Features Released in Software Version 3.0.2-b1325160 (March 2025)
Caution:
Prior to patching or upgrading to the latest release, ensure that all compute nodes are in the provisioned state.
Platform Images
New platform images are made available for Compute Enclave users through Private Cloud Appliance installation, upgrade, and patching.
The following platform images are delivered with this Private Cloud Appliance release:
|
Oracle Linux 9 |
|
|
Oracle Linux 8 |
|
|
Oracle Linux 7.9 |
|
|
Oracle Solaris 11.4 |
|
|
Kubernetes Engine |
|
GPU Expansion
Add a GPU expansion rack to Private Cloud Appliance and build a scalable platform for AI and graphics intensive applications. The minimum expansion rack configuration contains a single GPU node with 4 NVIDIA L40S GPUs. More nodes can be installed after initial deployment. Two full racks, each with up to 6 GPU nodes, can be connected to the base rack.
When the racks are interconnected, GPU nodes become an integral part of the system and are managed like any other compute node. They are added to the same 3 fault domains, but the server families operate separately. Users must deploy compute instances with a dedicated shape to take advantage of the GPUs. These instances do not support live migration.
To learn about adding GPU capacity to a Private Cloud Appliance, see Optional GPU Expansion in the "Oracle Private Cloud Appliance Installation Guide".
New Disaster Recovery Service
A new disaster recovery (DR) service is introduced, with orchestration of DR operations built directly into the Service Enclave. This new implementation, also called the Native DR Service, requires a mutual, symmetrical peer connection between two Private Cloud Appliance systems, and uses a separate uplink from the spine switches. DR configurations and plans can be managed from the active or standby rack, and are automatically replicated to the peer system's DR service. If one of the peered systems goes down due to a site level incident, the failover operations defined in the DR plans are triggered from the standby system. Other DR plan operations can be run from either rack, because the peer connection allows both DR services to exchange data and instructions.
See Disaster Recovery in the Concepts Guide for an overview. A link to detailed information and instructions in the Administrator Guide is provided.
Existing installations can continue to use their first-generation disaster recovery configuration, orchestrated through Oracle Enterprise Manager with Oracle Site Guard. A path is provided to migrate existing configurations to the new Native DR Service.
New Limit Service
The Limit service enables you to view limits that are currently set for Private Cloud Appliance resources, and change those resource limits if the limit definition allows. See the Viewing and Setting Resource Limits chapter in the Oracle Private Cloud Appliance Administrator Guide.
Improved Upgrade and Patching Procedures
Both the Upgrade Guide and Patching Guide have been restructured to make it easier and clearer to follow the appropriate instructions based on active software version and target version. In the main instructional sections, it is assumed that the system is already at a minimum required version. For systems on earlier software versions, the procedures to get to the minimum required version are provided in a separate chapter. See Upgrading from Earlier Software Versions in the "Oracle Private Cloud Appliance Upgrade Guide".
Uplink Reference Topologies
The integration of a Private Cloud Appliance into the data center network can be challenging due to existing configurations and local requirements and standard practices. There are various ways to configure the uplinks, and Oracle provides assistance completing the relevant checklists in advance. To simplify and speed up the process, the Installation Guide provides examples of commonly used topologies, with switch configuration examples to guide network administrators in mapping their specific setup. See "Appliance Networking Reference Topologies".
Change to VM Console Port
As of this release, the VM console needs access to port 1443. In previous releases, port 443 was used for VM console access. For more information, see Port Matrix in Oracle Private Cloud Appliance Security Guide.
Kubernetes Engine
The following new features are added for Oracle Private Cloud Appliance Kubernetes Engine (OKE):
-
Node Doctor. The Node Doctor utility helps you troubleshoot a cluster worker node that is not Active or Running. Node Doctor can identify potential problem areas and provide information to help you address those problem areas, and the utility can collect node system information into a support bundle to enable you to get help from Oracle Support. See Using Node Doctor to Troubleshoot Worker Node Issues in the Oracle Private Cloud Appliance Kubernetes Engine user guide.
-
Node cycling. By default when you update a node pool, only new nodes that are added during this update or that are added later receive the updates. Node cycling enables you to replace existing nodes with new nodes that use updated settings. Node cycling performs an in-place update of all existing nodes in the node pool to the latest specified configuration. New nodes are created, workloads moved onto them from existing nodes, current node pool updates applied, and the original nodes terminated. See Node Cycling an OKE Node Pool in the Oracle Private Cloud Appliance Kubernetes Engine user guide.
-
Node labels. A node label is a key/value pair that enables you to target pods for scheduling on specific nodes or groups of nodes. See the OCI CLI procedure in Creating an OKE Worker Node Pool in the Oracle Private Cloud Appliance Kubernetes Engine user guide.
-
Persistent storage. New ability to create high performance block volume storage and create file system storage by using the CSI (Container Storage Interface) plugin. See the Adding Storage for Containerized Applications chapter in the Oracle Private Cloud Appliance Kubernetes Engine user guide.
SR-IOV Support
Single root I/O virtualization (SR-IOV) technology enables virtual machines to achieve low latency and high throughput simultaneously on 1 or more physical links. This technology is ideal for low-latency workloads such as video streaming, real-time applications, and large or clustered databases. Hardware-assisted (SR-IOV) networking uses the VFIO driver framework.
See SR-IOV in the Concepts Guide for an overview. Configuration information is available at "Configuring SR-IOV for Virtual Networking" in the Networking section of the User Guide.
Dynamic Routing Gateways Limit Increase
As of this release, the limits for Dynamic Routing Gateways (DRG) have increased to 32 total DRGs across all tenancies, with up to 16 SR-IOV DRGs.
Flex Network Limit Increase
As of this release, the maximum number of flex networks per port is now 8 with a maximum of 32 flex networks per rack. The following restrictions apply:
- The relationship between flex networks and DRGs is 1:1.
- A flex network cannot share a DRG with another flex network.
- A flex network cannot share a DRG with one or more direct attach Exadata networks.
When upgrading to this release, if you have existing flex or Exadata networks that share a DRG with other flex networks for Exadata direct connect, they can no longer share a DRG. Each flex network must have its own DRG.
Bugs Fixed in This Release
For a list of bugs fixed in each release, see Oracle Support Document 2906831.1 ([PCA 3.x] Private Cloud Appliance: Software Updates) can be found at: https://support.oracle.com/knowledge/Sun%20Microsystems/2906831_1.html.