QemuNet uses standard QEMU disk file in raw format. So you can follow any tutorial available for QEMU. Here is a simple example on how to install a debian8 system for QEMU.

Depending on your host configuration, it is recommanded to build light VM system in text mode, without graphical interface and heavy desktop manager!

# download the iso for your favorite system
wget http://cdimage.debian.org/debian-cd/8.x.x/amd64/iso-cd/debian-8.x.x-amd64-netinst.iso

# create a disk image for your VM (8G)
qemu-img create debian.img 8G

# install your VM from cdrom
qemu-system-x86_64 -enable-kvm -hda debian.img -cdrom debian-8.3.0-amd64-netinst.iso -boot d -m 512

# after installation is complete, reboot your VM and test it
qemu-system-x86_64 -enable-kvm -hda debian.img -m 512

Nota Bene : If you would like to copy/paste this latter line, unfortunately this is not possible with the standard QEMU display based on SDL. However, the xterm QemuNet option, described in the previous section will help you

In order to use your debian8 in QemuNet, you need to add a new entry in qemunet.cfg.

qemunet.cfg

# IMGDIR="/absolute/path/to/raw/system/images" 
if [ -z "$IMGDIR" ] ; then IMGDIR="$(dirname $(realpath $0))" ; fi
 
# debian testing
SYS[debian8]="linux"
FS[debian8]="$IMGDIR/images/debian8/debian.img"
QEMUOPT[debian8]="-localtime -m 512"

Now, you can test your system in QemuNet in standalone mode. This mode is useful to complete your installation before to use it for virtual networking. In this mode, you have a direct access to disk image in raw format.

./qemunet.sh -l debian8

If you need to resize your VM image file, for instance, to add 1GB…

qemu-img resize debian.img +1G

Then after booting the VM, you can use fdisk to remove all partitions (command 'd') and create a new Linux one (n) with default options. Check /etc/fstab and correct it if required. The reboot your VM.

After that, use resize2fs on the / partition:

df -h /                # -> /dev/sda1
resize2fs /dev/sda1

More details: https://gist.github.com/larsks/3933980

Removing the root password in your VM will make your life easier…

passwd -d root

# networking
apt-get install emacs-nox jed vim nano python-scapy tcpdump tshark dnsmasq dsniff xterm apache2 telnetd ftpd openvpn openssl vlan bridge-utils util-linux manpages bash bash-completion locales-all tmux screen expect zip unzip git moreutils net-tools socat rlwrap wget curl tcpdump netcat-openbsd
# python & scapy
apt-get install python3 python3-pip python python-pip python3-scapy python-scapy ipython ipython3
pip install --upgrade pip
pip3 install --upgrade pip

First, edit “/etc/apt/sources.list” and replace it with the following lines to use an unstable release.

deb http://ftp.fr.debian.org/debian/ unstable main contrib non-free
deb-src http://ftp.fr.debian.org/debian/ unstable main contrib non-free

Then update your system…

apt-get update
apt-get upgrade
apt-get install linux-image-xxx  # to get the most recent linux kernel
apt-get autoremove
apt-get autoclean

Edit /etc/network/interfaces and remove all the configuration lines for eth0. It will prevent some system boot problem with different network topolgy, that are not connected to Internet! However, as it will disable DHCP, you will require to launch the dhclient manually in standalone mode:

dhclient eth0

rm /etc/resolv.conf
history -c
poweroff

QemuNet provides some advanced options, that are specific for “linux” system.

In order to use the classic naming convention (eth0, eth1, …)

$ echo 'GRUB_CMDLINE_LINUX="net.ifnames=0"' >>/etc/default/grub
$ grub-mkconfig > /boot/grub/grub.cfg

The mount option (-m, by default) is useful to share files between host and guest system files. In your QemuNet session directory (-s session_dir), you will find a directory for each VM that is connected to /mnt/host in the VM system file. For instance, let consider a VM named “myvm”: the host directory “session/myvm/” is directly linked to the “/mnt/host/” directory of the myvm's system file and you got read/write permission.

