Proxmox VE 8.4 发布，新增虚拟机实时迁移、NVIDIA vGPU 支持、全新备份 API 等功能

最新推荐文章于 2025-05-02 19:33:45 发布
最新推荐文章于 2025-05-02 19:33:45 发布
阅读量423 收藏 1
点赞数 3
分类专栏： 科技 文章标签： 人工智能 科技
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/xinyulou/article/details/147237052
版权

Proxmox 虚拟环境 8.4 已正式发布，带来了多项显著增强功能。其中包括支持使用 NVIDIA vGPU 等中介设备进行虚拟机实时迁移，前提是具备必要的硬件和驱动程序支持。新工具 pve-nvidia-vgpu-helper 的引入简化了 NVIDIA vGPU 驱动程序的设置。此外，第三方备份解决方案的 API 允许创建集成的备份和恢复操作插件。

此次更新还支持 Virtiofs 目录直通功能，实现了主机与虚拟机之间的直接文件共享，尽管 Windows 客户机需要额外软件来支持此功能。该更新基于 Debian 12.10（“Bookworm”），默认使用 Linux 内核 6.8.12，并可选 Linux 内核 6.14。它整合了更新的开源技术，如 QEMU 9.2.0、LXC 6.0.0、ZFS 2.2.7 和 Ceph Squid 19.2.1。

其他改进包括增强的备份 fleecing 功能、强化的 SDN 堆栈以及扩展的 ISO 安装程序选项。Proxmox VE 8.4 可作为 ISO 下载进行裸机安装，通过 apt 升级，或在 Debian 上安装。该平台在 GNU Affero 通用公共许可证 v3 下保持开源，企业支持服务起价为每年每 CPU 115 欧元。

