KVM虚拟化

http://blog.chinaunix.net/xmlrpc.php?r=blog/article&uid=20776139&id=3279579

KVM笔记！
kvm原理学习：


Linux as a Hypervisor：
By adding virtualization capabilities to a standard Linux kernel, we can enjoy all
the fine-tuning work that has gone (and is going) into the kernel, and bring that
benefit into a virtualized environment. Under this model, every virtual machine is a regular Linux process scheduled by the standard Linux scheduler. Its memory is allocated by the Linux memory allocator, with its knowledge of NUMA and integration into the scheduler.


A normal Linux process has two modes of execution: kernel and user. Kvm adds a third mode: guest mode (which has its own kernel and user modes, but these do not interest the hypervisor at all).
三种工作模式如下：


 

kvm Components：
The simplicity of kvm is exemplified by its structure; there are two components:
● A device driver for managing the virtualization hardware; this driver exposes its capabilities via a character device /dev/kvm
● A user-space component for emulating PC hardware; this is a lightly modified qemu process,QEMU is a well known processor emulator written by French computer wizard Fabrice  Bellard.
见附件！




What is the difference between kvm and QEMU?  ##kvm和QEMU的不同点
Qemu uses emulation; kvm uses processor extensions for virtualization.




What is the difference between kvm and Xen?   ##xen和kvm的不同点 
Xen is an external hypervisor; it assumes control of the machine and divides resources among guests. On the other hand, kvm is part of Linux and uses the regular Linux 
scheduler and memory management. This means that kvm is much smaller and simpler to use.




On the other hand, Xen supports both full virtualization and a technique called paravirtualization, which allows better performance for modified guests. kvm does not 
at present support paravirtualization.


KVM 所使用的方法是通过简单地加载内核模块而将 Linux 内核转换为一个系统管理程序。这个内核模块导出了一个名为 /dev/kvm 的设备，它可以启用内核的客户模式(除了传统的内核模式和用户模式)。有了 /dev/kvm 设备，VM 使自己的地址空间独立于内核或运行着的任何其他 VM 的地址空间。设备树(/dev)中的设备对于所有用户空间进程来说都是通用的。但是每个打开 /dev/kvm 的进程看到的是不同的映射(为了支持 VM 间的隔离)。KVM安装 KVM 之后，您可以在用户空间启动客户操作系统。每个客户操作系统都是主机操作系统(或系统管理程序)的一个单个进程。 


KVM 是指基于 Linux 内核的虚拟机（Kernel-based Virtual Machine），2006年10月，由以色列的 Qumranet 组织开发的一种新的“虚拟机”实现方案。 2007年2月发布的 Linux 2.6.20内核第一次包含了KVM 。增加KVM到Linux内核是 Linux发展的一个重要里程碑，这也是第一个整合到Linux主线内核的虚拟化技术。
KVM 在标准 Linux 内核中增加了虚拟技术，从而我们可以通过优化的内核来使用虚拟技术。在 KVM 模型中，每一个虚拟机都是一个由 Linux 调度程序管理的标准进程，你可以在用户空间启动客户机操作系统。
一个普通的 Linux 进程有两种运行模式：内核和用户。 KVM 增加了第三种模式：客户模式（有自己的内核和用户模式）


作为VMM，KVM分为两部分，分别是运行于kernel模式的KVM和运行于User模式下的Qemu模块，这里的kernel模式和User模式，实际上指的是VMX根模式下的特权级0和特权级3，KVM将虚拟机所在的运行模式称为Gues模式 当一起工作的时候，KVM管理CPU和MEM的访问，QEMU仿真硬件资源（硬盘，声卡，USB，等等）.而qemu-kvm是为了针对KVM专门做了修改和优化的QEMU分支.




 
 KVM的所有IO虚拟化工作时借助Qemu完成的，显著降低了实现的工作量！这是KVM的优势之一！ ##看下图即可理解
KVM是Linux的一个模块。可以用modprobe去加载KVM模块。加载了模块后，才能进一步通过其他工具创建虚拟机。但仅有KVM模块是远远不够的，因为用户无法直接控制内核模块去作事情：还必须有一个用户空间的工具才行。这个用户空间的工具，开发者选择了已经成型的开源虚拟化软件QEMU。说起来QEMU也是一个虚拟化软件。它的特点是可虚拟不同的CPU。比如说在x86的CPU上可虚拟一个Power的CPU，并可利用它编译出可运行在Power上的程序。KVM使用了QEMU的一部分，并稍加改造，就成了可控制KVM的用户空间工具了






   


