# SUSE Virtualization / Harvester 日跟踪 · 2026-09-03
> 日期：2026-09-03（Asia/Shanghai）。原料：archive 原文，非旧日报。目录对应 archive `suse-virtualization/`。

## 摘要
- **SUSE Virtualization 1.8**：口号「More Control. More Paths Off VMware. No Forced Bundling.」——可组合、不捆绑，针对 Broadcom 套餐反感。
- **边缘两数据节点 HA**：两节点 + arbiter 跑高可用虚拟化（远边缘功耗/空间约束）。
- **GitHub**：Harvester `v1.9.0-rc7` 仍在测（升级 addon/nvidia-driver-toolkit 等 bug），正式 1.9 未 GA。
- **SUSE AI Factory**：开源底座 vs 「with NVIDIA」增强栈并列。
- 产品页维持「100% 开源 HCI、VM+容器、可预期定价」。

## 原文摘录

### [SUSE Virtualization 1.8 More Control. More Paths Off VMware. No Forced Bundling.](https://www.suse.com/c/suse-virtualization-1-8-enterprise-controls-migration-composable-platform/)
- 来源：SUSE 博客 / `suse-virtualization/suse-virtualization-1-8-more-control-more-paths-off-vmwa-b456111e.txt` / 抓取 2026-09-03T00:39:49Z
> SUSE Virtualization 1.8 More Control. More Paths Off VMware. No Forced Bundling. Organizations running VMware may face a familiar set of problems. Licensing costs may have climbed along with inflexible renewal terms. The decision to move off legacy hypervisors is easy. The question is how? SUSE Virtualization 1.8 is built for the moment. This release adds the controls enterprises need to run production workloads with confidence, three supported migration paths to move at your own pace, and a composable platform model that lets you buy only what you need. Teams governance as a need and SUSE Virtualization 1.8 closes that gap. Granular RBAC with specific roles. Admins can now assign fine-grained permissions that are API-enforced within Rancher Manager. This means multi-tenant environments get proper access controls without external tooling or manual workarounds. Guest cluster backup and restore. Workload protection is now built in at the storage level through the CSI driver. Teams can back up and restore guest Kubernetes clusters (like RKE2 and K3s) running on SUSE Virtualization…

### [Running highly available virtualization at the Edge with just two data nodes](https://www.suse.com/c/ha-virtualization-at-the-edge-two-data-nodes/)
- 来源：SUSE 博客（tag 页同样收录） / `suse-virtualization/running-highly-available-virtualization-at-the-edge-with-just-two-data-f7283e50.txt` / 抓取 2026-09-03T00:40:05Z
> Running highly available virtualization at the Edge with just two data nodes Tags: cloud native, edge computing, Everpure, Far Edge, High availability, infrastructure, Kubernetes, NVMe, Portworx, Storage Architect, Supermicro, SUSE, SUSE Virtualization, Tech Reference, TNA, Two Node with Arbiter, Virtualization Categories: Supermicro, SUSE Documentation, SUSE Virtualization, Technical Solutions, Virtualization Workload Repatriation Is a Resilience Strategy, Not an IT Project Tags: Business Strategy, Cloud Migration, Cloudbase Solutions, Compliance, Coriolis, cost optimization, Cyber Resilience, data sovereignty, digital sovereignty, digital transformation, Enterprise IT, Harvester, Hybrid Cloud, Infrastructure Governance, Infrastructure Strategy, IT Resilience, KVM, Multi-cloud Strategy, Open Source, Risk Management, SAP Environments, SAP HANA, SUSE, SUSE Virtualization, vendor lock-in, Virtualization, VMware migration, Workload Repatriation, zero downtime Categories: Alliance Partners, Cloud Computing, Compliance, Configuration Management, Containers, Digital Sovereignty, Digital…

