Linux integration

1.5.4 KVM/QEMU Virtualization

1. Introduction

KVM (Kernel-based Virtual Machine) is a full virtualization solution built into the Linux kernel. Combined with QEMU (Quick Emulator), it provides a high-performance, enterprise-grade hypervisor capable of running multiple operating systems simultaneously on a single physical host.

This guide covers installation, VM creation, networking, storage, snapshots, live migration, and day-to-day management on Ubuntu 20.04/22.04/24.04 LTS.

1.1 KVM vs. Other Hypervisors

KVM/QEMU > Type-1 (bare-metal), native Linux kernel, best performance

VirtualBox > Type-2 (hosted), beginner-friendly, lower performance

VMware > Commercial Type-1/2, enterprise features, licensing costs

Hyper-V > Windows-only, tight Microsoft integration

Xen > Type-1, used by AWS, more complex setup

1.2 Architecture Overview

KVM turns the Linux kernel into a hypervisor. Each VM runs as a regular Linux process, leveraging hardware-assisted virtualization (Intel VT-x / AMD-V). QEMU handles device emulation while KVM handles CPU and memory directly.

• KVM module (/dev/kvm) — CPU & memory virtualization via hardware extensions

• QEMU — device emulation (disk, NIC, USB, display)

• libvirt — management API and daemon (libvirtd)

• virsh — command-line management tool

• virt-manager — graphical front-end

2. Prerequisites & Hardware Check

2.1 Hardware Requirements

Component Requirement

CPU Intel VT-x or AMD-V virtualization extensions (enabled in BIOS/UEFI)

RAM Minimum 4 GB host RAM; 8 GB+ recommended for running multiple VMs

Disk Sufficient free space for VM images (20–100 GB per VM typical)

Network At least one NIC; dedicated NIC recommended for bridged networking

2.2 Verify Virtualization Support

Run the following command to confirm the CPU supports hardware virtualization:

egrep -c '(vmx|svm)' /proc/cpuinfo

A result greater than 0 confirms support. Then verify KVM acceleration is available:

sudo kvm-ok

# Expected output:

# INFO: /dev/kvm exists

# KVM acceleration can be used

Note: If kvm-ok is not found, install it with: sudo apt install cpu-checker

To fully utilize KVM performance, it is essential that your processor supports hardware virtualization extensions.

To check support type:

lscpu | grep Virtualization

# Expected output:

Virtualization: VT-x or AMD-V

If you don't see any virtualization lines, or if the previous command returns nothing, your CPU may not support it, or the option may be disabled in the BIOS/UEFI.

If KVM hardware extensions are not available, you will still be able to run virtual machines, but they will be significantly slower.

2.3 Check Kernel Modules

lsmod | grep kvm

# Should show: kvm_intel (or kvm_amd) and kvm

3. Installation

3.1 Install KVM and Required Packages

sudo apt update

sudo apt install -y \

qemu-kvm \

libvirt-daemon-system \

libvirt-clients \

bridge-utils \

virt-manager \

virtinst \

ovmf

Package summary:

qemu-kvm > QEMU with KVM acceleration

libvirt-daemon-system > libvirt daemon (manages VMs)

libvirt-clients > CLI tools including virsh

bridge-utils > Network bridge utilities

virt-manager > GUI management application

virtinst > virt-install command for CLI VM creation

ovmf > UEFI firmware for VMs

3.2 Add User to Groups

sudo usermod -aG libvirt $(whoami)

sudo usermod -aG kvm $(whoami)

# Apply group changes (or log out and back in)

newgrp libvirt

3.3 Enable and Start the libvirt Service

sudo systemctl enable --now libvirtd

sudo systemctl status libvirtd

3.4 Verify Installation

virsh list --all # Should show empty VM list

virsh net-list --all # Should show 'default' network

is@ACQUARIO112-UB:~$ virsh net-list --all
Name State Autostart Persistent
--------------------------------------------
default active yes yes

If the "default network" doesn't appear, restart Host computer.

Success: If virsh list --all returns an empty table without errors, KVM is properly installed.

