Redundant VOIP dial peers on H323 gateways.

Unanswered Question

Hi, I have a CM 5.1 system with 2 CM's in the cluster. I have 22 voice gateways all configured as H323. I have the following dial peers configured in some variation for incoming calls on each gateway:

dial-peer voice 44 voip

preference 1

destination-pattern 3205

voice-class codec 1

session target ipv4:

dtmf-relay h245-alphanumeric

no vad


dial-peer voice 45 voip

preference 2

destination-pattern 3205

voice-class codec 1

session target ipv4:

dtmf-relay h245-alphanumeric

no vad is my publisher and is my subscriber. Everything works just fine in this configuration when all my servers are up.

The problem: My publisher went down the other day ( and all incoming calls got a busy signal. All the phones were registered and could make outgoing calls just fine.

My Question: Do I have this backup dial peer for the subscriber configured correctly? Is there something in CM I need to enable for this backup dial peer to work properly? For whatever reason in this configuration calls would not get forwarded to my subscriber.

Thanks for your help.

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)

I am looking for people's input on WAAS Printing services...pros and cons discovered.

Has anyone encountered any problems or issues after configuration and implementation?

I am looking at implementing WAAS Printing (WAE-512/612's) too but want to know how reliable it is.

Do you happen to know how many printers it can service/support? I understand it only has 1 GB Spool limitation and wondered what would happen if many people began sending print jobs larger than that to it...would it crash?



tzunt Sun, 03/09/2008 - 21:19


Your redundant VOIP dial-peers to each CM look good except you also need to create a H323 voice class like the following and associate it with your dial-peers.

from Cisco gateway book:

NYCgateway(config)#voice class h323 10

NYCgateway(configs-class)#h225 timeout tcp establish 3

Then add the following to both the dial peers you listed so they associate with the new class:

NYCgateway(config-dial-peer)voice-class h323 10

The calls were not failing over to your SUB in the inital configuration because the default for this to occur is 15 seconds. (I think.) If your calls arrived on PRI, the ISDN setup would timeout before 15 seconds. The new class will shorten the timer to 3 seconds which will provide the redundancy you wish.



I tried that, but it did not seem to fix it. What I did try was removing my primary dial-peer so all I had left was one dial-peer pointing to my subscriber, and the call failed as well.

I almost seems like im missing a setting in CM, since my subscriber will not accept calls. The debug cause says user busy.

Is there something else I need to enable. I saomething about a Run H.225D On Every Node flag but could not find it anywhere.

mattcalderon Tue, 03/11/2008 - 18:44

With both dial-peers on your gateway, have you ran a debug on your voip dial-peer to verify that you are indeed hitting your dial-peer to your subscriber? You gateway may still be sending them to your publisher.

Run a debug voip dialpeer all

and see exactly what dialpeer you are using.

Here is a link to what the previous poster had said about the shorter time out value.

hmm...weird when you only have the one to your subscriber and the call fails. Stupid question probably, but have you verified that your ccm service is running on your sub?

I did run the dialpeer debug and verified that it was in fact hiting the correct dialpeer. I even changed the session target to my publisher, and it worked right away.

I did verify that the Call Manager service on the subscriber was in fact started and running. I didn't want to restart the service right now since there are about 300 phones in the cluster and they are all registered to my subscriber


This Discussion