KVM 客户机网络连接有两种方式：
用户网络（User Networking）：让虚拟机访问主机、互联网或本地网络上的资源的简单方法，但是不能从网络或其他的客户机访问客户机，性能上也需要大的调整。
虚拟网桥（Virtual Bridge）：这种方式要比用户网络复杂一些，但是设置好后客户机与互联网，客户机与主机之间的通信都很容易。


启动一台客户机后，主机上多出了一个 tap0 虚拟网络设备，这就是 qemu-kvm 为客户机虚拟的 TAP 网络设备。查看网桥可以看出 tap0 加入了网桥 br0 。客户机就是通过网桥访问的外网。


1：[root@h4 ~]# sh install.sh 
ERROR    Guest name 'linuxvm02' is already in use.
ERROR    A name is required for the virtual machine. (use --prompt to run interactively)
[root@h4 ~]# cat install.sh 
virt-install -n linuxvm02 -r 1024 --vcpus=1 -l http://192.168.1.66 --nographics --os-type=linux --os-variant=rhel5 -f /data/linuxvm02.img -s 20 -w bridge:br0   --extra-args='console=tty0 console=ttyS0,115200n8'   --connect qemu:///system
显然已经安装过了linuxvm02，如果是上一次没有安装成功，删除这个虚拟机，然后重新运行该脚本即可!
删除虚拟机有两种方式：
1.1：图形界面方式，在物理机器图形界面下，运行virt-manager ，找到自己要删除的vm，如果是active状态，先将其关闭，然后点delete按钮删除
1.2: virsh undefine linuxvm02 即可从/etc/libvirt/qemu下将配置文件删除。然后手动删除linuxvm02对应的img文件即可


2:centos下安装kvm


centos5下安装：
[root@h4 CentOS]# yum grouplist | grep  -i kvm
   KVM
[root@h4 CentOS]# cat /etc/issue
CentOS release 5.8 (Final)
Kernel \r on an \m
显然centos5下要
# yum -y groupinstall KVM ##查看KVM组的信息
必需的软件包：
Required Packages
You must install the following packages:
kmod-kvm : kvm kernel module(s)
kvm : Kernel-based Virtual Machine
kvm-qemu-img : Qemu disk image utility
kvm-tools : KVM debugging and diagnostics tools
python-virtinst : Python modules and utilities for installing virtual machines
virt-manager : Virtual Machine Manager (GUI app, to install and configure VMs)
virt-viewer: Virtual Machine Viewer (another lightweight app to view VM console and/or install VMs)
bridge-utils : Utilities for configuring the Linux Ethernet bridge (this is recommended for KVM networking)
 [root@h4 CentOS]# yum groupinfo KVM
Loaded plugins: fastestmirror, security
Setting up Group Process
Loading mirror speeds from cached hostfile
 * base: centos.ustc.edu.cn
 * extras: centos.ustc.edu.cn
 * updates: centos.ustc.edu.cn


Group: KVM
 Description: Virtualization Support with KVM
 Mandatory Packages:
   etherboot-zroms
   etherboot-zroms-kvm
   kmod-kvm
   kvm
   kvm-qemu-img
 Default Packages:
   Virtualization-en-US
   libvirt
   virt-manager
   virt-viewer
   virt-who
 Optional Packages:
   etherboot-pxes
   etherboot-roms
   etherboot-roms-kvm
   gpxe-roms-qemu
   iasl
   kvm-tools
   libcmpiutil
   libvirt-cim
   qspice
   qspice-libs-devel
老外的一篇文章：写的挺好的！
A Note About libvirt：
libvirt is an open source API and management tool for managing platform virtualization. It is used to manage Linux KVM and Xen virtual machines through graphical interfaces such as Virtual Machine Manager and higher level tools such as oVirt. See the official website for more information.
A Note About QEMU：
QEMU is a processor emulator that relies on dynamic binary translation to achieve a reasonable speed while being easy to port on new host CPU architectures. When used as a virtualizer, QEMU achieves near native performances by executing the guest code directly on the host CPU. QEMU supports virtualization when executing under the Xen hypervisor or using the KVM kernel module in Linux. When using KVM, QEMU can virtualize x86, server and embedded PowerPC, and S390 guests. See the official website for more information.
A Note About Virtio Drivers：
  Virtio is paravirtualized drivers for kvm/Linux. With this you can can run multiple virtual machines running unmodified Linux or Windows VMs. Each virtual machine has private virtualized hardware a network card, disk, graphics adapter, etc. According to Redhat:
  Para-virtualized drivers enhance the performance of fully virtualized guests. With the para-virtualized drivers guest I/O latency decreases and throughput increases to near bare-metal levels. It is recommended to use the para-virtualized drivers for fully virtualized guests running I/O heavy tasks and applications.
Host Operating System
Your main operating system such as CentOS or RHEL is known as host operating system. KVM is a Linux kernel module that enables a modified QEMU program to use hardware virtualization. You only need to install KVM under host operating systems.
KVM Domains
It is nothing but a guest operating system running under host operating system. Each kvm domain must have a unique name and ID (assigned by system).
Guest Operating Systems
KVM supports various guest operating systems such as
MS-Windows 2008 / 2000 / 2003 Server
MS-Windows 7 / Vista / XP
FreeBSD
OpenBSD
Sun Solaris
Various Linux distributions.
NetBSD
MINIX
QNX
MS DOS
FreeDOS
Haiku
Amiga Research OS
Important Configuration And Log Files (Directories) Location
The following files are required to manage and debug KVM problems:
/etc/libvirt/ - Main configuration directory.
/etc/libvirt/qemu/ - Virtual machine configuration directory. All xml files regarding VMs are stored here. You can edit them manually or via virt-manager.
/etc/libvirt/qemu/networks/ - Networking for your KVM including default NAT. NAT is only recommended for small setup or desktops. I strongly suggest you use bridged based networking for performance.
/etc/libvirt/qemu/networks/default.xml - The default NAT configuration used by NAT device virbr0.
/var/log/libvirt/ - The default log file directory. All VM specific logs files are stored here.
/etc/libvirt/libvirtd.conf - Master libvirtd configuration file.
/etc/libvirt/qemu.conf - Master configuration file for the QEMU driver.
TCP/UDP Ports
By default libvirt does not opens any TCP or UDP ports. However, you can configure the same by editing the /etc/libvirt/libvirtd.conf file. Also, VNC is configured to listen on 127.0.0.1 by default. To make it listen on all public interfaces, edit /etc/libvirt/qemu.conf file.
架构：




