Callmanger 6.1 Network Failover

[brown.paul]

I am trying to understand how the Callmanager fails over between NIC cards.

I have a Publisher/Subscriber; each CCM has a connection to switch A and to switch B.

When does Eth1 take over? Does Eth0 need to lose physical connection and go down hard or is it based on replies from its configured GW?

[William Bell · Accepted solution]

It is based on link loss.  So, the port/interface/cable/etc. would need to experience a fault in some way.

HTH.

Regards,
Bill

[brown.paul]

So that explains why the failover didn't happen. My Nexus didn't go down hard  and lingered in a altered state.

Any suggestions when that condition occurs?

[James Hawkins]

I think you will need to provide more information on your network setup to get nany replies.

One possible approach - if the CUCM server NIC were plugged into an IOS switch Ethernet port and you wanted to force the port to shut down if the switch became isolated from the network you could configure Embedded Event Manager to monitor an upstream IP host and, if it loses connectivity to shut the Ethernet port connecting the CUCM. I am not sure whether Nexus support EEM though.

[William Bell]

Well, from the CUCM side you can manually disable eth0 and force a failover:

admin:set network status eth0 state down

I suppose you could script this if you can auto-detect the fault condition (with minimal risk of a false positive) and were able to launch a script.

HTH.

Regards,
Bill

Please remember to rate helpful posts.

[brown.paul]

Thanks guys. I will play with those ideas