Programster's Blog

Tutorials focusing on Linux, programming, and open-source

KVM Cheatsheet

Related Posts

List Running Virtual Machines

virsh list

List All Virtual Machines

virsh list --all

Managing Guest State

Shutdown Guest

virsh shutdown $VM_ID_OR_NAME

Start Guest

virsh start $VM_ID_OR_NAME

Reboot Guest

virsh reboot $VM_ID_OR_NAME

Destroy a Guest

This command is an ungraceful shutdown, much like if you were to pull the power out of a computer. You should use this if a guest becomes unresponsive. It does not delete the guest. The disk image will remain and the guest can be restarted.

virsh destroy $VM_ID_OR_NAME

Suspension

Suspension is a way to immediately "pause" a guest so that it no longer uses the CPU, disk, or network. However, it will continue to reside in memory. You may want to save/load a session instead, which would mean it no longer takes up memory, but is not instant. such that it no longer takes up memory and can be restored to its exact state (even after a reboot), it is necessary to save and restore the guest.

A suspended session will be lost if the host system is rebooted. However, a saved guest does persist.

Suspend Guest

virsh suspend $VM_ID_OR_NAME

Resume Guest

virsh resume $VM_ID_OR_NAME

Defining

Defining a Guest

Defining a guest allows one to start it from its name, rather than having to find it's XML file and running virsh create $name.xml. This means that guests will also show in virsh list --all when they are shutdown.

virsh define filename.xml

Undefine a Guest

In order to use a name over again for a new guest, you have to undefine the old one. You need to remove it's storage system as well.

virsh undefine $VM_ID

Guest Configuration

You can manually edit the guest's xml configuration file with:

virsh edit $VM_ID

Changes will not take effect until the guest is rebooted

List OS Variants

When creating a guest with virt-install you need to specify the --os-variant. To get a list of acceptable values (on Ubuntu 16.04), install the libosinfo-bin package before running the command below:

osinfo-query os

Rename guest

virsh domrename $OLD_NAME $NEW_NAME

You can only do this whilst the guest is not running.

Guest Start on Boot (Autostart)

virsh autostart $VM_ID

To disable Guest Autostart

virsh autostart --disable $VM_ID

Resizing Memory

Run the following command to bring up the config for the guest.

virsh edit $VM_ID

Change the memory and currentMemory fields to be the size you want in KiB.

Now use virsh to shutdown and startup the container for the changes to take effect.

Resizing Memory With Script

VM_ID="my_vm_id"
NEW_AMOUNT="4000"

EDITOR='sed -i "s;[0-9]*</currentMemory>;$NEW_AMOUNT</currentMemory>;"' virsh edit $VM_ID
EDITOR='sed -i "s;[0-9]*</memory>;$NEW_AMOUNT</memory>;"' virsh edit $VM_ID

virsh shutdown $VM_ID
virsh start $VM_ID

Do not use virsh memtune. See here for more details.

CPU Management

Discover CPU Scheduling Parameters

virsh schedinfo $VM_ID

Permanently Set CPU Shares For Live Running Instance

virsh schedinfo $VM_ID \
  --set cpu_shares=[0-262144] \
  --live \
  --current \
  --config

Get the CPU Pinning Settings for a Guest

virsh vcpupin blog.programster.org

Example output:

VCPU: CPU Affinity
----------------------------------
   0: 0-3
   1: 0-3

I got the output above because I gave the guest access to 2 vCPUs but didn't pin anything.

Pin A CPU

If I wanted to set the cores that a guest can use, I could do the following:

virsh vcpupin blog.programster.org 0 2

That will set the first vCPU (the one with ID 0) to only run on core ID 2. Thus the output of virsh vcpupin blog.programster.org changes to:

VCPU: CPU Affinity
----------------------------------
   0: 2
   1: 0-3

Pinning could be a great way to limit the effect a certain guest has on others, or to give a guest a dedicated core etc.