(Fig.01: Our sample server setup - you need to scroll to see complete diagram)
Where,
Host Configuration
OS - RHEL / CentOS v5.5 is our host operating system.
Host has two interface eth0 and eth1
LAN - eth0 with private ip
Internet - eth1 with public IPv4/IPv6 address.
Disk - 73x4 - 15k SAS disk in RAID 10 hardware mode. All VMs are stored on same server (later I will cover SAN/NFS/NAS configuration with live migration).
RAM - 16GB ECC
CPU - Dual core dual Intel Xeon CPU L5320 @ 1.86GHz with VT enabled in BIOS.
Virtual Machine Configuration
Bridged mode networking (eth0 == br0 and eth1 == br1) with full access to both LAN and Internet.
Accelerator virtio drivers used for networking (model=virtio)
Accelerator virtio drivers for disk (if=virtio) and disk will show up as /dev/vd[a-z][1-9] in VM.
Various virtual machines running different guest operating systems as per requirements.




 








3：确定KVM模块是否被加载
在启动完成后，你应该可以看到KVM内核模块已经被加载了：
[root@h5 ~]# lsmod |grep kvm 
kvm_intel 85256 1 
kvm 224800 2 ksm,kvm_intel
可以通过下面的命令看看KVM是不是真正运行了：
[root@h5 ~]# virsh -c qemu:///system list 
Id Name State ----------------------------------


4：设置某个虚拟机开机随系统启动：


[root@host1 qemu]# chkconfig --list | grep libvirtd  ##确认libvirtd开机随系统启动
libvirtd 0:off 1:off 2:on 3:on 4:on 5:on 6:off


 
[root@host1 ~]# virsh autostart linuxvm02
Domain linuxvm02 marked as autostarted
[root@host1 ~]# cd /etc/libvirt/
[root@host1 libvirt]# ls
cim libvirtd.conf nwfilter qemu qemu.conf storage
[root@host1 libvirt]# cd qemu
[root@host1 qemu]# ls
autostart kvm01.xml linuxvm02.xml networks
[root@host1 qemu]# cd autostart/
[root@host1 autostart]# ls  ##执行上述命令后，会在autostart目录下生成一个连接文件！
linuxvm02.xml
[root@host1 autostart]# ll
total 4
lrwxrwxrwx 1 root root 31 Aug 9 16:47 linuxvm02.xml -> /etc/libvirt/qemu/linuxvm02.xml
重启机器后测试，linuxvm02 已经随系统启动！
将刚刚随系统启动的虚拟机设置成禁止随系统启动：
[root@host1 qemu]# cd autostart/
[root@host1 autostart]# ls  ##显然有该文件
linuxvm02.xml
[root@host1 autostart]# ll
total 4
lrwxrwxrwx 1 root root 31 Aug  9 16:47 linuxvm02.xml -> /etc/libvirt/qemu/linuxvm02.xml
[root@host1 autostart]# virsh autostart linuxvm02 --disable
Domain linuxvm02 unmarked as autostarted 


[root@host1 autostart]# ls ##执行该命令后删除了 linuxvm02.xml软连接
[root@host1 autostart]# ll
5：kvm下每启动一个虚拟机，top里对应多出一个qemu-kvm进程！
解释如下：Under KVM's model, every virtual machine is a regular Linux process scheduled by the standard Linux scheduler.






