v9.1.0 Release Notes¶

• Compatible Plugin API version: 1.4.1

• Compatible HPE Morpheus Software Worker version: 9.0.0

• Minimum upgrade version: v9.0.0. Upgrading to v9.1.0 requires a non-rolling upgrade.

Release Dates

• v9.1.0 TBD

New Features¶

Licensing:

• Introduced the Advanced license tier, expanding the platform to three tiers: VM Essentials, Advanced, and Enterprise. Advanced includes Ansible, Kubernetes (HKS), Workflows, Service Catalog, Personas, and additional automation and integration capabilities. See Feature Comparison by License Tier for the complete tier comparison.

HVM & VM Essentials:

• Virtual Switch management: add, edit, and remove virtual switches with advanced port group management and network reconfiguration directly from the UI

• Linux-bridge/virtSwitch integration for Standard, Private, and Data Networks, with virtual switch discovery, per-host uplink mapping, and multiple iSCSI VLAN segments on one switch

• Add Host wizard improvements: probe NICs, validate compatibility, and map the new host’s interfaces to cluster virtual switches

• NVIDIA vGPU slicing: create SR-IOV vGPU profiles on supported HVM Hosts and bind physical GPU types or vGPU Segment Types to Service Plans

• Host Profiles: define and apply standardized host configurations across cluster hosts, similar to VMware vCenter Host Profiles

• Host Configuration Backup & Restore: back up HVM host identity and configuration for faster hardware replacement

• Bulk operations: trigger actions on multiple VMs or hosts simultaneously (migrate, snapshot, power operations, tag, and more)

• Increased live migration concurrency during Maintenance Mode to reduce maintenance windows for dense environments

• Combined compute and storage migration: select a target host and datastore in one Move operation, including hosts without shared storage, within a cluster or between HVM clusters in the same Cloud. Cross-cluster moves also map destination networks. Running VMs use live migration; powered-off VMs move cold. See VM Migration & Failover

• Live migration uses the dedicated live-migration network segment when available

• Connect and disconnect virtual network interfaces on running VMs without requiring a full reconfigure, supporting backup recovery and security isolation workflows

• Extended advanced VM settings in Reconfigure, including nested virtualization editing, restart warnings, and validation that Secure Boot requires UEFI

• Centralized Affinity Placement management: view and manage VM affinity rules and pinning overrides from a single cluster-level interface

• Affinity rules support required (MUST) and preferred (SHOULD) strength, Host Groups, VM Groups, and combined VM-to-Host rules for HVM and VMware vCenter clusters

• Linked clones: create space-efficient clones that share a base disk image with the parent VM

• Datastore Evacuation / Maintenance Mode: automatically migrate VMs off a datastore being retired or replaced

• Raw Device Block Mapping (RDBM): attach raw host block devices directly to VMs for performance-sensitive workloads

• Virtual machine console access directly from HVM hosts

• Graceful shutdown and forced power-off choices for VMs, defaulting to graceful shutdown

• Legacy OS option using i440fx virtual hardware, with machine-type selection in provisioning and Virtual Images

• Boot menu enabled by default and configurable boot delay for HVM VMs

• Remove the primary CD-ROM device from HVM VMs

• SCSI-3 Persistent Reservations on block-type multi-attach volumes for guest clustering

• Boot From SAN support for HVM hosts, including Fibre Channel multipath fresh installations

• Add or remove multiple hosts, and apply tags or policies to many VMs, in a single operation

• CPU Ready-equivalent metric for monitoring vCPU scheduling contention

• Leverage QEMU Guest Agent to gather guest OS information (IP addresses, OS version, hostname) without network-based access

• VSS-aware application-consistent snapshots: when the QEMU Guest Agent is connected, snapshots trigger Windows Volume Shadow Copy Service for database and application consistency

• vNUMA provisioning: automatic or manual virtual NUMA topology for wide VMs spanning multiple CPU sockets, with Off/Auto/Manual modes

• vCPU Placement: cluster-level CPU affinity management with Unmanaged, Region, and Pinned modes for NUMA-aware workload performance

• Improved audit logging: administrative actions (datastore changes, cluster settings, VM promotions) now recorded in object History for full traceability

