Virsh, Libvirt, etc.

(July 2017)

virsh is the primary linux CLI tool for dealing with virtual machines.

What is virsh? How does it fit into linux virtualization?

See VIRSH(1).

# virsh list --all
 Id    Name                           State
 -     debian                         shut off
 -     freebsd-512ram-1cpu-20gb       shut off
 -     openbsd                        shut off
 -     openbsd59                      shut off
 -     rhel7                          shut off

Edit a guest/domain configuration:

# virsh edit my-vm

Starting, shuting down, rebooting, suspending/pausing, resuming/unpausing, and pulling the plug on a domain:

# virsh start my-vm
# virsh shutdown my-vm
# virsh reboot my-vm
# virsh suspend my-vm
# virsh resume my-vm
# virsh destroy my-vm

Set a guest to autostart or not autostart:

# virsh autostart my-vm
# virsh autostart --disable my-vm
# virsh list --autostart

Take a snapshot of a domain, list snaphots, restore/revert to a snapshot, and delete a saved snapshot:

# virsh snapshot-create my-vm
# virsh snapshot-list my-vm
# virsh snapshot-restore my-vm snapshot-1
# virsh snapshot-delete my-vm snapshot-2

Libvirt provides storage management on the physical host through storage pools and volumes.

A storage pool is a quantity of storage set aside by an administrator, often a dedicated storage administrator, for use by virtual machines. Storage pools are divided into storage volumes either by the storage administrator or the system administrator, and the volumes are assigned to VMs as block devices.

For example, the storage administrator responsible for an NFS server creates a share to store virtual machines’ data. The system administrator defines a pool on the virtualization host with the details of the share (e.g. should be mounted on /vm_data). When the pool is started, libvirt mounts the share on the specified directory, just as if the system administrator logged in and executed ‘mount /vmdata’. If the pool is configured to autostart, libvirt ensures that the NFS share is mounted on the directory specified when libvirt is started.

Once the pool is started, the files in the NFS share are reported as volumes, and the storage volumes’ paths may be queried using the libvirt APIs. The volumes’ paths can then be copied into the section of a VM’s XML definition describing the source storage for the VM’s block devices.

Although all storage pool backends share the same public APIs and XML format, they have varying levels of capabilities. Some may allow creation of volumes, others may only allow use of pre-existing volumes. Some may have constraints on volume size, or placement.

The top level tag for a storage pool document is ‘pool’. It has a single attribute type, which is one of dir, fs, netfs, disk, iscsi, logical, scsi (all since 0.4.1), mpath (since 0.7.1), rbd (since 0.9.13), sheepdog (since 0.10.0), gluster (since 1.2.0), zfs (since 1.2.8) or vstorage (since 3.1.0).

# virsh pool-list --all --details
 Name          State    Autostart  Persistent    Capacity  Allocation   Available
  data-libvirt  running  yes        yes         468.45 GiB  179.47 GiB  288.98 GiB
  default       running  yes        yes         422.38 GiB  213.00 GiB  209.39 GiB
  Downloads     running  yes        yes         422.38 GiB  213.00 GiB  209.39 GiB

# virsh pool-info virt
Name:           virt
UUID:           aa52fbc0-2088-43ce-9618-3a7993cb45e4
State:          running
Persistent:     yes
Autostart:      yes
Capacity:       468.45 GiB
Allocation:     234.22 GiB
Available:      234.23 GiB

# virsh vol-list --pool virt-lvm-pool
 Name                 Path
  home                 /dev/falstaff-vg/home
  lv-openbsd-30gb      /dev/falstaff-vg/lv-openbsd-30gb
  lv-pi2-test          /dev/falstaff-vg/lv-pi2-test
  root                 /dev/falstaff-vg/root
  swap_1               /dev/falstaff-vg/swap_1

# virsh pool-define-as test --type dir --target /data/virt
# virsh pool-start test
# virsh pool-autostart test
# virsh pool-dumpxml virt
<pool type='dir'>
  <capacity unit='bytes'>502996557824</capacity>
  <allocation unit='bytes'>251489689600</allocation>
  <available unit='bytes'>251506868224</available>
# virsh pool-destroy test
Pool test destroyed 
# virsh pool-create /etc/libvirt/storage/test.xml 
Pool test created from /etc/libvirt/storage/test.xml
# virsh pool-destroy test
Pool test destroyed 
# virsh pool-undefine test
Pool test has been undefined

