Virtual Machine and Host High Availability - Getting Started with Red Hat Enterprise Virtualization - page 112

Manual fencing
If the hypervisor host without the power management configured in a cluster or data center
went to the Non Responsive state due to some hardware failure, it will significantly affect
the overall performance of your virtual infrastructure, thereby causing all the virtual ma-
chines running on the host to change to the Unknown status. In such cases, we need to
manually fence or isolate the nonresponsive host by manually rebooting the host or power-
ing it down to recover all the virtual machines. Once the host is rebooted or powered down
manually, perform the following steps:
1. Log in to the Administrator console, and navigate to the Host tab.
2. Select the Non responsive host, right-click on it, and select the Confirm Host has
been rebooted button.
3. A warning window will pop up to ensure that the host has been shut down or re-
booted; select Approve Operation , and click on OK .
4. Post this, all the virtual machines marked highly available that were running on the
failed host will turn to the power-down state and allow us to get restarted on other
active hosts in the cluster.
Note
Do not use the Confirm host has been rebooted option unless you have manually
rebooted the host. Using this option while the host is still running can lead to a vir-
tual machine image corruption. For more information on the power management
configuration, please refer to Red Hat Enterprise Virtualization Admin Guide.