写在前面的话

如果我有一台基于云的虚拟机，并且运行了Ubuntu或CentOS的Linux服务器系统，而这些设备可能是虚拟化的，也可能不是虚拟化的，那我们应该如何去判断这台Linux Guest系统设备使用的是哪种虚拟化技术（VMWARE/ KVM/ XEN/ VirtualBox/ Container/lxc/Hyper-V等等）呢？那我们如何才能识别一台CentOS 7 Linux VPS所使用的虚拟化技术类型呢？

技术分析

我们需要使用virt-what程序（一个Shell脚本）来检测目标设备系统所使用的虚拟化技术类型，它可以打印出每一条关于目标设备虚拟化技术的相关信息。在这篇文章中，我们将介绍如何去判断Linux Guest VM虚拟化技术类型。

判断Linux Guest VM虚拟化技术类型

正如我们刚才说的，我们需要使用virt-what Linux命令来判断一个程序当前是否在一台虚拟机设备上运行。该脚本支持各种虚拟机管理程序，接下来，我们看看如何在不同的Linux发行版系统中安装virt-what。

virt-what：https:///~rjones/virt-what/

Debian或Ubuntu Linux安装

首先，运行下列apt命令/apt-get命令：

$ sudo apt-get install virt-what

或者

$ sudo apt install virt-what

在RHEL/CentOS/Scientific Linux VM中安装virt-what

我们可以使用yum命令完成安装：

$ sudo yum install virt-what

Fedora Linux VM安装virt-what来检测当前环境是否为虚拟机环境

执行dnf命令：

$ sudo dnf install virt-what

Suse/OpenSUSE安装

运行zypper命令：

sudo zypper in virt-what

判断远程服务器是否为虚拟机环境

我们可以直接输入并运行下列命令：

$ $ sudo virt-what xen xen-domU

输出结果：

$ $ sudo virt-what kvm

其他输出结果：

$ $ sudo virt-what lxc

下面给出的是AWS返回的结果：

$ sudo virt-what xen xen-hvm aws

其他可能的值

1、hyperv : 这是微软的Hyper-V管理程序；

2、parallels : 访客用户正在Parallels虚拟平台（Parallels Desktop、Parallels Server）中运行；

3、powervm_lx86 : 访客用户正在IBM powervm lx86 Linux/x86模拟器中运行；

4、qemu : 这是使用软件模拟技术的QEMU管理程序；

5、virtualpc : 访客用户正在微软VirualPC上运行；

6、xen-hvm : 这是一个Xen Guest完全虚拟化平台（HVM）；

7、uml : 这是一个用户模式Linux（UML）客户机；

8、openvz : 访客用户正在OpenVZ或Virtuozzo容器中运行；

9、linux_vserver : 此进程正在Linux VServer容器中运行；

10、ibm_systemz : 这是一个IBM SystemZ硬件分区系统；

如果没有任何输出结果，那意味着我们的程序可能是在裸机上运行的，或者程序是在一种我们无法检测到的虚拟机系统中运行的。

检测脚本源代码

下面给出的是脚本的源代码：

#!/bin/sh - # virt-what. Generated from virt-what.in by configure. # Copyright (C) 2008-2017 Red Hat Inc. # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program; if not, write to the Free Software # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. # 'virt-what' tries to detect the type of virtualization being # used (or none at all if we're running on bare-metal). It prints # out one or more lines each being a 'fact' about the virtualization. # # Please see also the manual page virt-what(1). # This script should be run as root. # # The following resources were useful in writing this script: # . http://dmo.ca/blog/detecting-virtualization-on-linux/ # Do not allow unset variables, and set defaults. set -u root='' skip_qemu_kvm=false skip_lkvm=false VERSION="1.20" have_cpuinfo () { test -e "${root}/proc/cpuinfo" } use_sysctl() { # Lacking /proc, on some systems sysctl can be used instead. OS=$(uname) || fail "failed to get operating system name" [ "$OS" = "OpenBSD" ] } fail () { echo "virt-what: $1" >&2 exit 1 } usage () { echo "virt-what [options]" echo "Options:" echo " --help Display this help" echo " --version Display version and exit" exit 0 } # Handle the command line arguments, if any. while test $# -gt 0; do case "$1" in --help) usage ;; --test-root=*) # Deliberately undocumented: used for 'make check'. root=$(echo "$1" | sed 's/.