To do this, you just need to add the following line in /etc/fstab of your VM system file:

host  /mnt/host  9p  trans=virtio,defaults,nofail  0  0

Dont forget to create /mnt/host directory. Here, nofail is important to enable the VM to boot without the '-m' option.

For more details on this option in QEMU, please visit : http://wiki.qemu.org/Documentation/9psetup

The xterm option (-x) is useful to launch your Linux VM in no-graphic mode inside xterm, instead of the native SDL mode. For instance, it allows copy/paste with other VM xterm windows. Moreover, you can resize the xterm window without any problem. Simply, type commands: resize and/or reset to update xterm.

To do this, you need to extract both kernel and initrd files from the VM raw disk image, and to provide it to QemuNet configuration file.

Basically, you can get these files by copying them from /boot to the /mnt/host shared directory (see previous tips). Easy

Else, you can try the following instructions on your host, that requires root privilege.

sudo losetup /dev/loop0 debian.img
sudo kpartx -a /dev/loop0
sudo mount /dev/mapper/loop0p1 /mnt
cp /mnt/boot/initrd.img-4.3.0-1-amd64 initrd
cp /mnt/boot/vmlinuz-4.3.0-1-amd64 vmlinuz

Then, update qemunet.cfg with KERNEL and INITRD arrays.

qemunet.cfg

# IMGDIR="/absolute/path/to/raw/system/images" 
if [ -z "$IMGDIR" ] ; then IMGDIR="$(dirname $(realpath $0))" ; fi
 
# debian8
SYS[debian8]="linux"
FS[debian8]="$IMGDIR/images/debian8/debian.img"
QEMUOPT[debian8]="-localtime -m 512"
KERNEL[debian8]="$IMGDIR/images/debian8/vmlinuz" # optional, but useful for -x option
INITRD[debian8]="$IMGDIR/images/debian8/initrd"  # optional, but useful for -x option

It is very useful to execute a user-defined script on startup, as for instance /mnt/host/start.sh. Indeed, it is a flexible way to configure each VM (hostname, ip, route, …) without using qcow2 image.

Make sure your start.sh script is executable and the first line is #!/bin/sh. Then create the start.service file in /etc/systemd/system as follow.

start.service

[Unit]
Description=startup script
Requires=mnt-host.mount
After=mnt-host.mount
 
[Service]
Type=oneshot
ExecStart=/mnt/host/start.sh
 
[Install]
WantedBy=multi-user.target     # or default.target?

Once you're done with the files, enable the service:

systemctl daemon-reload        # rerun all generators to recreate the entire dependency tree
systemctl enable start.service

Then you can launch it as follows:

systemctl start start.service

Pour afficher toutes les units :

systemctl list-units
systemctl list-units --all
systemctl list-units --type target --all

Check critical chains… and try to optimize it…

systemd-analyze critical-chain

systemd-analyze blame

For instance, you can change the default target to disable graphical boot…

systemctl get-default   # => graphical.target
systemctl set-default multi-user.target
systemctl get-default   # => multi-user.target

It is very useful to execute a rc.local script on startup, that launches a user-defined script /mnt/host/start.sh located in the shared directory (see previous tip). It is a flexible way to configure each VM (hostname, ip, route, …) without using qcow2 image.

/etc/init.d/rc.local

#!/bin/sh -e
MYSTART="/mnt/host/start.sh"
if [ -e "$MYSTART" ] ; then sh $MYSTART ; else exit 1 ; fi

/mnt/host/start.sh

#!/bin/bash
echo mydebian > /etc/hostname

Then set up the rc.local script…

chmod +x /etc/init.d/rc.local
chmod +x /mnt/host/start.sh
update-rc.d rc.local defaults

I want to resize my debian image from 2G to 3G…

First:

$ qemu-img resize debian10.img +1G

Then, boot Qemu on the raw image:

$ ./qemunet.sh -l debian10

Puis en tant que root:

$ cat /proc/partitions   
  /dev/sda => 3G
  /dev/sda1 => 2G
$ fdisk -l
$ cfdisk /dev/sda # delete partition and create a new one of 3G
$ resize2fs /dev/sda1
$ reboot