# virsh find-storage-pool-sources logical
	<device path='/dev/md0'/>
	<format type='lvm2'/>
# sudo vgscan
  Reading all physical volumes.  This may take a while...
  Found volume group "vg0" using metadata type lvm2
# vgs --all
  VG   #PV #LV #SN Attr   VSize VFree
  vg0    1   7   0 wz--n- 2.73t 1.91t
$ ls /dev | grep vg0
# virsh pool-define-as --name vg0 --type logical --target /dev/vg0
# virsh pool-start vg0
Pool vg0 started
# virsh pool-autostart vg0
Pool vg0 marked as autostarted
# virsh pool-list --all --details
Name     State    Autostart  Persistent    Capacity  Allocation   Available
default  running  yes        yes         110.00 GiB    9.70 GiB  100.30 GiB
vg0      running  yes        yes           2.73 TiB  841.56 GiB    1.91 TiB

# virsh vol-create-as vg0 myvm.img 10G
# virsh vol-delete --pool vg0 myvm.img

# lvs
LV           VG   Attr     LSize   Pool Origin Data%  Move Log Copy%  Convert
2012r2       vg0  -wi-a---  80.00g                                           
myvm.img     vg0  -wi-a---  10.00g                                           
lvroot       vg0  -wi-ao-- 111.76g                                           
lvswap       vg0  -wi-ao--  29.80g                                           
smbshare     vg0  -wi-ao-- 450.00g                                           
virtbackup   vg0  -wi-ao-- 320.00g                                           
# vgs
VG   #PV #LV #SN Attr   VSize VFree
vg0    1   8   0 wz--n- 2.73t 1.63t

--os-variant OS_VARIANT
	Optimize the guest configuration for a specific operating system (ex.  'fedora18', 'rhel7', 'winxp').
	While not required, specifying this options is HIGHLY RECOMMENDED, as it can greatly increase performance by specifying virtio among other guest tweaks.
	Use the command "osinfo-query os" to get the list of the accepted OS variants.

# apt-get install libosinfo-bin
$ osinfo-query os
 Short ID             | Name                                               | Version  | ID
 centos7.0            | CentOS 7.0                                         | 7.0      |
 debian9              | Debian Stretch                                     | 9        |
 debiantesting        | Debian Testing                                     | testing  |
 fedora25             | Fedora 25                                          | 25       |
 freebsd11.0          | FreeBSD 11.0                                       | 11.0     |
 openbsd5.8           | OpenBSD 5.8                                        | 5.8      |
 rhel6.8              | Red Hat Enterprise Linux 6.8                       | 6.8      |
 rhel7.0              | Red Hat Enterprise Linux 7.0                       | 7.0      |
 rhel7.1              | Red Hat Enterprise Linux 7.1                       | 7.1      |
 rhel7.2              | Red Hat Enterprise Linux 7.2                       | 7.2      |
 ubuntu16.04          | Ubuntu 16.04                                       | 16.04    |
 ubuntu14.04          | Ubuntu 14.04 LTS                                   | 14.04    |
 win10                | Microsoft Windows 10                               | 10.0     |
 win2k12              | Microsoft Windows Server 2012                      | 6.3      |
 win2k12r2            | Microsoft Windows Server 2012 R2                   | 6.3      |
 win2k3               | Microsoft Windows Server 2003                      | 5.2      |
 win2k3r2             | Microsoft Windows Server 2003 R2                   | 5.2      |
 win2k8               | Microsoft Windows Server 2008                      | 6.0      |
 win2k8r2             | Microsoft Windows Server 2008 R2                   | 6.1      |
 win7                 | Microsoft Windows 7                                | 6.1      |
 win8                 | Microsoft Windows 8                                | 6.2      |
 win8.1               | Microsoft Windows 8.1                              | 6.3      |
 winvista             | Microsoft Windows Vista                            | 6.0      |
 winxp                | Microsoft Windows XP                               | 5.1      |

# virt-install \
--name=centos7 \
--disk pool=data-libvirt,cache=none,format=qcow2,size=16 \
--os-variant=centos7.0 \
--cdrom /home/paulgorman/Downloads/CentOS-7-x86_64-Minimal-1611.iso \
--vcpus=2 \
--ram=1024 \
--graphics spice \
--network bridge=br0

Enable the Serial Console

In the guest, edit /etc/default/grub, and append this to the GRUB_CMDLINE_LINUX=“” value: console=tty0 console=ttyS0,115200. On Red Hat, run grub2-mkconfig -o /etc/grub2.cfg, and reboot. On Debian, run update-grub, and reboot.