*=//') shift 1 test -z "$root" && fail "--test-root option requires a value" ;; -v|--version) echo "$VERSION"; exit 0 ;; --) shift; break ;; *) fail "unrecognized option '$1'";; esac done test $# -gt 0 && fail "extra operand '$1'" # Add /sbin and /usr/sbin to the path so we can find system # binaries like dmidecode. # Add /usr/libexec to the path so we can find the helper binary. prefix=/usr/local exec_prefix=${prefix} PATH="${root}${exec_prefix}/libexec:${root}/sbin:${root}/usr/sbin:${PATH}" export PATH # Check we're running as root. EFFUID=$(id -u) || fail "failed to get current user id" if [ "x$root" = "x" ] && [ "$EFFUID" -ne 0 ]; then fail "this script must be run as root" fi # Try to locate the CPU-ID helper program CPUID_HELPER=$(which virt-what-cpuid-helper 2>/dev/null) if [ -z "$CPUID_HELPER" ] ; then fail "virt-what-cpuid-helper program not found in \$PATH" fi # Many fullvirt hypervisors give an indication through CPUID. Use the # helper program to get this information. cpuid=$(virt-what-cpuid-helper) # Check for various products in the BIOS information. # Note that dmidecode doesn't exist on all architectures. On the ones # it does not, then this will return an error, which is ignored (error # message redirected into the $dmi variable). dmi=$(LANG=C dmidecode 2>&1) # Architecture. # Note for the purpose of testing, we only call uname with -m option. arch=$(uname -m | sed -e 's/i.86/i386/' | sed -e 's/arm.*/arm/') # Check for VMware. # cpuid check added by Chetan Loke. if [ "$cpuid" = "VMwareVMware" ]; then echo vmware elif echo "$dmi" | grep -q 'Manufacturer: VMware'; then echo vmware fi # Check for Hyper-V. # http://blogs.msdn.com/b/sqlosteam/archive/2010/10/30/is-this-real-the-metaphysics-of-hardware-virtualization.aspx if [ "$cpuid" = "Microsoft Hv" ]; then echo hyperv fi # Check for VirtualPC. # The negative check for cpuid is to distinguish this from Hyper-V # which also has the same manufacturer string in the SM-BIOS data. if [ "$cpuid" != "Microsoft Hv" ] && echo "$dmi" | grep -q 'Manufacturer: Microsoft Corporation'; then echo virtualpc fi # Check for VirtualBox. # Added by Laurent Léonard. if echo "$dmi" | grep -q 'Manufacturer: innotek GmbH'; then echo virtualbox fi # Check for bhyve. if [ "$cpuid" = "bhyve bhyve " ]; then echo bhyve elif echo "$dmi" | grep -q "Vendor: BHYVE"; then echo bhyve fi # Check for OpenVZ / Virtuozzo. # Added by Evgeniy Sokolov. # /proc/vz - always exists if OpenVZ kernel is running (inside and outside # container) # /proc/bc - exists on node, but not inside container. if [ -d "${root}/proc/vz" -a ! -d "${root}/proc/bc" ]; then echo openvz fi # Check for LXC containers # http://www.freedesktop.org/wiki/Software/systemd/ContainerInterface # Added by Marc Fournier if [ -e "${root}/proc/1/environ" ] && cat "${root}/proc/1/environ" | tr '\000' '\n' | grep -Eiq '^container='; then echo lxc fi # Check for Linux-VServer if test -e "${root}/proc/self/status" \ && cat "${root}/proc/self/status" | grep -q "VxID: [0-9]*"; then echo linux_vserver if grep -q "VxID: 0$" "${root}/proc/self/status"; then echo linux_vserver-host else echo linux_vserver-guest fi fi # Check for UML. # Added by Laurent Léonard. if have_cpuinfo && grep -q 'UML' "${root}/proc/cpuinfo"; then echo uml fi # Check for IBM PowerVM Lx86 Linux/x86 emulator. if have_cpuinfo && grep -q '^vendor_id.