本来kvm01 虚拟机启动的时候，只有一个qemu-kvm进程，随后我又启动一个linuxvm02进程，top显示时就有两个qemu-kvm进程！
6：查看vm和node的信息：


[root@host1 autostart]# virsh dominfo linuxvm02
Id: 1
Name: linuxvm02
UUID: d340f42e-8611-52c5-949c-8013181e73d4
OS Type: hvm
State: running
CPU(s): 1
CPU time: 251.7s
Max memory: 1048576 kB
Used memory: 1048576 kB
Persistent: yes
Autostart: disable
[root@host1 autostart]# virsh nodeinfo 
CPU model: x86_64
CPU(s): 2
CPU frequency: 1603 MHz
CPU socket(s): 1
Core(s) per socket: 1
Thread(s) per core: 2
NUMA cell(s): 1
Memory size: 4008652 kB
[root@host1 autostart]#
7:[root@host1 ~]# virsh resume linuxvm02
error: Failed to resume domain linuxvm02
error: Timed out during operation: cannot acquire state change lock
原因：很有可能是linuxvm02 已经在启动了，只是没启动起来而已，还处于启动过程中，只需要耐心等待几分钟即可！
Login as a root user and kill the libvirtd.
killall -9 libvirtd




Remove the libvirtd pid file.
rm /var/run/libvirtd.pid
Restart libvirtd.
/etc/init.d/libvirtd  restart
8:每个虚拟机启动的时候，Host上都会多出一个vnet相关网卡，例如：


[root@host1 ~]# ifconfig
br0 Link encap:Ethernet HWaddr 00:E0:6F:12:AC:8A
......
eth0 Link encap:Ethernet HWaddr 00:E0:6F:12:AC:8A
......
lo Link encap:Local Loopback
......


virbr0 Link encap:Ethernet HWaddr 00:00:00:00:00:00
......
[root@host1 ~]# virsh start kvm01
Domain kvm01 started


[root@host1 ~]# virsh start linuxvm02
Domain linuxvm02 started
[root@host1 ~]# ifconfig
br0 Link encap:Ethernet HWaddr 00:E0:6F:12:AC:8A
......
eth0 Link encap:Ethernet HWaddr 00:E0:6F:12:AC:8A
......
lo Link encap:Local Loopback
.....


virbr0 Link encap:Ethernet HWaddr 00:00:00:00:00:00
.....
vnet0 Link encap:Ethernet HWaddr FE:52:00:7F:0D:CB
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:500
RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)
vnet1 Link encap:Ethernet HWaddr FE:52:00:3B:54:33
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:500
RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)
显然物理机上多出来两种虚拟网卡！ 总之每一个虚拟机都会对应一个物理网卡！
9：error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': 




[root@localhost network-scripts]# virsh list


error: Failed to reconnect to the hypervisor
error: no valid connection
error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': No such file or directory


        
[root@localhost network-scripts]# lsmod  | grep kvm
kvm_intel              50380  0 
kvm                   305081  1 kvm_intel
[root@localhost network-scripts]# /etc/init.d/libvirt
libvirtd        libvirt-guests  
[root@localhost network-scripts]# /etc/init.d/libvirt
libvirtd        libvirt-guests  
[root@localhost network-scripts]# /etc/init.d/libvirtd  start
Starting libvirtd daemon:                                  [  OK  ]
[root@localhost network-scripts]# virsh list
 Id    Name                           State
----------------------------------------------------


[root@localhost network-scripts]# chkconfig  --list | grep  libvirt
libvirt-guests  0:off   1:off   2:on    3:on    4:on    5:on    6:off
libvirtd        0:off   1:off   2:off   3:on    4:on    5:on    6:off
[root@localhost network-scripts]# 
10：有时候在物理机里用virsh shutdown domain 不工作


Why doesn't 'shutdown' doesn't seem to work?
If you are using Xen HVM or QEMU/KVM, ACPI must be enabled in the guest for a graceful shutdown to work. To check if ACPI is enabled, run:
virsh dumpxml $your-vm-name | grep acpi
If nothing is printed, ACPI is not enabled for your machine. Use 'virsh edit' to add the following XML under :
If your VM is running Linux, the VM additionally needs to be running acpid to receive the ACPI events.