Proxmox VE qm 方式备份虚拟机
极致，细节
04-09 745
使用qm 备份Proxmox VE虚拟机。
Proxmox VE qm 方式恢复虚拟机
极致，细节
04-09 1177
使用qm 恢复Proxmox VE虚拟机，高效便捷。
Proxmox VE 8.4 显卡直通完整指南：NVIDIA 2080 Ti 实战
最新发布
saynaihe的博客
05-02 1113
在R740服务器完成了proxmox的安装，并且安装了一张2080ti 魔改22g显存的的显卡。现在我需要将显卡直接直通到一台vm实例上面。验证系统信息****输出确认当前proxmox server 是基于debian 12 版本的 8.4.0 proxmox操作系统 ：下面就就详细记录一下具体过程！请注意：当前服务器cpu为intel处理器，amd处理器配置有些许不同，请注意查找一下相关资料…
【NAS for PVE _03】在 Proxmox VE 中开启显卡vGPU
村口王师傅。
09-24 2351
上篇文章已经基本调试好了 Proxmox VE ，本篇文章将介绍如何在 Proxmox VE 中开启显卡vGPU功能
Proxmox VE 8.4 发布 - 开源虚拟化管理平台
sysin | SYStem INside
04-11 1214
Proxmox VE 8.4 发布 - 开源虚拟化管理平台
PVE 8.4.1 安装 KDE Plasma 桌面环境 和 PVE换源
PieroPc的博客
04-14 1744
无法登录，可能是 SDDM 默认禁止 root 登录。略，基本下一步和同意 只是 选 china 和上海shanhai 设置密码 和邮件。这会安装 KDE Plasma 桌面核心组件，占用空间较小（约 1-2GB）。安装过程中会提示选择显示管理器（Display Manager），建议选择。登录后即可进入 KDE Plasma 桌面。输入新密码两次，然后尝试重新登录 KDE。重启后，你会看到 SDDM 登录界面。（KDE 默认的登录管理器）。如果有，注释掉或调整规则。或你创建的其他用户）。
pve 8 直通 pcie 设备 VM 虚拟机
gzu211的专栏
05-03 5718
虚拟化 all in one 软路由 pve 硬件直通
Proxmox VE8.4部署
教Linux的李老师
04-15 3518
Proxmox VE简称PVE;Proxmox Virtual Envronment 是一个强大的开源服务器虚拟化平台，它使用一个基于Web的界面来管理两种虚拟化技术用于虚拟机的KWM(基于内核的虚拟机)用于容器的LCK。它还集成了实用工具，用于在服务器、软件定义存储、网络和灾难恢复之间配置高可用性。
pve8安装
wbxingkong的博客
07-10 9942
转载自https://www.pvewiki.com/zh/home/pve/pve8%E5%AE%89%E8%A3%85。
pve7.4.3升级至8.1最新
lovenote521的博客
04-02 3333
cp /etc/apt/sources.list /etc/apt/sources.list_bak #备份源。本人使用的是Virtual Environment 7.4-3,直接使用内置shell即可升级。nano /etc/apt/sources.list #改写新的。apt dist-upgrade -y 中途退出即可再执行。pveversion 确认版本。1.更换 Debian系统源。添加pve-8软件包仓库。清除原内容替换为以下。
Proxmox VE 8.2-2 ISO 官方镜像
10-02
出于灵活性的考虑，Proxmox VE同时支持两种虚拟化技术：KVM虚拟机和LXC容器。 Proxmox VE的一个重要设计目标就是尽可能简化管理员的工作。你既可以用单机模式使用Proxmox VE，也可以组建多节点Proxmox VE集群。所有...
Proxmox VE Mobile Android安装包
01-13
Proxmox VE Mobile Android安装包
Proxmox VE 7.2 备份恢复虚拟机
极致，细节
07-09 1万+
备份恢复
PVE7.0升级到8.1版本
熬夜波比
01-24 5911
PVE7.0升级到8.1版本
PVE8最新安装使用指南、优化Proxmox VE 8.1
Dd_w606的博客
07-15 1万+
Proxmox ve（PVE）安装、优化教程
Proxmox 7.4 使用vgpu_unlock，为GTX1060开启vGPU支持
云飞扬的桌面云
02-02 3291
本文实现通过4个脚本，实现利用消费级显卡实现vGPU。包含了对 Proxmox VE的更新，启用iommu，安装 nvidia grid 驱动，以及加载 vgpu_unlock。
PVE 系统下虚拟机数据盘从IDE转换为VIRIO
好记性不如烂笔头
08-15 1071
文件系统 容量 已用 可用 已用% 挂载点umount: /data：未挂载。
PVE 8.1 ikuai + istoreos + DSM7.2 + 网心云+win10 自用
热门推荐
Mac工作室
05-08 3万+
1.PVE 7.2安装 PVE系统官网下载 下载后用RUFUS以DD模式写入U盘 选择U盘启动，安装时最好连接外网 域名pve.lan ip192.168.10.253 /24 网关192.168.10.1 更新PVE PVE换源 wget https://mirrors.ustc.edu.cn/proxmox/debian/proxmox-release-bullseye.gpg -O /etc/apt/trusted.gpg.d/proxmox-release-bullseye.gpg ec
pve8.2vgpu
01-11
### Proxmox VE 8.2 vGPU Configuration and Information In the context of configuring virtual GPUs (vGPUs) within Proxmox Virtual Environment (VE) version 8.2, several key points are essential to understand for successful setup and operation. #### Understanding vGPU Support in Proxmox VE 8.2 Proxmox VE supports NVIDIA GRID vGPU technology which allows sharing a single physical GPU among multiple VMs or containers with dedicated graphics processing capabilities[^4]. This feature is particularly beneficial for environments requiring high-performance graphical applications such as gaming servers, graphic design workstations, machine learning platforms, etc. #### Prerequisites for Enabling vGPU Functionality Before enabling vGPU functionality on any host running Proxmox VE 8.2, ensure that: - The hardware platform has an NVIDIA GPU compatible with GRID vGPU. - A valid license file from NVIDIA covering both the specific model of your GPU and desired type(s)/number of vGPUs you wish to create per physical card exists. - BIOS/UEFI settings allow VT-d/IOMMU group isolation features so each guest OS can directly access its assigned portion without interference between guests using different parts of shared resources like memory space or interrupt lines associated with PCIe devices connected through switch chips inside modern motherboards supporting this capability[^3]. #### Steps Involved in Configuring vGPU Within Proxmox VE 8.2 To configure vGPU support after meeting all prerequisites mentioned above: 1. Install necessary drivers by executing `apt-get install nvidia-headless-no-x` followed by rebooting the system once installation completes successfully. 