I'm not sure I follow what you're asking... But it sounds like you're misunderstanding how VCMP works.
At it's simplest, VCMP is simply a linux server (The Hypervisor - I think that's what you're referring to as the 'host') running kvm. Each kvm (Pretty sure that's what I think you're calling a guest) is running a BigIP instance
Where I think you're misinterpreting this is
- There is no concept of active/standby at the hypervisor level. A hypervisor is always active. It is (**usually) always running the VCMP instances/guests that are configured on it
- HA at the BigIP guest level is performed by DeviceGroups. A DeviceGroup has 1 or more virtual servers in it. A DeviceGroup is active on 1 of the BigIP's in the 'group' of BigIP Devices (In your case a VCMP guest/instance but may be a standalone appliance or a virtual big instance) that it is associated with. A DeviceGroup may be sync-only or sync-failover.
Failover is performed at the DeviceGroup level. You decide which devices a device group is associate with. And they can have priorities. You can select which device the device group is active on. Each Device may have multiple devicegroups. This allows you to run as active/standby or active/active, or active/active/active/active... (Depending on how many devices in your cluster)
H