4. Networking

4.1 Default NAT Network

you can check the default status configuration with these commands:

virsh net-list --all

virsh net-info default

# Start and auto-start the default network

virsh net-start default

virsh net-autostart default

4.2 Network Modes Comparison

NAT > Default, VMs share host IP > Best for Outbound internet, isolated from LAN

Bridge > VMs get own LAN IP > Best for Full LAN access, servers, lab use

Isolated > No external access > Best for Secure internal lab networks

Routed > VMs routed through host > Best for Advanced setups, no bridge needed

4.3 Create a Bridged Network (br0) (only if you want to create a new custom configuration different from the default)

Bridged networking gives VMs direct access to the physical network. Edit the netplan configuration:

sudo nano /etc/netplan/01-netcfg.yaml

------------------ # Edit with these lines

network:

version: 2

ethernets:

eno1:

dhcp4: no

bridges:

br0:

interfaces: [eno1]

dhcp4: yes

parameters:

stp: false

forward-delay: 0

----------------

sudo netplan apply

4.4 Create a Custom libvirt Network (only if you want to create a new custom configuration different from the default)

sudo nano isolated-net.xml

------------------ # Edit with these lines
<network>

<name>isolated</name>

<dhcp>

</dhcp>

</ip>

</network>

-------------------

virsh net-define isolated-net.xml

virsh net-start isolated

virsh net-autostart isolated

5. Storage

5.1 Storage Formats

Format Notes

qcow2 QEMU Copy-On-Write v2. Supports snapshots, thin provisioning, compression. Recommended.

raw Direct disk image. Fastest I/O, no snapshots, largest file size.

vmdk VMware format. Use for compatibility with VMware tools.

vdi VirtualBox format. Limited KVM support.

5.2 Create a Disk Image

# Create a 50 GB qcow2 image (thin-provisioned)

qemu-img create -f qcow2 /var/lib/libvirt/images/vm01.qcow2 50G

# Create a raw image

qemu-img create -f raw /var/lib/libvirt/images/vm01.raw 50G

# Check image info

qemu-img info /var/lib/libvirt/images/vm01.qcow2

5.3 Storage Pools

# List existing pools

virsh pool-list --all

# Create a directory-based pool

virsh pool-define-as mypool dir --target /data/vms

virsh pool-build mypool

virsh pool-start mypool

virsh pool-autostart mypool

# List volumes in a pool

virsh vol-list mypool

5.4 Convert Disk Images

# Convert raw to qcow2

qemu-img convert -f raw -O qcow2 vm.raw vm.qcow2

# Convert vmdk to qcow2 (import from VMware)

qemu-img convert -f vmdk -O qcow2 vm.vmdk vm.qcow2

# Compress an existing qcow2

qemu-img convert -O qcow2 -c input.qcow2 compressed.qcow2

6. Creating Virtual Machines

To install from a local ISO:

sudo virt-install \

--name ubuntu-server \

--ram 4096 \

--vcpus 2 \

--disk path=/var/lib/libvirt/images/ubuntu-server.qcow2,size=50 \

--os-variant ubuntu24.04 \

--network bridge=br0 \

--graphics vnc,listen=0.0.0.0 \

--cdrom /tmp/ubuntu-24.04.4-desktop-amd64.iso

Single command line works better:

sudo virt-install --name ubuntu-server --ram 4096 --vcpus 2 --disk path=/var/lib/libvirt/images/ubuntu-server.qcow2,size=50 --os-variant ubuntu24.04 --network bridge=default --graphics vnc,listen=0.0.0.0 --cdrom /tmp/ubuntu-24.04.4-desktop-amd64.iso

6.2 Common virt-install Options

Option Description

--name VM name (must be unique)

--ram Memory in MB (e.g., 4096 = 4 GB)

--vcpus Number of virtual CPUs

--disk Disk config: path, size (GB), format

--os-variant OS type for tuning (osinfo-query os)

--network network=NAME or bridge=BRIDGE

--graphics vnc, spice, or none

--cdrom Path to ISO for installation

--import Import existing disk image (no install)

