October 2025 Patch 1 Release
This patch improves system reliability, enhances security, and fixes critical issues that affect Private Cloud Director operations. This includes support for self-hosted deployments.
Upgrade Notes
CPU Mode and Model customization:
PCD now supports configuring a CPU mode for new clusters. It can be one of custom (current default), host-model and host-passthrough . When using custom mode, you would also need to specify the CPU model to be used by all VMs in that cluster.
For existing clusters, If you had a non-default CPU mode/model configured in the nova_override.conf file on hypervisors prior to this release, please ensure to update that in the respective cluster object with the below API calls post the upgrade to this patch release.
# Query the current cluster speccurl -X GET -H "X-Auth-Token: <TOKEN>" https://<FQDN>/resmgr/v2/clusters/<cluster_name># Update the CPU mode and model in the above response and apply the patchcurl -X PUT -H "X-Auth-Token: <TOKEN>" https://<FQDN>/resmgr/v2/clusters/<cluster_name> -d '<modified-cluster-spec>'New Features and Enhancements
Pre-configured Volume-backed VM Flavors: Zero-disk flavors are now required for creating volume-backed VMs. Six diskless flavors now available out-of-the-box for volume-backed VMs and hotplug VMs, eliminating the need for manual flavor creation.
Volume Import from Storage Arrays via pcdctl:
Storage volumes can now be imported from backend arrays (Pure Storage, HPE Primera, etc.) using pcdctl, enabling workflows like database refreshes from production replicas without recreating volumes.
Enhanced Audit Logging for Core Services: API audit events from key system components now persist to filesystem, accessible via the audit PVC on the management node for compliance and troubleshooting workflows.
Customizable CPU Virtualization Settings: Cluster creation now supports configuring CPU mode (host-model, host-passthrough, custom) and model, enabling standardized CPU features across hosts for optimized live migration compatibility.
Automated Image Storage Cleanup: Deleted images now automatically removed from storage backends (filesystem, shared storage) without manual intervention.
Database Load Distribution for Percona Clusters: Percona HAProxy now supports round-robin request distribution across PXC nodes through replicas service, improving read performance and enabling read/write workload separation for high-concurrency database deployments.
Bug Fixes
VM High Availability for Hard Affinity Groups: VMs with hard affinity/anti-affinity policies now evacuate successfully during host failures, preventing error states caused by affinity constraint violations.
Proxy Configuration for Compute Service: Fixed inconsistent proxy environment variable handling across hypervisor services, eliminating manual configuration file edits previously required for hypervisor-to-DU communication in proxied environments.
Network Visibility for Self-Service Users: Shared external networks now appear in VM creation network list for self-service users, enabling seamless deployment in multi-tenant environments.
Root Volume Selection for Delete on Termination: Fixed volume deletion logic to correctly identify and delete root volumes when VMs are terminated. Attached volumes now default to preserve on VM deletion.