Guest Console

Enter Guest's Console

virsh console $VM_ID

Exit Guest's Console

Use the following keyboard shortcut (not a command):

Cntrl-]

Save Guest

virsh save $VM_ID $FILENAME

Saves the RAM (not including disk) of a running guest to a "state file" at the specified file name/path, so that it can be restored later. Once saved, the domain will no longer be running on the system, thus the memory allocated for the domain will be free for other domains to use. virsh restore (in "Load Guest" below) restores from this state file.

Optional parameters

  • --xml $FILEPATH - Usually omitted, but can be used to supply an alternative XML file for use on the restored guest with changes only in the host-specific portions of the domain XML . For example, it can be used to account for file naming differences that are planned to be made via disk snapshots of underlying storage after the guest is saved. If you forgot to provide this, you can make use of save-image-define later to achieve the same effect. You may also wish to refer to save-image-dumpxml and save-image-edit
  • --bypass-cache - the save will avoid the file system cache, although this may slow down the operation.
  • --running | --paused - Normally, restoring a saved image will decide between running or paused based on the state the domain was in when the save was done; passing either the --running or --paused flag will allow overriding which state the restore should use.
  • --verbose

Progress may be monitored using domjobinfo virsh command and canceled with domjobabort (or just Ctrl-C).

Load Guest (restore)

virsh restore $FILENAME

The filename here is the same file that you saved to in the previous command, not one of the other guest files!

Cloning

Below is a simple example of cloning a guest.

virt-clone \
  --original $VM_TO_CLONE \
  --auto-clone \
  --name $NEW_VM_NAME

Networking

List Running Network Configs

virsh net-list

List All Network Configs

virsh net-list --all

You can find network configs stored in /home/stuart/network-configs/

Edit Network Config

virsh net-list $NETWORK_NAME

Create Temporary Network Config

virsh net-create --file $ABSOLUTE_FILE_PATH

Create Permanent Network Config

virsh net-define --file $ABSOLUTE_FILE_PATH

Example Bridge Network Config File

<network>
  <name>examplebridge</name>
  <forward mode='route'/>
  <bridge name='kvmbr0' stp='on' delay='0'/>
  <ip address='192.168.1.1' netmask='255.255.255.0' />
</network>

Start Network Config

virsh net-start $NETWORK_ID

Enable Network Autostart

net-autostart --network $NETWORK_ID

Disable Network Autostart

net-autostart \
  --network $NETWORK_ID \
  --disable

Example Manual Network Config With Bridge

This is an example /etc/network/interfaces file for Ubuntu users. If you do this, make sure you have installed the bridge-utils package beforehand.

# The loopback network interface
auto lo
iface lo inet loopback

# The primary network interface
auto p17p1
iface p17p1 inet manual

auto kvmbr0
iface kvmbr0 inet static
    address 192.168.1.19
    netmask 255.255.255.0
    network 192.168.1.0
    broadcast 192.168.1.255
    gateway 192.168.1.254
    bridge_ports p17p1
    bridge_stp off
    bridge_fd 0
    bridge_maxwait 0

Here is a netplan version:

# This file describes the network interfaces available on your system
# For more information, see netplan(5).
network:
  version: 2
  renderer: networkd
  ethernets:
    enp39s0:
      dhcp4: no
  bridges:
    kvmbr0:
      addresses: [ 192.168.1.186/24 ]
      gateway4: 192.168.1.1
      nameservers:
        addresses:
          - 8.8.8.8
          - 8.8.4.4
      interfaces:
        - enp39s0

Example Netplan DHCP Config

The following config will set up a bridge where your host is using DHCP.

network:
  version: 2
  renderer: networkd
  ethernets:
    enp3s0:
      dhcp4: no
      dhcp6: no
  bridges:
    kvmbr0:
      dhcp4: yes
      dhcp6: no
      nameservers:
        addresses:
          - 8.8.8.8
          - 8.8.4.4
      interfaces:
        - enp3s0