--autostart Auto-start VM on host boot

--cpu host-passthrough Pass physical CPU features to VM

6.3 Using virt-manager (GUI)

1. Open virt-manager: Applications → System Tools → Virtual Machine Manager

2. Click File → New Virtual Machine

3. Select installation source (ISO, network, existing disk)

4. Set memory and CPU resources

5. Configure storage: create new or use existing disk

6. Set network: NAT (default) or Bridge

7. Review and click Finish

See image: Ubuntu- VM- Virtual Machine Manager

6.4 Get List of Supported OS Variants

osinfo-query os | grep -i ubuntu

# Install if needed:

sudo apt install libosinfo-bin

7. VM Management with virsh

7.1 Essential virsh Commands

Command Description

virsh list --all List all VMs (running and stopped)

virsh start <vm> Start a VM

virsh shutdown <vm> Graceful shutdown

virsh destroy <vm> Force stop (like power off)

virsh reboot <vm> Reboot a VM

virsh suspend <vm> Pause a VM

virsh resume <vm> Resume a paused VM

virsh autostart <vm> Enable auto-start on host boot

virsh autostart --disable <vm> Disable auto-start

virsh dominfo <vm> Show VM information

virsh dumpxml <vm> Show VM XML configuration

virsh edit <vm> Edit VM XML configuration

virsh undefine <vm> Remove VM definition (keeps disk)

virsh console <vm> Connect to serial console

7.2 CPU and Memory Management

# Adjust vCPUs (hot-plug if VM supports it)

virsh setvcpus vm01 4 --live --config

# Adjust memory (balloon driver required)

virsh setmem vm01 8388608 --live --config # 8 GB in KB

# Set max memory

virsh setmaxmem vm01 16777216 --config # 16 GB in KB

7.3 Connect to VM Console

# Serial console (requires console= in guest kernel cmdline)

virsh console vm01

# Exit with: Ctrl+]

# VNC viewer

virsh vncdisplay vm01 # Shows :0 or :1

vncviewer localhost:0

# SSH (once VM is running and has SSH)

ssh user@$(virsh domifaddr vm01 | grep -oP '\d+\.\d+\.\d+\.\d+')

7.4 Attach and Detach Devices

# Attach a disk

virsh attach-disk vm01 /data/extra.qcow2 vdb --subdriver qcow2 --persistent

# Detach a disk

virsh detach-disk vm01 vdb --persistent

# Attach a USB device

virsh attach-device vm01 usb-device.xml --live

# Add a NIC

virsh attach-interface vm01 network default --model virtio --persistent

8. Snapshots

8.1 Internal Snapshots (qcow2 only)

Internal snapshots are stored inside the qcow2 image. The VM must use qcow2 format.

# Create a snapshot

virsh snapshot-create-as vm01 snap1 'Before update' --disk-only

# List snapshots

virsh snapshot-list vm01

# Show snapshot details

virsh snapshot-info vm01 snap1

# Revert to a snapshot

virsh snapshot-revert vm01 snap1

# Delete a snapshot

virsh snapshot-delete vm01 snap1

8.2 External Snapshots

# Create external disk snapshot (VM can remain running)

virsh snapshot-create-as vm01 external-snap \

--diskspec vda,snapshot=external,file=/var/lib/libvirt/images/vm01-snap.qcow2 \

--disk-only --atomic

⚠️ Warning: Deleting external snapshots requires manual blockcommit/blockpull operations. Plan snapshot strategy before use in production.

8.3 Blockcommit — Merge Snapshot Back

# Commit external snapshot changes back to base image

virsh blockcommit vm01 vda --active --verbose --pivot

# Verify the active disk path

virsh domblklist vm01

9. Cloning & Templates

9.1 Clone a VM

# Shutdown the VM before cloning

virsh shutdown vm01

# Clone the VM

virt-clone \

--original vm01 \

--name vm01-clone \

--file /var/lib/libvirt/images/vm01-clone.qcow2

# Clone with auto-generated name and path

virt-clone --original vm01 --auto-clone

9.2 Create a Template VM

8. Install and configure the base OS

9. Remove machine-specific data:

# Inside the VM

sudo cloud-init clean

sudo truncate -s 0 /etc/machine-id

sudo rm -f /var/lib/dbus/machine-id

sudo ln -s /etc/machine-id /var/lib/dbus/machine-id

sudo history -c && sudo shutdown -h now

10. On the host, mark the VM as a template:

virsh dumpxml vm-template > template.xml

# Edit template.xml: remove MAC addresses, UUIDs

# Use virt-sysprep for automated cleanup

sudo virt-sysprep -d vm-template

11. Clone the template for each new VM

9.3 virt-sysprep — Automated Cleanup

# Install virt-sysprep

sudo apt install libguestfs-tools

# Reset VM to pristine state

sudo virt-sysprep -d vm01

# Customize: set hostname, password, SSH keys

sudo virt-customize -d vm01 \

--hostname newserver \

--root-password password:MySecret \

--ssh-inject root:file:/home/user/.ssh/id_rsa.pub

10. Live Migration

10.1 Prerequisites for Live Migration

• Both hosts must have KVM and libvirt installed

• Shared storage (NFS, GlusterFS, Ceph) for disk images OR use virsh migrate with --copy-storage-all