*PowerVM Lx86' "${root}/proc/cpuinfo" then echo powervm_lx86 fi # Check for Hitachi Virtualization Manager (HVM) Virtage logical partitioning. if echo "$dmi" | grep -q 'Manufacturer.*HITACHI' && echo "$dmi" | grep -q 'Product.* LPAR'; then echo virtage fi # Check for IBM SystemZ. if have_cpuinfo && grep -q '^vendor_id.*IBM/S390' "${root}/proc/cpuinfo"; then echo ibm_systemz if [ -f "${root}/proc/sysinfo" ]; then if grep -q 'VM.*Control Program.*KVM/Linux' "${root}/proc/sysinfo"; then echo ibm_systemz-kvm elif grep -q 'VM.*Control Program.*z/VM' "${root}/proc/sysinfo"; then echo ibm_systemz-zvm elif grep -q '^LPAR' "${root}/proc/sysinfo"; then echo ibm_systemz-lpar else # This is unlikely to be correct. echo ibm_systemz-direct fi fi fi # Check for Parallels. if echo "$dmi" | grep -q 'Vendor: Parallels'; then echo parallels skip_qemu_kvm=true fi # Check for oVirt/RHEV. if echo "$dmi" | grep -q 'Manufacturer: oVirt'; then echo ovirt fi if echo "$dmi" | grep -q 'Product Name: RHEV Hypervisor'; then echo rhev fi # Check for Xen. if [ "$cpuid" = "XenVMMXenVMM" ] && ! echo "$dmi" | grep -q 'No SMBIOS nor DMI entry point found, sorry'; then echo xen; echo xen-hvm skip_qemu_kvm=true elif [ -d "${root}/proc/xen" ]; then echo xen if grep -q "control_d" "${root}/proc/xen/capabilities" 2>/dev/null; then echo xen-dom0 else echo xen-domU fi skip_qemu_kvm=true skip_lkvm=true elif [ -f "${root}/sys/hypervisor/type" ] && grep -q "xen" "${root}/sys/hypervisor/type"; then # Ordinary kernel with pv_ops. There does not seem to be # enough information at present to tell whether this is dom0 # or domU. XXX echo xen elif [ "$arch" = "arm" ] || [ "$arch" = "aarch64" ]; then if [ -d "${root}/proc/device-tree/hypervisor" ] && grep -q "xen" "${root}/proc/device-tree/hypervisor/compatible"; then echo xen skip_qemu_kvm=true skip_lkvm=true fi elif [ "$arch" = "ia64" ]; then if [ -d "${root}/sys/bus/xen" -a ! -d "${root}/sys/bus/xen-backend" ]; then # PV-on-HVM drivers installed in a Xen guest. echo xen echo xen-hvm else # There is no virt leaf on IA64 HVM. This is a last-ditch # attempt to detect something is virtualized by using a # timing attack. virt-what-ia64-xen-rdtsc-test > /dev/null 2>&1 case "$?" in 0) ;; # not virtual 1) # Could be some sort of virt, or could just be a bit slow. echo virt esac fi fi # Check for QEMU/KVM. # # Parallels exports KVMKVMKVM leaf, so skip this test if we've already # seen that it's Parallels. Xen uses QEMU as the device model, so # skip this test if we know it is Xen. if ! "$skip_qemu_kvm"; then if [ "$cpuid" = "KVMKVMKVM" ]; then echo kvm elif [ "$cpuid" = "TCGTCGTCGTCG" ]; then echo qemu skip_lkvm=true elif echo "$dmi" | grep -q 'Product Name: KVM'; then echo