I'm preparing the heartbeat installation for a customer, but they have multiple networks on the vCenter server:
- Production network (for vSphere client connections): 172.x
- Backend network (ESXi servers): 192.x
- 2 x Channel network: 10.x
The vCenter managed IP is from the backend network (192.x), and the services are bind to this network. With this knowledge you can say: this is the "production" network for heartbeat. But from client/administrator perspective the production vlan (172.x) is the production network, and the production IP should be reconfigured on the passive server after a failover, because we don't want DNS reconfiguration. But the backend IP should be moved as well.
Is this possible with heartbeat? With Neverfail it is, but I'm not sure if 2 failover IP's is supported and possible to configure. I'm looking for the supported configuration, no fancy workarounds 😉
Otherwise, if it's not possible/supported, we can solve it with the loadbalancer:
- VIP from loadbalancer to active vCenter (healthcheck on vCenter Web service)
But this isn't my preferred option...