Modern virtualization platforms face the paradox that defines infrastructure management: the need to be adaptable enough to serve unknown future workloads, while maintaining the ability to serve existing mission business-critical applications – all with accountability and resilience.
The August 2026 release of Private Cloud Director addresses these needs directly, driven by customer and partner feedback on what they need from their virtualization platform. A few of the highlights of the August release include:
- Granular Role-Based Access Control
- Platform9 OS (Rocky Linux by CIQ)
- VM-to-Host Affinity for Stretched Clusters
- vGPU Live Migration
- Cross-Cluster Cold Migration
What does this mean for our customers and how do we provide the functionality you demand? Let’s dig into the details…
Granular Role-Based Access Control (Beta)
Previous versions of Private Cloud Director had a role model that worked for smaller environments but didn’t scale cleanly to the organizational structures of larger enterprise accounts. This release introduces three new administrative roles backed by a new policy engine that enforces scope boundaries rather than relying on convention. The new RBAC model is built on Open Policy Agent (OPA), allowing dynamic policy definitions. Building on OPA means PCD now supports policy as code, defining access policies alongside your existing infrastructure-as-code workflows.
Platform9 OS powered by Rocky Linux from CIQ (GA)
Platform9 OS is a hardened, RHEL-compatible Linux distribution that your security team will recognize and your compliance framework will accept. The GA release also introduces the Host Management Console, a UI for post-install host configuration that keeps the no-shell-required promise intact through the entire setup process. The hardened OS is fully supported by Platform9 with all updates and patches tested before being released.
VM-to-Host Affinity for Stretched Clusters
Private Cloud Director now supports stretched clusters across two physical sites, improving application availability by automatically recovering VMs on the surviving site in case of a complete site failure. Stretched clusters allow for greater resilience for site-level outages while preserving workload policies across sites.
vGPU Live Migration
With this release, vGPU VMs can live-migrate on Ubuntu 24.04 hosts, keeping workloads running while VMs move between hosts and bringing parity with existing CPU-based VM live migration. Expanded multi-vGPU flavor support, broader hardware compatibility through mediated devices, and more flexible GPU passthrough configuration make it easier to standardize and optimize GPU resources for demanding AI workloads, significantly enhancing our GPU infrastructure support capability.
Cross-Cluster Cold Migration
Cross-cluster cold migration enabled powered-off VMs to move between PCD clusters through the standard management interface. This gives administrators a straightforward way to relocate workloads during planned maintenance, rebalance capacity across clusters, consolidate infrastructure, and support cluster lifecycle changes without having to create the VM or rely on a separate migration workflow.
Additional Security, Networking, and Resiliency Enhancements
Those are the highlights of the release, but there’s plenty of additional features and improvements available. Volume encryption now integrates with HashiCorp Vault, allowing organizations to keep encryption keys within their existing key-management infrastructure and better support regulated environments. Simple (L2) Networking is now generally available, with native API support that eliminates manual port pre-creation, while a redesigned, guided workflow makes configuring storage backends such as NetApp and Pure Storage faster and more intuitive.
The release also introduces new ways to improve resiliency and operational awareness. Configurable VM HA response modes and an optional NFS-based storage heartbeat, both available in Beta, let administrators tune evacuation behavior and reduce false failovers caused by transient network issues. We’ve also improved observability. Continuous, component-level monitoring across etcd, the management plane, Kubernetes clusters, and nodes provides user-facing alerts when health degrades, helping teams identify issues before they affect users.
For the complete list of features, enhancements, and bug fixes, see the full release notes.
One More Thing: vJailbreak and VDDK
If you’ve seen the news about Broadcom withdrawing VDDK (Virtual Disk Development Kit) licensing for third-party migration tools, you may have wondered how this impacts vJailbreak. Short answer: it doesn’t. vJailbreak continues to operate without access to the VDDK. The migration workflows your team is counting on are intact, and the path to Private Cloud Director is still the most automated, lowest-cost option in the market.
Author
