Manjaro Difference between revisions of "Virt-manager/tr"

Difference between revisions of "Virt-manager/tr"

From Manjaro
(Created page with "= Genel Bakış =")
Tags: Mobile web edit Mobile edit
 
(Updating to match new version of source page)
Tags: Mobile web edit Mobile edit
Line 82: Line 82:
  sudo systemctl enable libvirtd.service
  sudo systemctl enable libvirtd.service
  sudo systemctl start libvirtd.service
  sudo systemctl start libvirtd.service
Add user to ''libvirt'' group to use the '''system'''-level virtual machines (qemu:///system)
Add user to ''libvirt,'' ''libvirt-qemu'' and ''kvm'' groups to use the '''system'''-level virtual machines (qemu:///system)
  sudo usermod -a -G libvirt $USER
  sudo usermod -a -G libvirt,libvirt-qemu,kvm $USER
{{note|
{{note|
# You don't need this step to run system-level virtual machines. However, virt-manager will prompt for sudoer's password when launch if the user is not in the ''libvirt'' group   
# You don't need this step to run system-level virtual machines. However, virt-manager will prompt for sudoer's password when launch if the user is not in the ''libvirt'' group   

Revision as of 06:38, 13 May 2024

Other languages:
English • ‎Türkçe • ‎русский

Genel Bakış

Topbar logo.png


tip
Have you heard about Virtualbox to virtualize operating systems like Linux and Windows or Mac? It 'a good software, but not the only one! Don't forget VMware, Gnome Boxes and ...virt-manager.



Virt-manager uses libvirt and it's a manager of many hypervisors, including the one that we want to use here: QEMU/KVM.



Why do I need to virtualize?

  • To learn about a new O.S.
  • To configure a hardware that has a setup only for that operating system
  • To use a software that only works on another

I want to know more

What we absolutely must know


1. What is my CPU. Identify it and make sure it's at least a quad core. TAKE A LOOK AT CPU-World
2. Check if the 'virtualization parameters' are enabled on BIOS using

LC_ALL=C lscpu | grep Virtualization

3. How much memory I have. Check the RAM and verify that is at least 4GB.
4. The amount of free space on my hard drive. The virtual machine can use a disk image file so extra partitions are not necessary.

Note
You can also use raw partition specific for virtual machines. See partitioning for more information

5. The minimum hardware requirements of the operating system you want to install as a virtual machine.

Install virt-manager, qemu and all dependencies


From terminal:

sudo pacman -S --needed virt-manager qemu-desktop libvirt edk2-ovmf dnsmasq iptables-nft

For TPM support:

sudo pacman -S --asdeps swtpm

Enable and start service

sudo systemctl enable libvirtd.service
sudo systemctl start libvirtd.service

Add user to libvirt, libvirt-qemu and kvm groups to use the system-level virtual machines (qemu:///system)

sudo usermod -a -G libvirt,libvirt-qemu,kvm $USER
Note
  1. You don't need this step to run system-level virtual machines. However, virt-manager will prompt for sudoer's password when launch if the user is not in the libvirt group
  2. You can also create user-level virtual machines (qemu:///session) and use without sudoer's privelige. However, some features such as VirtioFS file sharing may be unavailable in qemu:///session

Using Virt-Manager for guest creation

0. Prepare installation image. If you're going to install Windows, prepare the Virtio driver image too.


Tip
You can also find the virtio-win package in AUR. The image is located in /var/lib/libvirt/images, which is the default directory pool of qemu:///system

1. Launch menu Virtual Machine Manager. It should already have a LXC connection. You can disconnect and remove it if you don't use LXC.

2. Go to File, choose Add Connection and choose hypervisor QEMU/KVM, or QEMU/KVM user session if you don't want system-level virtual machines. Click on connect.

3. You need directory pools to store the disk images of virtual machines or the ISO file of CD/DVD. Double click qemu/kvm, go on storage and add by clicking + the path to the folder where you have the iso and the folder where create the virtual machine.

tip
My advice: use a different partition than the root. In case you need to reinstall your operating system you don't lose the VM that already ready-to-start.

4. Click on create a new virtual machine: select Local install media (ISO image or CDROM), and select the installation ISO image and OS type (if not detected).

5. How many CPU assign and how much memory? (check the recommended requirements of O.S. that you are installing)

6. Create the file system of the virtual machine by selecting Select or create custom storage and click Manage.... Under your desired directory pool, create the volume of the virtual machine (default in qcow2 format). How many GB? Check the recommended requirements O.S. you install.

7.Assign a name to the machine and flag customize configuration before install. You have access to the screen with all the hardware that will be virtualized, do a check if there is all that is needed to initialize and launch the installer.

8. In Overview, change the firmware to UEFI for future-proof capacity.

9. Change the type of SATA Disk 1 (the disk image of the creating virtual machine) to virtio for better performance. Change discard mode to unmap, and then apply the change. You should notice the device name would change from SATA Disk 1 to VirtIO Disk 1.

Warning
If you're installing Windows, make sure to create a new SATA CDROM device and select the virtio-win image as the media, so that the Windows Installer can load the driver to recognize the disk. The virtio storage driver should be located at somewhere like E:\viostor\w10\amd64

10. Set the NIC type to virtio too for better network performance.

Note
If you're installing Windows, once the system if ready, make sure to install all the virtio drivers so that your virtIO network can work normally

11. Add TPM chip, select TIS model through Emulated device backend.

12. Add a watchdog to reboot the guest when it hangs. Leave the settings as default.

13. And a hardware RNG, to get entropy from the host.

14. Click on the top to start installation.

Warning
All these steps are visible on youtube. >>> PLAY THIS
YouTube 1.png

Install guest additions

Once the VM is started and running you have to install the spice guest tools.

For Windows is a single package: spice-guest-tools-xxxx.exe

For linux are: spice-vdagent and xf86-video-qxl. If you visrtualize a linux distro you can install them with their package manager

Tune the display settings

The default model of display card is QXL. If your virtual machine is Linux system, you can change it to virtio and enable 3D acceleration for better graphic performance.
Windows virtual machine, however doesn't support virtio display yet. Nevertheless, we can increase its VGA memory from the default 16 MB to 64 MB to allow higher display resolution and slightly better 2D graphical performance.
To do so, we need to edit the XML file of the virtual machine.

  1. In virt-manager, go to edit > preference, and check Enable XML edit.
  2. In virtual machine details, go to display card. Under the XML tab, change the value of vgamem to 65536, then apply the change.


File sharing between host and guest

For Linux guests, Virtio-FS and 9p are available for file sharing. See ArchWiki for more information.
For Windows guest, the easiest way to share file between host and guest is through SAMBA.

  1. Create a SAMBA usershare
  2. In Windows guest, enter \\192.168.122.1 in the file explorer, and you should be able to see the usershare on your host.
  3. Map that SAMBA usershare to a new drive, and connect to it with the appropriate credential. That's it.

Check this article for more information.

Optimize vCPU

Check this article to customize vCPU topology for better performance.
For example, my CPU has 1 socket, 8 cores, and 16 threads in total. The automatic topology assigns 4 sockets, 1 core, and 1 thread to my guest. After changing it to 1 socket, 4 cores, and 1 thread, the performance increases significantly.

References

Cookies help us deliver our services. By using our services, you agree to our use of cookies.