• SSH key-based authentication between hosts

• Same libvirt version (or compatible versions) on both hosts

• Network connectivity on the migration port (default: 49152–49215)

10.2 Configure SSH Between Hosts

# On source host — generate and copy key

ssh-keygen -t ed25519 -N '' -f ~/.ssh/id_ed25519

ssh-copy-id root@destination-host

# Test

ssh root@destination-host virsh list

10.3 Perform Live Migration

# Live migration with shared storage

virsh migrate --live vm01 qemu+ssh://destination-host/system

# Live migration + copy storage (no shared storage needed)

virsh migrate --live --copy-storage-all vm01 \

qemu+ssh://destination-host/system

# Persistent migration (VM defined on destination)

virsh migrate --live --persistent vm01 \

qemu+ssh://destination-host/system

# Monitor migration progress

virsh domjobinfo vm01

ℹ️ Note: For production live migration, shared storage (NFS/Ceph) is strongly recommended. Copy-storage migration is slower and puts more load on the network.

11. Performance Tuning

11.1 CPU Tuning

# Use host-passthrough CPU model (exposes all CPU features)

virsh edit vm01

# Change: <cpu mode='custom'> → <cpu mode='host-passthrough' check='none'/>

# Pin vCPUs to physical cores (CPU pinning)

virsh vcpupin vm01 0 2 # Pin vCPU 0 to pCPU 2

virsh vcpupin vm01 1 3 # Pin vCPU 1 to pCPU 3

11.2 Memory Tuning

# Enable Huge Pages for better memory performance

echo 1024 | sudo tee /proc/sys/vm/nr_hugepages

# In VM XML — use hugepages

# <memoryBacking>

# <hugepages/>

# </memoryBacking>

# Enable KSM (Kernel Same-page Merging) for memory deduplication

echo 1 | sudo tee /sys/kernel/mm/ksm/run

11.3 Disk I/O Tuning

# Use virtio-scsi controller (best disk performance)

# In virt-install:

# --disk ...,bus=virtio

# Set disk I/O scheduler

virsh blkdeviotune vm01 vda --total-iops-sec 5000 --live

# Use native AIO for raw images

# In XML: <driver name='qemu' type='raw' cache='none' io='native'/>

11.4 Network Tuning

# Use virtio NIC (always prefer over e1000 or rtl8139)

# virt-install: --network ...,model=virtio

# Enable multiqueue virtio-net

# In VM XML:

# <driver name='vhost' queues='4'/>

# Use macvtap for near-native performance

# --network type=direct,source=eno1,source_mode=bridge,model=virtio

Optimization Description

virtio Paravirtual drivers — always use for disk/NIC/memory balloon

host-passthrough CPU mode — exposes host CPU features to guest

CPU pinning Prevents vCPU migration across pCPUs, reduces latency

Huge Pages Reduces TLB misses for memory-intensive workloads

cache=none Bypass page cache for raw disk images (use with io=native)

KSM Shares identical memory pages across VMs (good for many similar VMs)

12. Monitoring

12.1 VM Statistics

# CPU stats

virsh cpu-stats vm01 --total

# Memory stats

virsh dommemstat vm01

# Block I/O stats

virsh domblkstat vm01

# Network stats

virsh domifstat vm01 vnet0

# Overall VM info

virsh dominfo vm01

12.2 Host-Level Monitoring

# Processes per VM

ps aux | grep qemu

# Memory usage

free -h

cat /proc/meminfo | grep -i huge

# Disk usage of images

du -sh /var/lib/libvirt/images/*

qemu-img info /var/lib/libvirt/images/vm01.qcow2

# Network interfaces

ip addr show

brctl show

12.3 virt-top — Real-Time Monitoring

# Install virt-top

sudo apt install virt-top

# Run (similar to top, but for VMs)

virt-top

13. Backup & Restore

13.1 Backup Strategy

Method Notes

Cold backup (offline) Shutdown VM, copy qcow2 image. Simplest, 100% consistent.

Live snapshot backup Create snapshot, backup snapshot file, delete snapshot. VM stays online.