• HVM OS 26.04 and QEMU 10.2 machine-type support with cluster layout 2.0

• Host NTP configuration and improved guest timekeeping with RTC catchup and kvmclock for Linux guests

Monitoring & Administration:

• Monitoring tab time ranges for instance, server, and cluster details, allowing analysis over configurable periods

• Native threshold alarms for host and VM CPU/memory usage, host reachability and memory overcommit, and datastore capacity/offline state, with configurable thresholds and cluster overrides. See Cluster Alarms

• GPU monitoring for HKS clusters and worker nodes, and KVM/HVM hosts and instances. Agent-collected device metrics expose utilization, memory, temperature, and power where supported, including NVIDIA MIG slices. AMD and Intel GPUs are supported alongside NVIDIA

• Storage-volume performance charts for IOPS, throughput, and latency

• Log rotation and retention controls from the GUI, with workload logging settings available to VM Essentials. Logging Settings help distinguishes workload/host controls from appliance logging

• Optional agent log throttling with Off, Monitor, and Enforce modes, per-server policy overrides, cooldown-based re-enablement, and duplicate-message suppression. Throttling defaults to Off

• Read-only access level for Appliance Settings permission, allowing auditors and operators to view configuration without modification rights

• Key-value metadata tags and tag-based filtering/search for Instances, VMs, Hosts, Networks, Datastores, and Clusters

• Global Search results grouped by object type, with richer identifying context, filtering, sorting, selectable columns, and CSV export

• Enhanced data tables with sorting, column visibility and reordering, and persistent preferences

• Role Details limits Instance Types to those available under the appliance license

Storage & Backup:

• Datastore Groups: group shared HVM file-based datastores as a provisioning target with automatic member selection. Storage DRS supports capacity-based rebalancing or recommendations-only operation. See Storage Lifecycle

• HPE Alletra MP synchronous and asynchronous periodic replication volume types for VM and volume provisioning

• Integrated HPE Alletra MP volume and GFS2 datastore provisioning workflow

• Unified File Support Phase 2: create, grow, and inventory Morpheus File Shares, with per-filesystem ransomware settings

• Datastore Explorer shows VM inventory and volume usage, with browsing, upload, download, and deletion for supported file-based datastores

• Snapshot size displayed in Instance and Server snapshot tabs

• Storage-migration guards for VMs with snapshots to protect snapshot data

• Host Storage Interfaces view for Fibre Channel and iSCSI discovery, asynchronous storage rescans, and datastore multipath status

• Backup-job visibility and Tenant permissions can be managed through the REST API

BMaaS / Bare Metal:

• PXE/Network Boot support: boot Bare Metal hosts via customer PXE infrastructure instead of virtual media, with optional Config ISO for post-deployment automation

• Brownfield import in Boot From SAN (BFS): discover and manage existing BFS-provisioned servers

• Support for multiple networks (VLANs) with Unmanaged Networking on a single physical or bond interface

• Bare Metal support bundles include server diagnostics and Active Health System (AHS) log retrieval

• Convert to Managed supports Linux servers using active-backup bonding and discovers boot-volume information

HKS / Kubernetes:

• Kubernetes Gateway API support replaces NGINX Ingress Controller in updated layouts, with an optional Envoy Gateway controller package and offline installation support

• HKS cluster packages updated to latest Kubernetes versions (1.33, 1.34, 1.35) with CVE evaluation

• Updated containerd 2.x LTS support and cluster runtime upgrades

• Hybrid layouts support VM control-plane nodes and Bare Metal workers, including a single-worker topology for PCO Lite

• Optimized cluster refresh

Networking & SDN:

• Change the primary network role for vCenter Instances in the UI

• Tenant-aware Networks API responses for Zerto integrations, including the includeTenants option

• SDN onboarding prechecks, controller/dataplane upgrade workflows, configurable host VTEP gateway at creation, and BGP attribute viewing and editing

• RFC 8365 Route Type 2 interoperability and DPDK NIC validation for supported HPE QLogic and Mellanox adapters

• Aruba CX LACP-fallback support and virtual switch integration for Aruba CX DSS Networks

• Optional embedded DNS resolver for small PCAI internal zones, with local A and CNAME records. The resolver requires restricted network exposure because answers are not Tenant-scoped

Platform & Orchestration:

• PCO Support for Host addition and removal within existing VMware/VME clusters

• Unified & Contextual Approval Framework for consistent approval workflows across provisioning operations

• Import existing systems into HPE Morpheus Software for management without reprovisioning

• Provider-driven update actions for supported clusters, compute, storage, and network servers

• Simplified HVM installation and a unified Windows/macOS installer for VM Essentials Manager and Morpheus deployments

• Distributed Worker support for iLO consoles and HVM console connections

• Embedded Commvault, Veeam, and UpCloud integrations are removed from the core product and are delivered as plugins

• Optional preservation of Windows static IP configuration during VMware-to-KVM migration, using the QEMU Guest Agent on the destination VM

Security:

• Phase 2 STIG hardening and FIPS enablement for HVM OS and Morpheus Manager QCOW2/OVA images

• SDN dependency and security-hardening updates

CLI & Tooling:

• hvmcli: unified CLI tool for HVM host management including log viewing, VM/node/storage/cluster status, network configuration, and health validation

• hvmcli host software update commands for online and offline updates, status, and history

• Generate and manage support bundles through the UI and MCP tools, including NVIDIA diagnostic collection for HVM

• Aruba CX Support Bundle: collect diagnostic logs from Aruba CX switches as part of the HPE Morpheus Software support bundle

• Search plugins in the Marketplace

• Plugin ReadMe Standardization

• Exclude optional bundled plugins from installation during reconfigure using ui['excluded_plugins'] in morpheus.rb. Current-version installed plugins are retained; excluded older versions are removed during upgrade. Protected plugins cannot be excluded. A removed plugin may remain greyed out until its record is deleted from the UI

Fixes¶

HVM:

• Fixed an issue where Dynamic Placement continuously attempted to move VMs to a memory-exhausted host, causing repeated migration failures

• Fixed allocated memory displaying incorrect values after cluster syncs

• Fixed network statistic information not displaying in the cluster view

• Fixed failure to attach a volume to a VM in certain configurations

• Addressed combined migration destination-datastore handling, cross-cluster host preparation, and premature storage-copy timeouts

• Improved affinity enforcement during provisioning, maintenance, and HA failover, including required Host Group constraints, pinned placement, and violation reporting

• Fixed host addition and network initialization issues on layout 2.0 clusters, and upgrade finalization issues in mixed VMware/HVM systems

• Fixed legacy machine-type preservation during backup restore, vNUMA memory-resize handling, and CD-ROM validation blocking Reconfigure

• Fixed vTPM snapshot metadata preservation during datastore migration and QEMU Guest Agent OS-information refresh

• Fixed stretched-cluster and GFS2 heartbeat fencing that could stop VMs or prevent HA failover, and VMs not auto-recovering after a network switch reboot

• Fixed UEFI, vTPM, Secure Boot, and boot-option settings being lost between VMs, Virtual Images, and provisioning

• Fixed disks added to running VMs disappearing from inventory

• Fixed cluster layout finalization being blocked by agent installation, and incorrect host status after reboot from Maintenance Mode

• Fixed Primary CD-ROM removal being blocked in Reconfigure, and datastore evacuation follow-up issues

Monitoring:

• Fixed old open incidents not being automatically closed when all health checks return 100% healthy

• Fixed health-log date filtering, performance-chart time windows, CPU Ready value/severity inconsistencies, and alarm acknowledgement from detail views

HKS / Kubernetes:

• Fixed HKS cluster provisioning failures in proxy environments

• Addressed Calico proxy timeouts, upgrade version inconsistencies, worker lifecycle issues, Hybrid HA joins, and Helm App Blueprint upgrades

• Fixed license-related cluster creation failures and preservation of Hybrid layout during upgrades

BMaaS / Bare Metal:

• Fixed physical-interface discovery during Convert to Managed, brownfield switch cleanup, and attaching existing shared volumes

• Fixed Bare Metal datastore creation and boot-volume discovery failures

Storage & Backup:

• Addressed HPE Alletra MP clustered-datastore snapshot revert, cloning, and backup restore failures

• Fixed replicated-volume cloning, replication status after switchover, and unsafe resizing while replication links are down

• Fixed file-share capacity growth and ransomware-state synchronization, as well as datastore selection of snapshot paths instead of base shares

• Fixed Datastore Group threshold validation, inventory visibility, and recommendations in the VM Essentials tier