### [Releases · harvester/harvester](https://github.com/harvester/harvester/releases)
- 来源：GitHub Releases / `suse-virtualization/releases-harvester-harvester-88e16c30.txt` / 抓取 2026-09-03T00:43:07Z
> 💿 https://releases.rancher.com/harvester/v1.9.0-rc7/harvester-v1.9.0-rc7-amd64-net-install.iso 📝 https://docs.harvesterhci.io/v1.9/install/net-install/ - [BUG] After upgrade to v1.9.0-rc6, addon deployment job failed for nvidia-driver-toolkit and pcidevices-controller - #11469 - [BUG] upgrade from v1.9.0-rc6 stuck downloading image - scratch PVC not created - #11466 - [BUG][UI] vGPU on RKE2 VM created by Rancher does not show on UI - #11456 - [BUG] Harvester VIP lost after upgrade complete to v1.9.0-rc5 - #11440 - [BUG] refresh install ARM cluster stuck on setting up node - #11395 - [BUG] Fail to migrate VM in v1.9.0-rc4 due to "CPU doesn't match specification: missing features: pdcm" - #11358 - [BUG] VM does not LiveMigrate but Off during v1.8.1 to v1.9.0-rc4 upgrade - #11356 - [FEATURE] add harvester-kernel-module-devel image - #11263 - [BUG] V2 Disk Provisioning Fails Due to Conflicting IOMMU Group - #11041 - [ENHANCEMENT] Support provisioning of guest clusters on Harvester VMs attached to overlay networks - #9682 - [FEATURE] block registration of new nodes while upgrade is in…

### [SUSE Virtualization 产品页](https://www.suse.com/products/rancher/virtualization/)
- 来源：SUSE 产品 / `suse-virtualization/suse-virtualization-42c80152.txt` / 抓取 2026-09-03T00:35:45Z
> SUSE Virtualization Kubernetes-Native Infrastructure Benefits Open-source hyperconverged infrastructure (HCI) for organizations that need to run VMs and containers together in the datacenter, public clouds, AI, and at the edge. With built-in backup and restore, enterprise governance, and predictable pricing. 100% open source with no per-CPU or per-VM fees. FIS-ASP avoided VMware renewal cost increases of more than 100% by moving 1,000 SAP systems to SUSE Virtualization. Thales Alenia Space cut virtualization costs 3x. Your costs stay predictable. Guest cluster backup and restore protects workloads at the storage level. Live migration moves running VMs between nodes with zero downtime. VM Auto Balancing optimizes placement automatically. Switch runs mission-critical ops at 100% uptime. Granular RBAC with SUSE Virtualization-specific roles, API-enforced within Rancher Manager. Native NVIDIA MIG vGPU support for AI workloads (ie A100, H100, H200). CSCS runs 500 nodes with just 2 engineers and cuts infrastructure management time 80%. Manage VMs and containers from one interface…

### [SUSE AI Factory](https://www.suse.com/products/ai/)
- 来源：SUSE AI 产品页 / `suse-virtualization/suse-ai-factory-…-a49b82ac.txt` / 抓取 2026-09-03T00:41:37Z
> Compare the open foundation of SUSE AI Factory with the enhanced performance of SUSE AI Factory with NVIDIA. | Standardized, automated software stack | Natively embedded NVIDIA AI Enterprise | | Pre-validated blueprints for key use cases | Accelerated inference & agentic control (NVIDIA NIM & NeMo) | | Sandbox to production parity | Unified "single hand to shake" support for both NVIDIA and SUSE components | Seamlessly promote AI workloads from development to production without environmental drift across core data centers, cloud, or edge. Eliminate the guesswork and technical debt of building custom AI stacks from scratch. Deploy with confidence using our opinionated, pre-validated blueprints for high-demand use cases. Bridge the AI Ops divide with continuous workflows and pipelines to build and evolve AI workloads. Fluidly scale from ClickOps simplicity to industrial GitOps. Keep your proprietary data and core intellectual property strictly under your control with a predictable economic model. Eliminate unbounded public cloud billing and egress fees. Access free technical courses…

## 对基础档案的影响
- **版本**：商业品牌 SUSE Virtualization **1.8**；上游 Harvester **1.9.0-rc7**（未 GA）。二者版本号不要混写。
- **差异化**：反捆绑、可组合、两节点边缘 HA。
- **AI**：AI Factory 是上层，不等于 Harvester 发行说明。