Product SiteDocumentation Site

Lab 6 - High Availability Scenarios

The Red Hat Enterprise Virtualization Manager offers various high availability features which can be applied in a granular manner, from the level of a single virtual machine up to protection against multiple host failure scenarios. In addition, you can protect your virtual machines against various failures by combining virtual machine high availability with out of band power management, Red Hat Enterprise Virtualization Manager's failure detection and failure recovery solutions.
This lab enables you to configure virtual machine high availability, and demonstrates its use in several common enterprise scenarios. It is assumed that you have successfully completed the basic labs of Section 1, “Track A: Standard Setup”, meaning that you have correctly installed and configured Red Hat Enterprise Virtualization, and have several running virtual machines.
In addition, to successfully complete this lab you must have a power management card for each of your hosts. This lab uses the Intelligent Power Management Interface (IPMI) device as an example. If you have a different device see the Red Hat Enterprise Virtualization Administration Guide.
Lab 6 - Objectives
This lab takes you through the best practices of configuring and testing a reliable Red Hat Enterprise Virtualization environment where virtual machines running critical workloads are not easily interrupted. This lab should take you about 30 minutes.
Section 1, “Configure Power Management” shows you how to enable power management for your hosts. (3 minutes)
Section 2, “Configure Virtual Machine High Availability” shows you how to mark a virtual machine as highly available so that it automatically restarts after being unexpectedly terminated. (2 minutes)
Section 3, “High Availability - Host Initiated Reboot” demonstrates virtual machine high availability when a host is manually fenced and rebooted before the virtual machines running on it are properly shut down. (3 minutes)
Section 4, “High Availability - Virtual Machine Interruption” demonstrates how a highly available virtual machine will automatically restart on the same host if it experiences a crash. (6 minutes)
Section 5, “High Availability - Non-Operational Host” demonstrates virtual machine high availability when a host becomes non-operational due to a storage disconnection. (8 minutes)
Section 6, “High Availability - Non-Responsive Host” demonstrates virtual machine high availability when a host has lost communication with the Red Hat Enterprise Virtualization Manager. (8 minutes)
Lab 6 - Requirements
Before you attempt this lab, you must have a working Red Hat Enterprise Virtualization environment. The requirements for setting it up are stipulated in Section 1.1, “Track A Requirements”. To successfully complete this lab, you must have:
  • An out of band power management device for each host. This lab uses the Intelligent Power Management Interface (IPMI) device.
  • Access to the administration portal and access to your hosts (for simulating failure scenarios). You will be instructed on how to enable remote login for your hypervisors later in this lab.

1. Configure Power Management

At this point, you should have two hosts and at least two virtual machines, however this lab uses six virtual machines. You can use as many virtual machines as you want, but for optimal demonstration of high availability features, it is recommended that you add four new virtual machines to your environment.
Highly available virtual machine settings will only be effective if power management is enabled on the hosts. However before configuring power management, recall that you have previously defined a power saving cluster policy in Lab 3 - Live Migration Scenarios. Cluster policies and high availability can be used concurrently; however to best demonstrate Red Hat Enterprise Virtualization's high availability features for this lab, reset the cluster policy. This ensures that the virtual machines remain where they are staged before each demonstration, so the migration is triggered by high availability rather than the cluster's load balancing policies.
To disable cluster policy
  1. On the administration portal, navigate to the Tree pane, click the Expand All button, and click the Default cluster. The Cluster tab displays, select the Default cluster to display its details pane.
  2. On the General subtab, you can see that the policy is set to Power Saving. Click Edit Policy.
    Edit cluster policy
    Figure 42. Edit cluster policy

  3. The Edit Policy dialog displays. Select the None button to remove the previously configured policy. Click OK.
Now, you can configure power management on your hosts. Power management enables the system to fence a troublesome host using an additional interface such as an Intelligent Platform Management Interface (IPMI) device. Perform this procedure for each host.
To set up power management on a host
  1. On the Tree pane, click the Hosts icon under the Default cluster. The Hosts tab displays a list of available hosts.
  2. Select a host, in this example the Atlantic host is used. Notice that there is an exclamation mark next to the hostname which you were asked to ignore in previous labs. Click the Edit button to display the Edit Host dialog.
  3. Select the Power Management tab. Tick the Enable Power Management checkbox and provide the required information in the following fields:
    • Address: The address of the power management card. This card does not have to be on any of the logical networks defined on the Red Hat Enterprise Virtualization Manager. However, all hosts in the cluster must have a valid route to the card's IP. This ensures that fencing is correctly configured, as fencing is done by any available host in the cluster.
    • User Name: The user allowed to log in to the power management device.
    • Password: The password of the user allowed to log in.
    • Type: The type of management device. For this lab, select ipmilan.
    • Options: These additional parameters depend on the specific implementation of each device. Detailed documentation of the options available is provided in the man page for each fence agent. For this lab, enter power_wait=4,lanplus=yes in the provided textbox.
    Click the Test button to test the settings. If the power management options can be verified, the text Test Succeeded, Host Status is: on displays.
    Edit cluster policy
    Figure 43. Edit cluster policy

  4. Click OK. You are returned to the list of hosts. Note that the exclamation mark next to the host's name has now disappeared, signifying that power management has been successfully configured.
You have now configured power management for your hosts, meaning that your hosts' power status can be verified and controlled by the Red Hat Enterprise Virtualization Manager. Power management checks that a host is properly powered down, then restarts its virtual machines on another host in the same cluster. However, if the host's status cannot be verified, the virtual machines that were originally running on it will not be restarted.
Next, configure high availability for several virtual machines in the system.