Red Hat developed RHEV as a complete platform that includes advanced features, such as high availability, and does...
not require additional components. However, to enable high availability, you'll need to plan ahead and configure power management as you architect an environment for production.
The fundamentals of configuring RHEV high availability
Every Red Hat Enterprise Virtualization (RHEV) environment should have shared storage available to all hosts within the data center. By default, every host has access to associated storage domains. If hosts cannot access the same storage, it won't be possible for virtual machines (VMs) to failover to another host.
Each RHEV installment also needs at least one cluster, regardless of the number of nodes. The cluster consists of a group of hosts you can configure for high availability. RHEV automatically configures hosts with the same network connections, which is required for high availability.
No power management capabilities? No high availability
RHEV also includes a capability called power management, which has nothing to do with managing the energy consumption level on host computers. Instead it dictates how RHEV manages power-on cluster hosts. Power management is a vital part of configuring RHEV high availability because the cluster must be able to shut down a failing host before you can restart the VMs somewhere else.
More resources on configuring RHEV
The Internet Explorer requirement for RHEV 3.0
Connecting shared storage to RHEV in three easy steps
Red Hat Enterprise Linux virtualization architecture
It may seem far-fetched, but without power management two hosts could think the other is unavailable and, as a result, both will try to run the same VM. This scenario would corrupt services on all VMs involved, which is why RHEV must shut down before another host takes over its resources.
To set up power management, you must have supported hardware. Typically, this would be an integrated management board such as HP iLO or Dell DRAC, but you could also use a power distribution unit (PDU). A PDU allows you to shut off a specific port, which stops the server connected to that port. You use the same method to set up power management on either hardware.
After installing the required hardware, you need to configure power management for each host. Selecting the Enable Power Management link from the RHEV-M management console will display all hosts. Under power management properties, include the IP address, user name and password for the power management device. Depending on the device type selected, other options might appear as well. To configure a PDU called APC MasterSwitch, for example, you must specify the port and slot number as well.
After you configure power management for each host, the high-availability cluster becomes available, which means VMs will automatically migrate from a failing host to a remaining host if the cluster encounters a problem.
Enabling power management allows you to easily configure a RHEV environment for high availability. Power management guarantees the integrity of connected VMs; it also forces you to decide which PDU you'll use before purchasing servers for a virtual infrastructure.