VM Recovery Manager HA concepts

The VM Recovery Manager HA solution uses the following concepts:

Controller system (KSYS)

The controlling system, also called KSYS, is a fundamental component that monitors the production environment for any unplanned outage. If an unplanned outage occurs, the KSYS analyzes the situation, notifies the administrator about the failure, and can automatically move the failed virtual machines to another host in the host group. The KSYS interacts with the Hardware Management Console (HMC) to collect configuration information of managed systems. The KSYS subsystem also collects VIOS health information through the HMC.

The KSYS subsystem runs in an AIX® logical partition (LPAR). You can customize the security level for the KSYS LPAR according to the AIX security requirements of your organization. In addition, the KSYS LPAR can be protected for failure by using other products such as PowerHA® SystemMirror® for AIX. The KSYS subsystem must remain operational even if the site fails. Ensure that you periodically receive KSYS health reports. You can also check the KSYS subsystem health in the VM Recovery Manager HA GUI dashboard.

Host group

Hosts are grouped together to be backup for each other. When failures in any of the hosts are detected, VMs in the failed host are relocated and restarted on other healthy hosts within the group of hosts. This group of hosts is called a host group.

Host monitor

The host monitor daemon is shipped with the Virtual I/O Server (VIOS) and is deployed during the VIOS installation. When you initialize the KSYS subsystem for high-availability feature, the host monitor module becomes active. The KSYS subsystem communicates with the host monitor daemon through the HMC to monitor the hosts for high availability. For information about the VIOS version that contains the host monitor daemon, see the Requirements section.

VM agent

You can optionally install the VM agent filesets, which are shipped along with the KSYS filesets, in the guest virtual machines. The VM agent subsystem provides high-availability feature at the VM and application level. The VM agent monitors the following issues in the production environment:

• VM failures: If the operating system of a VM is not working correctly, or if the VM has stopped working because of an error, the VM is restarted on another host within the host group. The KSYS subsystem uses the VM monitor module to monitor the heartbeat from the VM to the host monitor subsystem in a VIOS.
• Application failures: Optionally, you can register the applications in the VM agent to enable application monitoring. The VM agent uses the Application HA monitoring framework to monitor the health of the application periodically by running the application-specific monitor scripts, by identifying whether the application has failed, and by identifying whether the VM must be restarted in the same host or another host. This framework can also manage the sequence in which applications are started and stopped within a VM.

Note: The VM agent is supported on AIX and Linux® (RHEL and SLES) guest VMs only. Currently, the VM agent subsystem is not supported for the IBM® i and Ubuntu VMs. Therefore, IBM i and Ubuntu VMs are relocated from one host to another host within the host group only after a host failure.

The following figure shows the detailed architecture of the VM Recovery Manager HA solution: