High availability management is a critical component of business continuity planning. Any downtime to the software stack can result in loss of revenue and disruption of services—making the speed and consistency of recovery a priority for many organizations running IBM Power environments.

Disaster recovery planning often focuses on two metrics: Recovery Time Objective (RTO) and Recovery Point Objective (RPO). Organizations define targets, design architectures to meet them, and test periodically to validate.

But there’s a third dimension that sometimes gets less attention: how much human involvement the recovery process requires.

This blog explores VM Recovery Manager (VMRM) and how its infrastructure-level orchestration model is designed to change the operational profile of DR on IBM Power—with the goal of reducing the manual steps, specialist dependency, and execution variability that can affect even well-designed recovery architectures.

The Challenge of Manual DR Orchestration

Traditional disaster recovery processes often depend heavily on documented runbooks and human execution. When a failover or DR rehearsal is triggered, the sequence typically involves:

This process can work well when it happens infrequently and when the right people are available. In practice, it can introduce variability: steps may be missed under pressure, execution times can vary between operators, and the process may be difficult to rehearse frequently enough to keep skills current. For organizations that need fast, consistent recovery, manual orchestration is a factor worth examining.

What VM Recovery Manager Is Designed to Do

IBM VM Recovery Manager HA for Power Systems is designed as a high availability solution that aims to be straightforward to deploy and provides an automated approach to recovering virtual machines (VMs), also known as logical partitions (LPARs).

VMRM is designed to operate at the infrastructure level, outside of individual LPARs. This means it can be configured to detect events—hardware failures, site-level outages, or planned maintenance triggers—and initiate recovery actions with reduced reliance on manual steps.

VM Restart Technology: A Different Approach

The VMRM solution implements recovery of virtual machines based on VM restart technology. This approach relies on an out-of-band monitoring and management component that is designed to restart VMs on another server when the host infrastructure experiences a failure.

This architecture is different from conventional cluster-based technology, which deploys redundant hardware and software components for near real-time failover when a component fails. Because VMRM operates outside the LPAR, it may cover a broad range of workloads without requiring changes to individual applications or the complexity associated with traditional clustering approaches.

In terms of architecture, a set of hosts is grouped to serve as backup for each other. When failures are detected, VMs are designed to be relocated and restarted on other healthy hosts within the group—without requiring manual intervention at the time of the event.

Infrastructure-Level Event Detection

VMRM is designed to monitor the environment at the hypervisor level. Rather than waiting for an alert to reach an on-call administrator, it is intended to detect conditions that may require action and begin orchestrated recovery based on defined policies.

Automated VM Restart and Relocation

When recovery is triggered, VMRM is designed to support restarting VMs on alternate hardware or relocating them to a secondary site based on defined policies. The orchestration is defined in advance, not improvised at the time of the event.

Site-Level Recovery Orchestration

For multi-site environments, VMRM is designed to support coordination of the recovery sequence across sites. This includes managing dependencies between workloads and handling the sequencing that would otherwise require manual coordination.

Broad Workload Coverage

Because VMRM operates at the infrastructure layer, it is designed to cover workloads running in the virtualized environment without requiring changes to individual applications.

Designed for Simpler Management

One consideration that often comes up when evaluating HA/DR solutions is operational complexity. Traditional cluster-based approaches can introduce clustering complexities—configuration, tuning, and specialist knowledge requirements that accumulate over time.

The VMRM solution is designed to be easier to manage because it does not carry those clustering complexities. For organizations looking to extend high availability coverage across a large number of VMs without adding significant operational overhead, this is a design consideration worth exploring.

DR Rehearsal as a More Routine Activity

One of the operational considerations with VMRM is how it may affect DR rehearsal frequency. When recovery orchestration is manual, rehearsals can be disruptive, time-consuming, and difficult to schedule. The operational cost of a rehearsal may be high enough that organizations test infrequently.

With automated orchestration, the rehearsal process is designed to follow the same automated sequence as an actual recovery event. This may make rehearsals faster to execute and easier to repeat. Teams may be able to validate their recovery posture more frequently—and regular rehearsal is one way to surface misconfigurations, policy gaps, or sequencing issues before they matter.

Where VMRM Fits in the IBM Power HA/DR Landscape

VMRM is not a replacement for PowerHA SystemMirror. The two solutions are designed to address different recovery scenarios.

PowerHA provides full-stack high availability with tight integration between the cluster and the application layer—appropriate for environments where fine-grained control and continuous availability of specific workloads is a priority.

VMRM provides infrastructure-level recovery orchestration that is designed to operate independently of the application layer—appropriate for environments where automated VM restart or site-level failover is the focus, and where simpler management is a consideration.

Many organizations use both: PowerHA for their most critical workloads, VMRM for broader infrastructure coverage, or VMRM DR alongside PowerHA for site-level orchestration.

VMRM DR is included with AIX Enterprise Edition, making it available to many organizations already running AIX without additional licensing complexity.

👉 Learn more about VM Recovery Manager.

#IBMPower #VMRecoveryManager #DisasterRecovery #highavailability #PowerHA #DR #AIX