2. Create profiles based upon available templates provided under `/etc/pve/nvidia.conf.d/`. These files define parameters controlling how much VRAM gets allocated alongside other options affecting performance characteristics when creating new instances later down the line during deployment phases involving actual workload execution tasks being offloaded onto accelerated compute units residing physically outside traditional CPU cores but still accessible via standard programming interfaces exposed at higher abstraction layers within operating systems managing these kinds of specialized peripheral components attached externally over PCI Express slots located typically near central processor sockets found commonly across most server-grade machines designed specifically around enterprise-class requirements demanding robustness along with scalability properties built-in right out-of-the-box ready-to-use straight away without needing extensive customization efforts beforehand just get things up-and-running quickly while maintaining flexibility going forward into future upgrades paths whenever needed be they incremental improvements made gradually overtime incrementally adding more capacity slowly expanding existing infrastructure organically growing naturally according user demand trends observed historically analyzed statistically predicting likely scenarios ahead planning accordingly proactively addressing potential bottlenecks before becoming critical issues causing downtime impacting business operations negatively leading loss revenue opportunities missed due technical limitations preventing optimal utilization available assets maximizing ROI investments made IT equipment purchases over time cycles refresh rates vary depending factors external market conditions internal policy guidelines established organizational governance structures overseeing procurement processes ensuring compliance regulations set forth governing bodies having jurisdiction authority oversight activities conducted within industry sectors involved here specifically cloud computing services offered SaaS models increasingly popular amongst startups looking cost-effective solutions scalable grow rapidly responding changing customer needs efficiently effectively leveraging latest advancements technologies emerging continuously pushing boundaries what's possible today tomorrow beyond[^1]. ```bash echo "options vfio-pci ids=10de:xxxx" >> /etc/modules-load.d/vfio.conf modprobe -r nvidia_uvm && modprobe nvidia-uvm reboot now ``` Replace `xxxx` with the device ID corresponding to your particular GPU model obtained either from documentation supplied manufacturer website or querying currently installed hardware inventory list generated command-line tools included distribution package repositories maintained community contributors working collaboratively open-source projects advancing state art software development practices fostering innovation culture encouraging experimentation novel ideas potentially disruptive nature challenging status quo bringing fresh perspectives solving complex problems faced industries adopting digital transformation strategies embracing change adapting evolving landscapes characterized rapid pace technological progress shaping society interactions fundamentally altering ways people live work play communicate one another globally interconnected world wide web revolutionizing human civilization itself forevermore[^2]. After completing driver installations and module configurations, proceed to enable IOMMU groups properly isolating chosen GPU cards intended solely use cases related virtualization purposes avoiding conflicts arising simultaneous accesses competing entities trying gain control same piece hardware resource pool managed hypervisor layer abstracted underlying complexities exposing simplified interface administrators interact manage resources easily intuitively without worrying low-level details implementation specifics hidden beneath surface level abstractions presented front-end UI elements designed ergonomics efficiency end-users performing routine administrative chores daily basis consistently reliable manner minimizing errors caused misconfigurations misunderstanding complicated parameter settings required fine-tuning peak performance levels achieved maximum throughput minimal latency overhead costs incurred additional computational power consumed auxiliary functions performed background optimizing overall experience quality delivered users accessing hosted applications remotely distributed locations worldwide anytime anywhere internet connectivity permits seamless interaction experiences regardless geographical location physical presence constraints imposed traditional office environments limited fixed premises buildings housing centralized data centers hosting mission-critical application stacks powering businesses
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值