Agent-based backup Use qemu-guest-agent for filesystem-consistent live backups.

virsh dumpxml Always backup VM XML configuration alongside disk images.

13.2 Backup a VM (Offline)

#!/bin/bash

VM=vm01

BACKUP_DIR=/backup/vms

mkdir -p $BACKUP_DIR

# Save VM XML

virsh dumpxml $VM > $BACKUP_DIR/$VM.xml

# Shutdown VM

virsh shutdown $VM

sleep 30

# Copy disk image

cp /var/lib/libvirt/images/$VM.qcow2 $BACKUP_DIR/$VM.qcow2

# Restart VM

virsh start $VM

echo 'Backup complete: '$BACKUP_DIR

13.3 Live Backup with Snapshot

VM=vm01

SNAP_IMG=/backup/$VM-snap.qcow2

# Create external snapshot

virsh snapshot-create-as $VM livesnap \

--diskspec vda,snapshot=external,file=$SNAP_IMG \

--disk-only --atomic

# Copy base image (now frozen)

cp /var/lib/libvirt/images/$VM.qcow2 /backup/$VM-backup.qcow2

# Merge snapshot back

virsh blockcommit $VM vda --active --verbose --pivot

# Clean up snapshot file

rm -f $SNAP_IMG

virsh snapshot-delete $VM livesnap --metadata

13.4 Restore a VM

VM=vm01

BACKUP_DIR=/backup/vms

# Restore disk image

cp $BACKUP_DIR/$VM.qcow2 /var/lib/libvirt/images/$VM.qcow2

# Re-define VM from XML

virsh define $BACKUP_DIR/$VM.xml

# Start the VM

virsh start $VM

14. Troubleshooting

14.1 Common Issues

Problem Solution

libvirtd not starting Check: sudo systemctl status libvirtd Fix: sudo systemctl restart libvirtd

Permission denied on /dev/kvm Add user to kvm group: sudo usermod -aG kvm $USER

VM fails to start Check logs: sudo journalctl -xe | grep libvirt Check: virsh start vm01 (shows error)

No network in VM Verify default network: virsh net-start default Check bridge: ip addr show virbr0

VM very slow Ensure virtio drivers installed in guest Check host load: top, vmstat

Cannot connect to VNC Check firewall: sudo ufw allow 5900/tcp Verify VNC port: virsh vncdisplay vm01

Snapshot fails Disk must be qcow2 format Check disk space on host

Migration fails Verify SSH connectivity Check firewall ports 49152–49215

14.2 Useful Log Files

# libvirt daemon log

sudo journalctl -u libvirtd -f

# QEMU log for specific VM

sudo cat /var/log/libvirt/qemu/vm01.log

# Live tail

sudo tail -f /var/log/libvirt/qemu/vm01.log

14.3 Diagnostic Commands

# Check libvirt connectivity

virsh -c qemu:///system version

# Validate VM XML

virsh define vm.xml --validate

# Check host capabilities

virsh capabilities

# Check domcapabilities for a specific machine type

virsh domcapabilities --machine pc-q35-6.2

# List all virtual networks

virsh net-list --all

# Check running QEMU processes

ps aux | grep qemu | grep -v grep

15. Quick Reference Cheat Sheet

Command Action Notes

virsh list --all List VMs All states

virsh start <vm> Start VM —

virsh shutdown <vm> Graceful stop Guest OS required

virsh destroy <vm> Force stop Immediate

virsh reboot <vm> Reboot —

virsh suspend / resume <vm> Pause / unpause RAM preserved

virsh snapshot-create-as <vm> <name> Create snapshot qcow2 only

virsh snapshot-revert <vm> <name> Restore snapshot —

virsh migrate --live <vm> qemu+ssh://... Live migrate Shared storage needed

virsh dumpxml <vm> > vm.xml Export XML config For backup/restore

virsh define vm.xml Import VM from XML —

virsh undefine <vm> Remove VM config Disk preserved

virt-clone --original <vm> Clone VM Shutdown first

qemu-img info <image> Image info —

qemu-img convert -O qcow2 in out Convert image —

qemu-img resize <image> +10G Resize image Expand only

📚 Further Reading: Official docs: libvirt.org | qemu.org | Ubuntu Server Guide (ubuntu.com/server/docs/virtualization-introduction)

Ubuntu-_VM-_Virtual_Machine_Manager