow

[jbyrne000]

We recently changed datacenters for our Primary site and we are now having trouble with calls overflowing from our primary to our DR site when our VoIP server is down -- this worked previously without issue at the old datacenter and on the old circuits without any changes to our Router configuration. Now calls are failing to fast busy. Our carrier is saying that because they have a valid SIP connection to the Router, the overflow will not be initiated.

We run straight SIP from our VoIP server -> Cisco Catalyst 3750 Switch (no SIP manipulation) -> Cisco ASR 1001 (CUBE) Router -> Carrier SIP trunks. When our VoIP server is down, we want calls to be directed by the carrier to our DR site (also straight SIP with the same/similar call flow), and then if the DR site fails we want calls directed to a third backup phone number we have configured with an outage message.  Currently, calls just fail to fast busy.  The carrier is trying to tell us we need to disable our interface on the Router to the VoIP server -- this would mean either disabling the connection to the carrier (our only data connection in to the Router), or disabling the connection to the Cisco Catalyst 3750 switch and therefore our entire datacenter network beyond the ASR 1001 router.  

What is best practice in configuring overflows? Is it reasonable for the carrier to ask us to shut down an entire interface if our VoIP server crashes? How should call overflows be initiated/configured from the CUBE/Carrier?

Thanks for any guidance you can provide!

Best regards,

Jack

[Mohammed al Baqari · Accepted solution]

On the ASR CUBE you need to enable SIP Options ping on the dialpeer pointing to your SIP server. This will put the dialpeer in busy-out status when the voip server is down. CUBE will return error message to your provider (most probably 404 Not Found because it can't locate active dialpeer). Your carrier should be able to route the call to DR site once this message is received. Same thing to be done on DR site.

[Dennis Mink · Accepted solution]

If you can't get your provider to anything semi-intelligently like set up redundancy, keep pushing them, cos what they are saying, although technically correct, is rubbish.

in the meanwhile, create a second dial peer on your CUBE with a higher preference to your DR site for when your primary VOIP server is down. And Like Mo said below, use OOB ping to track the status of your dial peers for automated fail over.  Its a very elegant redundancy, but it will work

[jbyrne000]

Thanks Mohammed and Dennis!