Wiki Enseignement

Writing /net/www/auesnard/teaching/data/cache/4/4c5ab830b7838a3f2f1202639e310249.xhtml failed
This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== How to install a Debian Linux System for QemuNet? ======

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. 

==== Build a Standard QEMU Image ====

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

<code>
# 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

</code>

__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//.

<code bash 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"
</code>

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.

<code>
./qemunet.sh -l debian8
</code>

==== Resize your Image File ====

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

<code>
qemu-img resize debian.img +1G
</code>

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:

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

__More details__: https://gist.github.com/larsks/3933980


==== Some Basic Tips for QemuNet ====

==Remove root password== 

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

<code>
passwd -d root
</code>

==Install your packages==

<code bash>
# 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
</code>

==Upgrade your system==

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...

<code>
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
</code>


==Disable dynamic configuration for eth0== 

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:

<code>
dhclient eth0
</code>

==Closing standalone session==

<code>
rm /etc/resolv.conf
history -c
poweroff
</code>

==== Some Advanced Tips for QemuNet ====

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

==Naming of Network Interfaces ==

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

== Mount shared directory ==

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

== Run VM in xterm terminal ==

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.

<code>
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
</code>

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

<code bash 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
</code>

== Startup script with SystemD ==

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.

<code text 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?
</code>

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


== Startup script with rc.local (deprecated) ==

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.

<code bash /etc/init.d/rc.local>
#!/bin/sh -e
MYSTART="/mnt/host/start.sh"
if [ -e "$MYSTART" ] ; then sh $MYSTART ; else exit 1 ; fi
</code>

<code bash /mnt/host/start.sh>
#!/bin/bash
echo mydebian > /etc/hostname
</code>

Then set up the rc.local script...
<code>
chmod +x /etc/init.d/rc.local
chmod +x /mnt/host/start.sh
update-rc.d rc.local defaults
</code>


==== Resize your Qemu Image ====

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