HVM Clusters¶
HVM clusters provide KVM-based virtualization, centralized lifecycle management, workload placement, shared storage, monitoring, migration, and high availability. This guide covers every supported HVM cluster layout. Use the layout matrix to identify the operating system, networking model, and management tools that apply to a cluster.
Supported Layouts¶
Layout |
Host operating system |
High availability |
Host networking |
Host management |
|---|---|---|---|---|
Legacy (1.2 and earlier) |
Ubuntu 22.04 |
Pacemaker and Corosync |
Open vSwitch (OVS) |
Standard Linux and |
1.3 |
HVM OS/Ubuntu 24.04 |
HPE Morpheus Software Agent quorum, Corosync, and DLM |
Open vSwitch (OVS) |
Standard Linux commands |
2.0 |
HVM OS 26.04 |
HPE Morpheus Software Agent quorum, Corosync, and DLM |
Virtual Switches (Linux bridge) |
|
Important
Layout and HVM OS versions are separate from the HPE Morpheus Software Manager version. Confirm the cluster layout on the cluster detail page before following a layout-specific procedure.
How to Use This Guide¶
Unless a section contains a layout notice, it applies to layouts 1.3 and 2.0. Legacy clusters remain supported, but their Pacemaker, Ceph, provisioning, and OVS differences are collected in Legacy HVM Clusters to keep current workflows singular.
For OVS networking on Legacy and layout 1.3 clusters, see HVM Networks.
For layout 2.0 networking, see Virtual Switches.
For layout 2.0 host commands, see HVMCLI.
For NVIDIA SR-IOV vGPU profiles, Host preparation, Service Plans, and guest licensing, see NVIDIA vGPU Slicing.
For datastore creation, Datastore Groups, file browsing, maintenance, and capacity operations, see Storage Lifecycle.
For VM snapshots and creating linked clone Virtual Images, see Snapshots.
For moving VMs between hosts or HVM clusters (Actions > Move), see VM Migration & Failover.
For layout and HVM OS upgrades, start at HVM Clusters After Manager Upgrade, then use Upgrading Clusters for detailed cluster procedures.
For two-Host GFS2 clusters that use a Distributed Worker witness and no Site Groups, see Two-Node Clusters with a Witness.
For multi-site stretch clusters that use Site Groups and a
siteWitness, see Stretch Clusters & Witness Nodes.For the quorum vote model and Quorum panel differences between those topologies, see Architecture.