Harvester 1.3.1: Elevating Cloud Native Virtualization, Optimizing AI Workloads and the Edge

Today, at SUSECON 2024, we are excited to share the latest release of Harvester: our 100% open source software to seamlessly manage containers and virtualized environments, and at the edge. This update brings a host of highly anticipated features, like NVIDIA vGPU support for cloud native virtualization to optimize AI workloads. For robust, highly-available virtualization in demanding edge scenarios deploy a Witness Node for two-node clusters. Even more reliability with optimization for devices abruptly powered off and on. We are also excited to announce a technical preview of ARM enablement and new cluster management capabilities using Fleet. Let’s dive into the standout features of this release.

Enterprises’ Need for Cloud Native Virtualization

In today’s fast-paced digital landscape, enterprises are increasingly seeking agile and scalable solutions to manage their IT infrastructure. Cloud native virtualization offers unparalleled operational flexibility, enabling businesses to efficiently manage both virtual machines (VMs) and containerized workloads. As enterprises search for solutions for skyrocketing virtualization licensing and subscription fees while remaining agile in the cloud-native world, Harvester addresses this critical need by providing a unified platform that enhances resource utilization, reduces costs and simplifies operations.

Latest 1.3.1 Features 

• NVIDIA vGPU Support: Harvester now allows users to leverage NVIDIA GPUs for SRIOV-based virtualization, enabling the sharing of GPU resources across multiple VMs. This feature enhances general performance for GPU-intensive workloads. For detailed instructions on configuring vGPU, please refer to the documentation.

• • ARM Support (Technical Preview): Harvester now supports installation on ARM-based servers, thanks to recent updates to KubeVirt and RKE2, which both support ARM64 architecture. This technical preview allows users to explore the benefits of Harvester on ARM platforms, broadening its applicability.

• Witness Node: Highly Available Two-Node Clusters This release introduces support for two-node clusters with a witness node, providing high availability without the need for larger deployments. This configuration is ideal for environments with limited resources requiring resilience through frequent interruption and relocation. The witness node helps maintain cluster operations, ensuring uptime and reliability. More details are available in the documentation.

• Optimized for Frequent Device Power-Off/Power-On: Harvester is now optimized for environments that experience frequent power interruptions or device relocations, such as at the edge or remote environments. The new optimizations ensure VMs are shut down safely and responsibly, ensuring the cluster remains stable even after abrupt shutdowns and restarts, reducing the operational burden on cluster administrators.

• Managed DHCP (Experimental Add-on): This experimental feature simplifies IP address management within clusters. Administrators can configure IP pools and automatically assign IP addresses to VMs, streamlining the deployment process. Managed DHCP uses the vm-dhcp-controller add-on to handle DHCP requests efficiently. See the documentation for setup details.

• Fleet Management (Technical Preview): Fleet is now integrated for managing and deploying objects, such as VM images and node settings in Harvester clusters. Fleet support is enabled by default and functions independently of Rancher, though it can also manage Harvester clusters imported into Rancher. This feature enhances scalability and simplifies cluster operations.

Harvester at SUSECON 2024

We are thrilled to announce that Harvester will be prominently featured at SUSECON 2024. This premier event is the perfect opportunity to see Harvester in action and learn more about its new features directly from our experts. Be sure to attend our sessions and visit our demo booths to get hands-on experience and deeper insights into the advancements of Harvester.

Don’t Miss These Exciting SUSECON Sessions:

• SUSE Enterprise Container Management Roadmap: Review the future of Rancher Manager, NeuVector, Harvester, Longhorn, and more.
• Plan your migration off vSphere Discover how Harvester can seamlessly replace your virtualization platform.
• Leveraging Power of SUSE Harvester & Dell PowerFlex: Explore how Dell PowerFlex and Harvester offer a unified solution for modern enterprises.
• Migrating from VMware to Harvester: Learn how a German university transitioned to Harvester, creating a cutting-edge HCI environment.
• Harvester for ARM Environments: Discover the technical preview of ARM-based virtualization support in Harvester.
• Deploying RKE2 Clusters on Harvester: Hands-on session showcasing Rancher’s support for Cluster API with Harvester and RKE2.
• Hyperconverged Infrastructure Solutions: Overview of SUSE’s Harvester HCI solution, exploring scalability and flexibility.

A big thank you to the Harvester development team for their tireless efforts in bringing these features to life. We invite you to explore Harvester 1.3.1 and share your feedback through our Slack channel or GitHub. Your input is invaluable in shaping the future of Harvester and 100% open source software.

Thank you for your continued support and engagement with the Harvester project!