Feature Request Overview

The user goal is to control the placement of migrated VMware virtual machines (VMs) onto specific OpenStack compute nodes using the Red Hat OpenStack VMware Migration Toolkit.

Currently, after flavor selection/creation, the VM could be scheduled onto any eligible compute host. This feature is needed to ensure that workloads requiring specific attributes—such as dedicated hardware, specific VLANs, or licensing restrictions—are reliably placed on compute nodes that possess those attributes, which are managed within OpenStack using Host Aggregates.

Business justification

This feature provides significant benefits by allowing customers to:

Enforce Compliance and Resource Separation: Customers can ensure that workloads with strict requirements (e.g., regulatory compliance, specific networking needs like certain VLANs, or specialized hardware like GPUs or SSD-only storage) are placed only on the compute hosts configured to meet those needs.

Optimize Resource Utilization: By isolating specialized or licensed hardware/resources into specific host aggregates, customers can prevent those resources from being consumed by general-purpose workloads, ensuring the resources are available for the critical workloads that require them.

Improve Performance Predictability: Workloads requiring dedicated, high-performance, or specialized hardware (e.g., a specific CPU architecture or high-end NICs) can be reliably placed in aggregates containing only that hardware, leading to more predictable performance.

Leverage Existing OpenStack Functionality: It allows customers to fully leverage the established Host Aggregates and associated Availability Zone functionality within OpenStack to manage and segment their compute resources effectively during the migration process.

Functional requirements

Host Aggregate Selection Input: The Red Hat OpenStack VMware Migration Toolkit must provide an option (e.g., a field, a dropdown list, or a configuration parameter) for the user to specify the target Host Aggregate or Availability Zone during the migration setup process.

Flavor Integration with Host Aggregates:

Existing Flavor Match: If use_existing_flavor is true, the best match algorithm must not only find the closest existing flavor but also check if that flavor is associated with the selected target Host Aggregate (via its metadata/properties). If the flavor is not compatible or associated, the user must be informed, and the process should offer to create a new, compatible flavor (Requirement 3).

New Flavor Creation: If a new flavor is being created (either because use_existing_flavor is false or no suitable existing flavor was found), the new flavor must be automatically created with the appropriate flavor extra-specs that enforce scheduling onto the specified target Host Aggregate.

Validation and Error Handling: The tool must validate that the specified target Host Aggregate exists and is available. If an invalid or unavailable aggregate is selected, the migration process must be blocked, and a clear error message must be presented to the user.

Placement During Migration: Upon final execution, the migrated OpenStack VM instance must be scheduled and booted only on a compute host that belongs to the specified target Host Aggregate.

Describe the customer impact

This feature has a high positive impact on customers who require fine-grained control over VM placement. It directly addresses the need to guarantee that migrated workloads land on compute nodes with specific, requisite hardware (e.g., dedicated GPUs, high-speed storage, or specific CPU features) or network configurations (e.g., specific VLAN connectivity). This control is critical for maintaining licensing compliance and performance Service Level Agreements (SLAs) for specialized workloads post-migration.

(Optional) Point of contact

• Carlos Franciosi 
•  

(Optional) Additional links

Click More > Link to add any links to issues, such as an outcome, that are related to this feature request.