autopkgtest-qemu(1) autopkgtest virtualisation server using QEMU


autopkgtest [...] -- qemu [options] image [ro-image ...]


autopkgtest-qemu provides an autopkgtest virtualisation server using QEMU.

Normally autopkgtest-qemu will be invoked by autopkgtest.

You always need to specify at least one QEMU VM image. If your VM needs more disks, you can specify arbitrarily many additional images which will be added in order.

autopkgtest-qemu does not modify the given images, but will instead create a temporary overlay for the primary image, and add all other images as read-only.

The first image without the overlay is always added as an additional read-only hard drive, which will be available for tests as /dev/baseimage. This allows tests that require nested VMs to reuse the same image. Be aware that /dev/baseimage will not be accessible between calling autopkgtest-reboot-prepare and the next boot, thus make sure to stop accessing it before.


autopkgtest-qemu assumes that you have already prepared a suitable Debian based QEMU image (see BUILDING IMAGES below for how to create them). But you can use any VM which has python3 or python installed and either

opens a root shell on ttyS1, or

has a serial console on ttyS0 where you can log in (getty or similar) with specified user and password. This will then be used to start a root shell on ttyS1, to reduce this to the first case and to not assume that ttyS0 stays operational throughout the whole test.


-u user | --user=user

User name for running tests that don't require root privileges.

If the VM is not already prepared to start a root shell on ttyS1 this will be also used for logging in on ttyS0. If this is not root, it must be able to run sudo without a password to become root for that purpose.

-p password | --password=password
Password for above user for logging into the VM on ttyS0. This is necessary if the VM is not already prepared to start a root shell on ttyS1.

-q command | --qemu-command=command
QEMU command to run. This defaults to the qemu-system-* that matches your system architecture.

-o dir | --overlay-dir=dir
Directory where the temporary image overlay for the primary image is created. By default this will happen in a work directory in /tmp (or $TMPDIR), but you can use this to e. g. ensure that the overlay is placed on a tmpfs if your /tmp directory is not on tmpfs. This will greatly increase the speed.

-c num | --cpus=num
Number of (virtual) CPUs in the VM. Default is 1.

VM RAM size in MiB. Default is 1024, i. e. 1 GiB.

Show boot messages from serial console.

-d | --debug
Enable debugging output.

Pass through arguments to QEMU command; e. g. --qemu-options='-readconfig qemu.cfg'


If you use lots of options or images, you can put parts of, or the whole command line into a text file, with one line per option. E. g. you can create a file sid-desktop.cfg with contents like

-utestuser -ps3kr1t --ram-size=4096 /home/bob/autopkgtest/sid-desktop.img

and then run

autopkgtest [...] -- qemu @sid-desktop.cfg

The contents of the configuration file will be expanded in-place as if you would have given its contents on the command line. Please ensure that you don't place spaces between short options and their values, they would become a part of the argument value.


The behaviour of autopkgtest-qemu is as described by the AutomatedTesting virtualisation regime specification.


autopkgtest does not run apt-get update at the start of a package build, which can cause a build failure if you are using a development series template. You will need to run apt-get update in the template yourself (e. g. using --setup-commands).

If the --qemu-command option has not been specified and no custom CPU type was selected in --qemu-options, autopkgtest-qemu will try to enable nested KVM support by default on x86_64 platforms if the hardware supports this. To fully enable this, one needs to additionally set some module parameters on the host, by creating a file /etc/modprobe.d/nested_kvm.conf with the contents

options kvm_intel nested=1 options kvm_amd nested=1

and rebooting or reloading the KVM modules. It is still possible to use QEMU in tests without this, albeit without hardware acceleration. On Ubuntu systems these module options are typically already set.



For Debian you can use vmdebootstrap(8) to build a suitable image. E. g. for unstable:

vmdebootstrap --verbose --serial-console --distribution=sid \

   --customize=/usr/share/autopkgtest/setup-commands/setup-testbed \

   --user=test/test --size=10000000000 --grub --image=autopkgtest-sid.raw
qemu-img convert -O qcow2 autopkgtest-sid.raw autopkgtest-sid.img
rm autopkgtest-sid.raw

vmdebootstrap can only create a raw format image; it is recommended to convert it to qcow2 as that is much smaller (in the order of 700 MB for qcow2 vs. the 10 GB as specified above) and also supports additional features such as snapshots or compression. You can run that command with setting the environment variable AUTOPKGTEST_APT_PROXY to a proxy which will be used by apt in the VM. If you have an apt proxy configured on the host, this will be used automatically; otherwise you can run e. g.

AUTOPKGTEST_APT_PROXY= vmdebootstrap [...]


For Ubuntu, autopkgtest provides autopkgtest-buildvm-ubuntu-cloud(1) to build a VM based on the Ubuntu cloud images. To create an image for the current development series and the i386 architecture:

autopkgtest-buildvm-ubuntu-cloud -v -a i386

This will produce e. g. autopkgtest-trusty-i386-cloud.img.


Run the tests of the gdk-pixbuf source package, using an Ubuntu cloud image:

autopkgtest gdk-pixbuf -- qemu autopkgtest-trusty-i386-cloud.img


autopkgtest-qemu was written by Martin Pitt <[email protected]>

This manpage is part of autopkgtest, a tool for testing Debian binary packages. autopkgtest is Copyright (C) 2006-2014 Canonical Ltd and others.

See /usr/share/doc/autopkgtest/CREDITS for the list of contributors and full copying conditions.