July 2025 Release

The latest release of Platform9 Private Cloud Director includes new features, usability improvements, and resolved issues to enhance product stability and performance.

Enhancements

Enhanced DNS Configuration Across Networking Workflows

This release introduces improved DNS visibility and configuration options across the UI:

• View DNS Zone information on Network and Security > Physical Networks and Network and Security > Virtual Networks.

• Add DNS Domain and DNS Name when creating a Public IP from Network and Security > Public IPs > Create Public IP.

• Enable DNS publishing when creating or editing a physical network from Network and Security > Physical Networks

These enhancements streamline DNS setup and integration for floating IPs and subnet configurations.

DNS Domain Association for Networks

The PCD console now supports associating DNS domains with networks during creation and through updates. You can now specify DNS zones (domains) to attach to networks directly through the console interface, completing the designate service integration. This enhancement brings the previously CLI-only functionality to the console, enabling full DNS domain management capabilities for network resources through the graphical interface.

DNS Name Configuration for Ports

The PCD console now supports DNS name configuration for network ports. You can specify a DNS name when creating a new port or when updating existing ports. The system automatically creates corresponding DNS records in Designate using the specified DNS name.

This enhancement simplifies DNS management by allowing direct DNS name assignment through the console interface during port configuration workflows.

To use this feature, ensure your network has DNS publishing enabled (dns-publish-fixed-ip flag) and an assigned DNS zone.

App Catalog now supported for On-Prem deployments

Platform9 has enabled the App Catalog feature for on-premises environments. This update improves application lifecycle management for on-prem users.

Support for Retyping In-Use Volumes

You can now retype volumes that are currently attached or in use. When initiating a retype operation, the PCD console displays a warning that migration is driver-dependent and may fail based on the source or destination volume driver.

1

Select a new volume type

Select a new volume type from those enabled on active storage backends.

2

Acknowledge risks

Acknowledge the risks by confirming the checkbox before initiating the retype.

These changes provide more flexibility while preserving workload safety during active volume transitions.

Enhanced Service Health Checks for Critical Operations

The console now checks the health of key services, such as Compute, Storage, and Networking, before allowing critical operations. If any of these services are unresponsive, a banner will appear stating that critical services are not responding, helping you avoid failures during operations such as VM creation. This enhancement provides better visibility into control plane health before initiating key workflows.

Improved Grafana Login Experience

Grafana now supports login using Private Cloud Director credentials. On fresh installs, you can log in to Grafana using the same credentials as your PCD console by clicking on the Sign in with PCD button on the Grafana login screen. Alternatively, you can log in using Grafana's default admin credentials.

Please refer to the Upgrade Notes section for expected behavior and workarounds for older deployments upgraded to the July release.

BETA Enhanced Storage Support for Windows Clusters

Added support for shared storage that enables Windows Server clustering in OpenStack environments. This feature provides VMware-like capabilities, allowing multiple virtual machines to safely share the same storage devices.

Upgrade Notes

Volumes Placed on Incorrect NFS Backend

This release fixed an issue where, if you had multiple NFS backends on one host, volumes would be placed in the incorrect backends. If you are upgrading from a previous release and have multiple NFS backends, you should specify different directories where you want each volume to be created on the host for each of the backends in the blueprint in the field nfs_mount_point_base. The directories should be unique for each backend. Ensure that if you specify a directory that is not in /opt/pf9/pf9-cindervolume-base, you should pre-create it and set the owner to pf9:pf9group. Lastly, update the nfs_shares_config for each of the backends to be unique in the blueprint.

Grafana PCD Authentication

When you upgrade from the June release to the July release, you may encounter a user sync error when logging into Grafana with PCD credentials. This error occurs when a local Grafana user account exists with the same username as your PCD user account.

As a solution, perform these steps to resolve the user sync error:

1

Log in to Grafana

Log in to Grafana using your existing local credentials (for example, the default admin user).

2

Create a new admin user

• Navigate to the user management section.

• Create a new user account.

• Assign both the Grafana Admin and Organization Admin roles to this user.

3

Delete conflicting users

• Identify local users whose usernames match your PCD usernames.

• Delete these conflicting local user accounts from Grafana.

4

Test PCD login

Test PCD login using your PCD credentials with the previously conflicting username.

Optional, configure Dashboard Permissions

1

Navigate to permissions

Navigate to Dashboards > Edit > Settings > Permissions.

2

Add permissions

Add viewer or admin permissions for your SSO user account.

3

Repeat for each dashboard

Repeat this process for each dashboard that requires access.

This manual step applies to upgrades from both April and June releases.

Upgrade Behavior Summary

• Pre-June to July release upgrades: You can continue using Grafana's default credentials (admin/admin) or switch to PCD login with your PCD credentials.

• July release deployments: Both authentication methods work without additional configuration.

• June to July release upgrades: Follow the resolution steps above if you encounter user sync errors.

Bug Fixes

Identity, Storage, and Networking Services

• Resolved an issue where live migrations and hard reboots caused hotplug VMs to lose their configuration and behave as static VMs, breaking hotplug functionality.

• Resolved an issue where live migration and DRR operations caused packet loss in VMs. Upgraded the networking backend to OVN version 23.0, reducing packet drops to zero in most cases.