# virsh console centos7-test

Find the IP address of a guest (without a console connection)

# virsh domiflist centos7
Interface  Type       Source     Model       MAC           
vnet0      bridge     br0        virtio      52:54:00:80:ef:da
# nmap -sn | grep -i -B2 52:54:00:80:ef:da
Nmap scan report for                            
Host is up (-0.087s latency).                              
MAC Address: 52:54:00:80:EF:DA (QEMU virtual NIC)

(Or ip nei show dev br0?)

Attach More storage to a guest

# virsh attach-disk myguest /var/lib/libvirt/images/file.img vdb --cache none


Move VM to Another Host Manually

host-a~# virsh shutdown my-vm
host-a~# virsh dumpxml my-vm > my-vm.xml
host-a~# scp my-vm.xml host-b:
host-a~# scp /var/lib/libvirt/images/my-vm.qcow2 host-b:/var/lib/libvirt/images/
host-a~# virsh undefine my-vm
host-a~# rm /var/lib/libvirt/images/my-vm.qcow2

host-b~# virsh define my-vm.xml
host-b~# virsh start my-vm

Live Migration

Live migration should be possible, even without shared storage. Without shared storage, it’s necessary to pre-define identical storage at the destination. Know the root password for the destination (is there no way to “sudo” this? (So, remote root ssh logins need to be enabled?!) It may be necessary to open TCP ports 49152-49215 on the destination (or without opening the ports by passing the --tunnelled flag, if available).

host-b# iptables -A INPUT -p tcp --match multiport --dports 49152:49215 -j ACCEPT

host-a~# virsh vol-list vg0
Name                 Path
myvm.img             /dev/vg0/myvm.img
doorentry            /dev/vg0/doorentry
lvroot               /dev/vg0/lvroot
lvswap               /dev/vg0/lvswap
postoffice-data      /dev/vg0/postoffice-data
postoffice-os        /dev/vg0/postoffice-os
virtbackup           /dev/vg0/virtbackup
wolf                 /dev/vg0/wolf
host-b~# sudo virsh vol-create-as vg0 myvm.img 10G

host-a~# virsh migrate --copy-storage-all --persistent --live my-vm qemu+ssh://host-b/system

A migrate a non-running guest:

host-a~# virsh migrate --copy-storage-all --persistent --offline my-vm qemu+ssh://host-b/system

Spice/VNC console

# apt-get install spice-client-gtk
# virsh domdisplay my-vm
$ spicy --uri=spice://

# virsh vncdisplay my-vm

(“:1” seems to mean 5901.)

Once we know the guest console port, SSH forwarding works for Spice or VNC too.

$ ssh -L <local port>:localhost:<hypervisor port>
$ ssh -L 5900:localhost:5900
$ spicy --uri=spice://

Or, through a jump host:

$ ssh -A -t -L 7900:localhost:6900 ssh -A -t -L 6900:localhost:5900
$ spicy --uri=spice://

$ vncviewer


# apt-get install virt-top

“3” shows block devices.

VirtIO Drivers for Windows Guests

Installing the network driver is straightforward.

  1. Shut off the VM. (I guess we could add a second interface to the live VM with virsh attach-interface.)

  2. Change the network interface model from “e1000” (or “rtl8139” or whatever) to “virtio”. Do this via virsh edit myvm.

  3. Spin up the VM.

  4. Attach the VirtIO driver ISO to the image, and add the new drivers from inside the guest.

    $ wget

    virsh attach-disk myvm ~/virtio-win.iso hdb –type cdrom

    virsh domiflistm myvm

Installing the VirtIO disk bus driver is slightly less straightforward, but not a huge deal. We need to trick Windows into loading the driver by adding a small, temporary volume of the “VirtIO” type.

  1. Create a stub volume, and attach it to the VM. We probably need to restart the guest.
  2. Attach the VirtIO driver ISO to the image. In the guest’s Device Manager, load the VirtIO driver for the new storage controller device.
  3. Shut down the VM, and change the type of the main storage volume from IDE to VirtIO. Remove the stub volume.
  4. Start the guest. That’s it.
# virsh vol-create-as vg0 stub 1G
# virsh attach-disk doorentry /dev/vg0/stub vdc --config

# virsh detach-disk doorentry hdc --config
# virsh vol-delete stub --pool vg0

(Newer versions of virsh attach-disk can specify the bus explicitly with a --targetbus virtio flag. Older version only inferred the bus type based on the target name, like “hda”, “sdb”, or “vdc”.)

Delete an unwanted VM

#  virsh undefine --nvram --remove-all-storage --delete-snapshots myvm