2. Configure Virtual Machine High Availability
Now that you have configured power management on your hosts, they are ready to run highly available virtual machines. High availability means that a virtual machine will automatically be restarted if its process is interrupted. This interruption occurs when a virtual machine is terminated by methods other than shutting down within the guest or from the administration portal. When these events occur, the highly available virtual machine will be automatically restarted, either on its original host or another host in the cluster.
High availability can only be configured for virtual servers, not virtual desktops.
This lab uses six virtual machines, and configures three of them to be highly available. You can use different names, or a different number of virtual machines - just be aware that this lab uses examples which consistently refer to these allocated names. The virtual machines used in this lab are:
Highly Available Virtual Machines: RHEL6Nile, RHEL6RioGrande, RHEL6Thames
Non-Highly Available Virtual Machines: RHEL6Congo, RHEL6Erie, RHEL6Ganges
To configure High Availability for a virtual machine
On the Tree pane, click the VMs icon. On the Virtual Machines tab, select the virtual machine you wish to mark as highly available. This example uses RHEL6Thames
. Click Edit.
The Edit Server Virtual Machine dialog displays. Select the High Availability tab and tick the Highly Available checkbox. Change the Priority for Run/ Migrate Queue to High. This means that this virtual machine will take precedence in the queue of virtual machines to be migrated.
Click OK to save your settings. Repeat this procedure for other virtual machines you wish to mark as highly available, in this example they are RHEL6Nile
and RHEL6RioGrande
.
Now, you have three virtual machines which are highly available, and three which are not. Hold down the Shift
key and click the virtual machines to select all of them, then click the Run icon to start them all. The virtual machines may not be evenly distributed across the hosts, because one of them is the allocated Storage Pool Manager (SPM) which manages access between hosts and storage. The SPM's CPU resources are utilized more heavily than that of other hosts. Therefore, when virtual machines are started, the host with a lighter workload is selected to run the machines.
On the Tree pane, click Hosts to display the available hosts and their workloads on the Hosts tab.
You have now successfully configured power management for your hosts, and set several of your virtual machines to be highly available. You can now experiment with virtual machine high availability in four scenarios that may happen in an enterprise data center — host manual reboot (user error), virtual machine crash (unexpected fault), partial failure (non-operational) and host disconnection full failure (non-responsive).