kvm skip_lkvm=true elif echo "$dmi" | grep -q 'Manufacturer: QEMU'; then # The test for KVM above failed, so now we know we're # not using KVM acceleration. echo qemu skip_lkvm=true elif [ "$arch" = "arm" ] || [ "$arch" = "aarch64" ]; then if [ -d "${root}/proc/device-tree" ] && ls "${root}/proc/device-tree" | grep -q "fw-cfg"; then # We don't have enough information to determine if we're # using KVM acceleration or not. echo qemu skip_lkvm=true fi elif [ -d ${root}/proc/device-tree/hypervisor ] && grep -q "linux,kvm" /proc/device-tree/hypervisor/compatible; then # We are running as a spapr KVM guest on ppc64 echo kvm skip_lkvm=true elif use_sysctl; then # SmartOS KVM product=$(sysctl -n hw.product) if echo "$product" | grep -q 'SmartDC HVM'; then echo kvm fi else # This is known to fail for qemu with the explicit -cpu # option, since /proc/cpuinfo will not contain the QEMU # string. QEMU 2.10 added a new CPUID leaf, so this # problem only triggered for older QEMU if have_cpuinfo && grep -q 'QEMU' "${root}/proc/cpuinfo"; then echo qemu fi fi fi if ! "$skip_lkvm"; then if [ "$cpuid" = "LKVMLKVMLKVM" ]; then echo lkvm elif [ "$arch" = "arm" ] || [ "$arch" = "aarch64" ]; then if [ -d "${root}/proc/device-tree" ] && grep -q "dummy-virt" "${root}/proc/device-tree/compatible"; then echo lkvm fi fi fi # Check for Docker. if [ -f "${root}/.dockerinit" ]; then echo docker fi # Check ppc64 lpar, kvm or powerkvm # example /proc/cpuinfo line indicating 'not baremetal' # platform : pSeries # # example /proc/ppc64/lparcfg systemtype line # system_type=IBM pSeries (emulated by qemu) if [ "$arch" = "ppc64" ] || [ "$arch" = "ppc64le" ] ; then if have_cpuinfo && grep -q 'platform.**pSeries' "${root}/proc/cpuinfo"; then if grep -q 'model.*emulated by qemu' "${root}/proc/cpuinfo"; then echo ibm_power-kvm else # Assume LPAR, now detect shared or dedicated if grep -q 'shared_processor_mode=1' "${root}/proc/ppc64/lparcfg"; then echo ibm_power-lpar_shared else echo ibm_power-lpar_dedicated fi # detect powerkvm? fi fi fi # Check for OpenBSD/VMM if [ "$cpuid" = "OpenBSDVMM58" ]; then echo vmm fi # Check for LDoms if [ "${arch#sparc}" != "$arch" ] && [ -e "${root}/dev/mdesc" ]; then echo ldoms if [ -d "${root}/sys/class/vlds/ctrl" ] && \ [ -d "${root}/sys/class/vlds/sp" ]; then echo ldoms-control else echo ldoms-guest fi MDPROP="${root}/usr/lib/ldoms/mdprop.py" if [ -x "${MDPROP}" ]; then if [ -n "$($MDPROP -v iodevice device-type=pciex)" ]; then echo ldoms-root echo ldoms-io elif [ -n "$($MDPROP -v iov-device vf-id=0)" ]; then echo ldoms-io fi fi fi # Check for AWS. # AWS on Xen. if echo "$dmi" | grep -q 'Version: [0-9]\.[0-9]\.amazon'; then echo aws # AWS on baremetal or KVM. elif echo "$dmi" | grep -q 'Vendor: Amazon EC2'; then echo aws fi

如何使用dmidecode命令来寻找相同的信息

配合Bash for循环并使用dmidecode命令判断目标虚拟化技术：

for i in system-manufacturer system-product-name do sudo dmidecode -s $i done

样本输出：

Red Hat KVM

其他用来判断虚拟化技术的命令

我们还可以使用systemd-detect-virt命令来检测虚拟化环境：

systemd-detect-virt