51工具盒子
写在前面的话 {#i}
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 {#i-2}
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 {#linux-guest-vm}
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。
Debian或Ubuntu Linux安装
首先，运行下列apt命令/apt-get命令：
[sourcecode language="plain"]
$ sudo apt-get install virt-what
[/sourcecode]
或者
[sourcecode language="plain"]
$ sudo apt install virt-what
[/sourcecode]
在RHEL/CentOS/Scientific Linux VM中安装virt-what
我们可以使用yum命令完成安装：
[sourcecode language="plain"]
$ sudo yum install virt-what
[/sourcecode]
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：
[sourcecode language="plain"]
$ sudo dnf install virt-what
[/sourcecode]
Suse/OpenSUSE安装
运行zypper命令：
[sourcecode language="plain"]
sudo zypper in virt-what
[/sourcecode]
判断远程服务器是否为虚拟机环境 {#i-3}
我们可以直接输入并运行下列命令：
[sourcecode language="plain"]
$ $ sudo virt-what
xen
xen-domU
[/sourcecode]
输出结果
[sourcecode language="plain"]
$ $ sudo virt-what
kvm
[/sourcecode]
其他输出结果
[sourcecode language="plain"]
$ $ sudo virt-what
lxc
[/sourcecode]
下面给出aws反回结果
[sourcecode language="plain"]
$ sudo virt-what
xen
xen-hvm
aws
[/sourcecode]
其他可能的值 {#i-4}
[sourcecode language="plain"]
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；
[/sourcecode]
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 {#i-5}
下面给出的是脚本的源代码：
[sourcecode language="plain"]
#!/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 "$?" in
-
;; # not virtual
-
Could be some sort of virt, or could just be a bit slow.
echo virt
esac
fi
fi
Check for QEMU/KVM.
Parallels exports KVMKVMKVM leaf, so skip this test if we've already
seen that it's Parallels. Xen uses QEMU as the device model, so
skip this test if we know it is Xen.
if ! "$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
[/sourcecode]
如何使用dmidecode命令来寻找相同的信息 {#dmidecode}
配合Bash for循环并使用dmidecode命令判断目标虚拟化技术：
[sourcecode language="plain"]
for i in system-manufacturer system-product-name
do
sudo dmidecode -s $i
done
[/sourcecode]
样本输出
[sourcecode language="plain"]
Red Hat
KVM
[/sourcecode]
其他用来判断虚拟化技术的命令 {#i-6}
我们还可以使用systemd-detect-virt命令来检测虚拟化环境：
[sourcecode language="plain"]
systemd-detect-virt
[/sourcecode]
资源获取 {#i-7}
1、virt-what：【点我获取】
2、Bash for循环：【参考文档】
3、yum命令：【参考文档】
4、apt命令：【参考文档】
5、apt-get命令：【参考文档】
