Legacy HVM Clusters

This appendix covers HVM layout 1.2 and earlier. Legacy clusters remain supported, but their provisioning, high-availability, storage, and networking architecture differs from layouts 1.3 and 2.0. For operations that are not listed as Legacy-specific here, use the shared HVM cluster guide.

Important

Confirm the cluster layout on the cluster detail page before using this appendix. Do not use the Pacemaker or Ceph procedures on layouts 1.3 or 2.0.

Legacy Differences

Area

Legacy layout

Current layouts

Host operating system

Ubuntu 22.04

HVM OS/Ubuntu 24.04 for layout 1.3; HVM OS 26.04 for layout 2.0

High availability

Pacemaker and Corosync

HPE Morpheus Software Agent quorum, Corosync, and DLM

Cluster diagnostics

pcs and standard Linux commands

Agent quorum endpoint, corosync-quorumtool, and dlm_tool

Host networking

Open vSwitch (OVS)

OVS on layout 1.3; Virtual Switches on layout 2.0

Converged storage

Ceph on HCI layouts

HPE Clustered Datastores and supported external storage

Host management CLI

Standard Linux, virsh, and pcs commands

Standard Linux commands on layout 1.3; hvmcli on layout 2.0

Legacy Requirements

  • Operating System: Ubuntu 22.04

  • CPU: One or more 64-bit x86 CPUs with Intel VT or AMD-V enabled

  • Memory: 4 GB minimum, plus 4 GB per Ceph disk for converged layouts

  • Storage: A data disk of at least 500 GB for test HCI deployments; production requires additional capacity

  • Networking: Static Host IP addresses, DNS resolution of the HPE Morpheus Software appliance, and access to package repositories

  • Cluster Size: At least three Hosts for an HCI layout

Note

Ubuntu 22.04 uses Netplan. Configure static Host networking before provisioning the cluster. Clustered storage and combined traffic configurations require at least 10 Gbps interfaces with jumbo frames enabled end-to-end.

Legacy Network Ports

Description

Source

Destination

Port

Agent communication

HVM Host

HPE Morpheus Software appliance

TCP 443

Host configuration

HPE Morpheus Software appliance

HVM Host

TCP 22

Interhost communication

HVM Host

HVM Host

TCP 22

Ceph Monitor

HVM Host

HVM Host

TCP 3300, 6789

Ceph MDS and OSD

HVM Host

HVM Host

TCP 6800-7300

Provisioning a Legacy Cluster

  1. Provision the required Ubuntu 22.04 Hosts and configure static networking.

  2. Create or select a HPE Morpheus Software-type Cloud for the cluster.

  3. Navigate to Infrastructure ‣ Clusters and click + Add Cluster.

  4. Select HVM and choose the required Legacy HCI or non-HCI layout.

  5. Enter each Host’s SSH address and credentials for a user with passwordless sudo access.

  6. For an HCI layout, select the Ceph data device on each Host.

  7. Select the management, storage, and compute interfaces and configure the required compute VLANs.

  8. Complete the wizard and monitor the cluster History tab until the KVM, Ceph, OVS, and Pacemaker configuration tasks finish.

Warning

Do not reuse the operating-system disk as a Ceph data device. Verify the selected device on every Host before completing the wizard.

Legacy Networking

Legacy layouts use OVS bridges and port groups. The compute VLANs selected during cluster provisioning create OVS-backed networks for workloads. For network lifecycle concepts and supported OVS network types, see HVM Networks.

Legacy High Availability

Pacemaker manages Legacy cluster resources, fencing, and failover. Use pcs status to inspect cluster state and pcs resource to inspect managed resources. Do not apply the agent quorum interpretation from the current architecture or troubleshooting pages to a Legacy cluster.

Before removing a Host or performing maintenance:

  1. Confirm Pacemaker and Ceph report a healthy state.

  2. Evacuate or stop VMs that cannot migrate.

  3. Enter maintenance mode from the Host Actions menu.

  4. Complete the maintenance or removal operation.

  5. Confirm Pacemaker resources, Ceph placement groups, and workload placement return to a healthy state.

Warning

Never remove a Ceph OSD, Monitor, or Host while the Ceph cluster is unhealthy or rebalancing. Contact HPE Support when a failed Host cannot be cleanly evacuated.

Shared Operations

The following current guide topics also apply to Legacy clusters unless a page contains a more restrictive layout notice:

Uploading Multi-Disk QCOW2 Images

HVM/KVM supports Virtual Images made from multiple QCOW2 disk files. These uploads require a metadata.json manifest on the same Virtual Image to map each QCOW2 file to its disk size, guest device, order, and boot role. Upload all QCOW2 files before uploading the manifest. Without valid metadata, HPE Morpheus Software treats the image as a single-disk image.

For the required manifest format and complete upload procedure, see Multi-Disk QCOW2 Images for HVM/KVM.

Upgrading from Legacy

Layout 1.3 replaces Pacemaker with the HPE Morpheus Software Agent quorum service while retaining Corosync and DLM. Use the supported cluster layout upgrade rather than manually removing Pacemaker. See Upgrading Clusters for the migration behavior and verification steps.