• Fixed backup creation fields, provider filtering, and Veeam subtenant backup execution

• Fixed missing provisioning-time backups for ESXi and Rubrik RSC integrations

• Fixed multi-disk image deployment to HPE Alletra MP and high-disk-count VM creation on GFS2 datastores

Networking:

• Fixed VLAN-0 and hybrid-network connectivity, private-network cleanup, and Aruba CX VLAN provisioning/cleanup

• Fixed Aruba CX DSS network selection and port-group resource-pool validation

• Fixed SDN overlay connectivity after VM migration, duplicate PowerDNS records, and IP pool deletion after partial EfficientIP sync failures

• Fixed SDN Mellanox SR-IOV onboarding and offboarding after VTEP communication failures

Clouds & Provisioning:

• Fixed Azure disk-name/size handling and Marketplace synchronization, concurrent SCVMM cloud sync, and OLVM guest/worker provisioning

• Fixed VMware affinity synchronization and validation, and API Proxy selection during cloud creation

• Fixed Windows rename/domain-join regressions, ISO filenames containing spaces, and expired multi-disk image download tokens

• Fixed Terraform resource-key length errors and Scribe imports duplicating form inputs

• Fixed migration detail pages returning HTTP 403 when the source VM still exists

• Fixed post-provision workflows not running on KVM, Xen, OpenStack, and Nutanix Prism, VCD provisioning wizard stalls, and Stop Instance not stopping its container

Administration & API:

• Fixed inherited cloud/group grants on locked subtenant roles, subtenant cloud access, and remembered-login API authorization

• Fixed ServiceNow approval links and saved flow/workflow selection, plus Ansible Tower inventory Tenant permissions

• Improved cluster, compute inventory, and dashboard load performance; fixed pagination resets and custom whitelabel rendering

• Fixed SCVMM overbilling, workflow-layout invoicing, pricing precision, and saved service-plan memory ranges

• Fixed repeated HKS refresh history entries, billing API timeouts, and saved password-field display

CLI & Appliance:

• Fixed hvmcli help handling that could execute an action, invalid-argument handling, large-log viewing, and cluster membership reporting

• Fixed software-update history/status and rollback selection, FIPS package handling, and Host Profile NTP remediation

• Fixed hvmcli audit-log access for non-root users

• Fixed installer image-size issues, fresh-install seed failures, and Distributed Worker connection initialization

• Fixed upgrade failures on 3-node HA appliances with large usage tables, and Windows agent upgrades from the UI on Windows Server 2022

• Fixed cluster creation and Manager deployment on FIPS/STIG-hardened hosts

Security:

• Hardened MCP destination validation and category gating, command allowlist enforcement, read-only permissions, and credential redaction in hvmcli audit logs

• Fixed the OLVM plugin disabling TLS certificate verification process-wide

Known Issues¶

HVM & Storage:

• Moving an image-store disk from HPE Alletra MP to a GFS2 datastore on a stretched cluster can time out

• A failed NFS pool start can cause cluster storage to report 100% usage

• VM provisioning can fail with qemu-img: Could not open after the local image-cache volume is purged

• Image upload to a new HVM HCI 1.3 cluster can fail

• Some nodes can fail to upgrade during a cluster layout upgrade

Installer:

• The HVM installer can crash when Use Entire disk is selected

• The GUI installer can select the wrong management virtual switch on hosts with multiple virtual switches

Networking:

• OVS Bridge Domain router creation accepts an interface already attached to another OVS bridge and returns a generic error

• Router deletion blocked by associated Networks and VMs returns an unclear error message

Integrations:

• The Rubrik plugin is not packaged with 9.1.0 builds

• Veeam backup restore can display incorrect options

• Creating a snapshot of a VMware Cloud Director VM can fail

• A Distributed Worker can drop a Windows agent session on reconnect

• Hostname updates can intermittently fail during server build

Appliance & Agent Updates¶

Node Packages:

• HPE Morpheus Software Node and VM Node packages updated to 3.3.11

Linux Agent:

• Updated to v4.0.0

• Agent runtime updated to bundled Java 25 for native GPU-library integration

Windows Agent:

• v2.6.1.0

macOS Agent:

• v2.4.0

Plugin API:

• Updated to 1.4.1