Configure VM To Use Manual Bridge

If you manually set the bridge up with the section above rather than through using the virsh net commands, this is how to configure deployed guests make use of it:

virsh edit $VM_ID

Find the following section

    <interface type='network'>
      <mac address='52:54:00:4d:3a:bd'/>
      <source network=''/>
      <model type='virtio'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/>
    </interface>

Change it to be like so:

    <interface type='bridge'>
        <mac address='52:54:00:4d:3a:bd'/>
        <source bridge='[bridge name here]'/>
        <model type='virtio'/>
        <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/>
    </interface>

Now run the following two commands since reboots wont work.

virsh shutdown $VM_ID
virsh start $VM_ID

Add Network Interface to VM

I used the command below to add a NIC to my guest that uses my host's bridge interface called kvmbr1.

virsh attach-interface \
  --domain guest1 \
  --type bridge \
  --source kvmbr1 \
  --model virtio \
  --config

If your guest is running at the time, you need to add the --live parameter.

You could specify a mac address with --mac but without it, one will be generated randomly.

Snapshotting

Create Internal Snapshot

virsh snapshot-create $VM_ID

You can take snapshots of guests whilst they are running. Whilst the snapshot is being taken, the guest will be "paused". The "state" of the guest is also saved.

Create Internal Snapshot With Name

virsh snapshot-create-as $VM_ID $SNAPSHOT_NAME

Create Internal Snapshot With Name and Description

virsh snapshot-create-as $VM_ID $SNAPSHOT_NAME $DESCRIPTION

Create Internal Snapshot With Name and Description Using File

If you just love writing xml, then you can create a file like so:

<domainsnapshot>
    <name>Name for the snapshot</name>
    <description>Description for the snapshot</description>
</domainsnapshot>

... then pass it to virsh snapshot-create to create the snapshot

virsh snapshot-create $VM_ID $FILEPATH

Create External Snapshot

Refer here.

List Snapshots

virsh snapshot-list $VM_ID

Snapshot-list defaults to being in alphabetical rather than chronological order. If you want to find out what your latest snapshots are, you may wish to add the optional --tree or --leaves parameters.

Restore Snapshot

virsh snapshot-revert $VM_ID $SNAPSHOT_NAME

Delete Snapshot

virsh snapshot-delete $VM_ID $SNAPSHOT_NAME

More snapshot functionality can be found in Qcow2 Conversion and Snapshotting

Edit Snapshot

If you use virsh with internal qcow2 snapshots and you decide to move the file to another location, you will not be able to restore those snapshots. This is easily fixed by editing the snapshots and updating the filepath.

virsh snapshot-edit $VM_ID_OR_NAME $NAME_OF_SNAPSHOT

Blockcommit

Use blockcommit to merge a qcow2 file down into its backing file.

Blockpull

  • Use blockpull to flatten a disk/image chain by moving data up the chain.
  • The principal here is to "flatten" a chain, so it is impossible to target the removal of a single node in the chain except if your chain happens to be small enough.
  • The original files will be kept unless you provide the --delete option.
    • Be careful about providing --delete as you might have other guests using disk images that reference them if you use thin-provisioning.
  • Full list of options can be found here.

This example will pull all the images up into the current active/head image, for disk vda.

virsh blockpull $GUEST_ID vda --wait
  • --wait causes the command to block until the operation completes.

The following example will flatten a chain by merging all images above the specified base, so that the base will be referenced by the current top/active disk image. E.g. your chain will only contain the active, the base, and any images the base references.

virsh blockpull \
  $GUEST_ID \
  vda \
  --base /path/to/disk-image.qcow2 \
  --verbose \
  --wait

Disable AppArmor

If you are trying to use external snapshots, you may find it easier to disable apparmor. Do this by editing your qemu configuration:

sudo vim /etc/libvirt/qemu.conf

... and adding the following line for the security_driver

security_driver = "none"

Then restart libvirt for the changes to take effect.

sudo service libvirtd restart

References

Last updated: 13th October 2020
First published: 16th August 2018