Dial peer problem in call manager failover

srosenthal · ‎04-01-2010

I have a two server call manager cluster. I have my router with PRI setup with two dialpeer's, the first one sends calls to the publisher and the second one sends calls to the subscriber.

This morning we took the publisher down to test failover to the subscriber. When we did that, inbound calls did not work. If I shut down the dial-peer that points to the publisher, then inbound calls worked.

Is there some way of forcing this automatically?

Seth

Aaron Harrison · ‎04-01-2010

Hi

On your gateway:

Voice class h323 1

h225 timeout tcp establish 5

Then on each dial-peer:

Voice class h232 1

Reduces the long TCP timeout to something more appropriate...

Aaron

Aaron Please remember to rate helpful posts to identify useful responses, and mark 'Answered' if appropriate!

William Bell · ‎04-01-2010

Seth,

Could you possibly post the relevant configurations from you voice gateway?

A sample dial-peer config that I know works:

dial-peer voice 2001 voip

description Connection to Primary CM

preference 1

destination-pattern 202555....

session target ipv4:10.3.4.21

dtmf-relay h245-alphanumeric

codec g711ulaw

no vad

!

dial-peer voice 2002 voip

description Connection to Secondary CM

preference 2

destination-pattern 202555....

session target ipv4:10.3.4.20

dtmf-relay h245-alphanumeric

codec g711ulaw

no vad

!

The key is that you want to make sure that the destination-patterns are the same. If the dial-peer to your publisher was more specific and matched the ingress calls, it would always be preferred which would lead to a situation like the one you describe.

You could also look at tweaking the TCP setup timers (using "h225 timeout tcp establish" command) though if you say the calls are completely failing, look at your dial-peer destination-patterns first.

HTH.

Regards,

Bill

Please remember to rate helpful posts.

HTH -Bill (b) http://ucguerrilla.com (t) @ucguerrilla

Please remember to rate helpful responses and identify