• Resolved an issue where VM creation failed No valid host found due to stale resource allocations in the compute scheduler. The compute service now avoids allocating resources during automated migration planning, preventing phantom allocations.

• Improved CPU model selection strategy on hypervisors to ensure the latest supported model is used. This fix prevents VMs from getting stuck in the Booting from the disk state due to unsupported CPU configurations.

• Fixed an issue that resulted in volumes being created in incorrect NFS backends when a host had two or more NFS backends configured.

PCD User Interface

• Fixed an issue where assigning new metadata to a flavor removed all existing metadata from that flavor.

• Any updates to the metadata during flavor edit caused the flavor to update automatically without selecting Assign Metadata.

• Fixed an issue where Virtual Machines > Images displayed only 25 images, causing existing images to disappear from view and blocking VM deployments via the console. After the fix, the list now displays all onboarded images, including those beyond the initial 25.

• Fixed an issue on Network and Security > Security Groups > Create Security Group, where, during Inbound Security Group Rules creation, the Custom ICMP Rule did not retain input. Although ICMP Type and ICMP Code were mandatory, the values were not saved, and the rule was incorrectly allowed to persist with empty fields.

• Improved error feedback and end-user messaging across various screens.

• Disabled migration for VMs in suspended state to prevent unsupported operations that previously failed without an error.

• Cluster selection is now mandatory when creating a VM using an existing volume or VM snapshot, which was previously missing.

• Added validation for volume backend name when configuring a cluster blueprint to allow only letters, digits, hyphens (-), and underscores (_), and disallow spaces or multiple words.

• Fixed an issue where hotplug-compatible flavors were incorrectly listed during VM creation, even when the Hot-plug Compatible option was not enabled.

• Enabled metadata editing for pf9-managed host aggregates on Infrastructure > Host Aggregates.

Self-Hosted PCD

• Fixed an issue where support bundle generation failed for hypervisor hosts in on-prem environments. Once password-less SSH access was configured from the Airctl host to the hypervisors, support bundles were successfully collected using the specified hypervisor IPs.

Here is an example:

airctl gen-support-bundle  --host-ips <HOST-IP>

Miscellaneous

• Resolved an issue where the GPU status was displayed as disabled for one of the GPU nodes in a multi-node cluster.

• Fixed an issue where GPU Passthrough mode validation was reporting IOMMU not configured, even though the GPU host was successfully configured with IOMMU and onboarded to PCD.

• Optimized the performance of the Virtual Machines page to be efficient at scale by adding pagination and enhancing backend query handling.

• Fixed an issue where deleted regions left residual metadata in the service catalog, causing them to appear in the OpenStack CLI. Region deletion now ensures complete cleanup of associated entries.

• Fixed an issue on Virtual Machines > Virtual Machines > Deploy New VM, where hotplug compatible flavors did not require CPU and memory values when the Hot-plug compatible option was enabled, causing VM creation to fail. Now, the CPU and memory fields are validated to ensure that the required values are set for hotplug-compatible configurations.

• Fixed an issue where cold migration temporarily showed a failed state before completing successfully.

• Fixed an issue that caused the new image upload dialog to close automatically when a previously triggered image upload completed. This interrupted ongoing uploads and prevented users from uploading additional images.

Known Limitations

• Cold Migration Unsupported for Hotplug VMs: Cold migration is not supported for hotplug-enabled VMs. Attempting cold migration causes the VM to lose its hotplug configuration and revert to a static configuration.

• GPU Passthrough Limitation for VM Creation: When using GPU passthrough mode, only one GPU host configuration is allowed per region.

• GPU VM Creation Fails with No Valid Host Was Found Error: You may see the message "No valid host was found." There are not enough hosts available when creating a VM using GPU passthrough flavors. This can occur if SR-IOV is not enabled for the GPU device. It is recommended to verify if the GPU supports SR-IOV and enable it before configuring GPU passthrough.

• Kubernetes Cluster Names Must Be Unique Across Regions: Two clusters cannot share the same name across regions within the same tenant.

• Tenant Name Restriction: Spaces are not supported in tenant names. Use only alphanumeric characters, dashes, or underscores.

Known Issues

• Re-enabling SSO after it has been disabled may fail with the error Identity provider already exists. This occurs because a check prevents the reuse of the same identity provider across domains.

• When you assign multiple storage backends to a host and then remove and re-add them, you may need to manually re-enable them.

1

Find the backends to enable

Run:

pcdctl volume service list

2

Re-enable the required backend(s)

Run:

pcdctl volume service set --enable <HOSTID>@<BACKENDNAME>

• VM HA does not honor the host liveness traffic network interface configured in the cluster blueprint in this release.

• VM HA and DRR do not support vTPM-enabled VMs. Live migration and evacuation are not possible, so these VMs will not be migrated automatically.

• If you are using NFS as the backend for block storage, set the image_volume_cache_enabled flag to false. If the flag is set to true , creating a VM from a cached image volume may lead to incorrect root disk sizing.

• SSO users are unable to create Heat orchestration stacks at this time.

• pcdctl config set command is not supported for a user with MFA enabled.

• Image upload to encrypted volumes is currently unsupported. Volume encryption only works with empty volumes at this time.

• Currently, rescue mode is only supported for VMs with ephemeral storage. The rescue operation does not work for instances backed by volumes. Users attempting to rescue a volume-backed instance will encounter failures.