Lab46 Wiki

This document will detail the steps needed to successfully create Xen VMs manually using the Debian (6.0) Squeeze distribution, which, at the time of writing, is neither supported by xen-tools nor officially released.

The procedure for installing other distributions will likely be quite similar.

Running a bleeding edge distribution in a VM is often far more likely than also running said bleeding edge distribution as the dom0. In the case of Debian, nothing past Lenny at the time of this writing has stable Xen dom0 support, so the dom0 is likely something older and more stable.

This will not stop us from running what we want in a VM, however.

In the LAIR, our Xen data directory structure is as follows:

/xen
    /boot
    /conf
    /domains
    /images
    /save

We configure xen-tools to create new VM images in /xen/domains, and place config files in /xen/conf, then typically migrate the VM images to /xen/images, and make the appropriate corrections in the config file.

Here, will just go make everything correctly from the start.

There are typically 2 images needed for a standard virtual machine- the main disk and the swap.

Here we will make a 4GB disk image:

vmserver:/xen/images# dd if=/dev/zero of=newvm.disk bs=1M count=4096
4096+0 records in
4096+0 records out
4294967296 bytes (4.3 GB) copied, 91.3332 s, 47.0 MB/s
vmserver:/xen/images# 

We'll also want to make swap (here we'll make a 256MB swap disk):

vmserver:/xen/images# dd if=/dev/zero of=newvm.swap bs=1M count=256
256+0 records in
256+0 records out
268435456 bytes (268 MB) copied, 7.20193 s, 37.3 MB/s
vmserver:/xen/images# 

Before we can install, we need to put a filesystem or some structure on our disks.

ere we will format our main disk image:

vmserver:/xen/images# mkfs.ext3 -j -O dir_index -v ./newvm.disk
mke2fs 1.41.3 (12-Oct-2008)
./newvm.disk is not a block special device.
Proceed anyway? (y,n) y
fs_types for mke2fs.conf resolution: 'ext3', 'default'
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
262144 inodes, 1048576 blocks
52428 blocks (5.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=1073741824
32 block groups
32768 blocks per group, 32768 fragments per group
8192 inodes per group
Superblock backups stored on blocks: 
        32768, 98304, 163840, 229376, 294912, 819200, 884736

Writing inode tables: done                            
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 27 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.
vmserver:/xen/images# 

And be sure to bless the swap disk:

vmserver:/xen/images# mkswap newvm.swap 
Setting up swapspace version 1, size = 268431 kB
no label, UUID=a22e7d04-d2ea-4a6b-99db-4941c291648f
vmserver:/xen/images# 

Once created, we can local loopback mount the disk image, and install our distro of choice (since we're going to use squeeze, we'll debootstrap it).

Using a loopback mount, we mount the disk image so we can dump files onto it (ie install).

vmserver:/xen/images# mount -o loop /xen/images/newvm.disk /mnt
vmserver:/xen/images# 

Once mounted, proceed with the installation:

vmserver:/xen/images# debootstrap squeeze /mnt http://mirror/debian
I: Retrieving Release
I: Retrieving Packages
I: Retrieving Packages
I: Validating Packages
I: Resolving dependencies of required packages...
I: Resolving dependencies of base packages...
I: Found additional required dependencies: insserv libbz2-1.0 libdb4.8 libslang2 
I: Found additional base dependencies: libnfnetlink0 libsqlite3-0 libudev0 
I: Checking component main on http://mirror/debian...
I: Retrieving libacl1
I: Validating libacl1
I: Retrieving adduser
I: Validating adduser
I: Retrieving apt-utils
I: Validating apt-utils
I: Retrieving apt
I: Validating apt
...
I: Configuring tasksel-data...
I: Configuring tasksel...
I: Base system installed successfully.
vmserver:/xen/images# 

Once we're done with the install, we should tend to some configuration matters.

Debian systems use /etc/network/interfaces for network configuration… if we're gunning for a standard VM, we'll want something like this (could likely just copy it from the VM server):

# This file describes the network interfaces available on your system
# and how to activate them. For more information, see interfaces(5).

# The loopback network interface
auto lo
iface lo inet loopback

auto eth0
iface eth0 inet dhcp

Debian's package management system builds its database from repositories configured in /etc/apt/sources.list; we should be mindful to set the VM's sources.list to use any local services (such as our in-lab repository).

#####################################################################
#
#  /etc/apt/sources.list
#
deb     http://mirror/debian    squeeze           main contrib non-free
deb-src http://mirror/debian    squeeze           main contrib non-free
#deb     http://mirror/security squeeze/updates   main contrib non-free
#deb-src http://mirror/security squeeze/updates   main contrib non-free
deb     http://mirror/lair      squeeze           main

In many cases this could also be straight copied from the VM server (but remember to change the distro to “squeeze”.. the VM server is likely running “lenny”).

Now we'll need to do some editing on the VM itself, so we're going to trick it into thinking it is running. We do this by using chroot to make the VM's filesystem the root of our shell session's filesystem:

vmserver:~# chroot /mnt
root@vmserver:/# 

At this point, an “ls /” will show the base of the VM's filesystem, and not the VM server. The prompt still says “vmserver”, but this is a lie. We will actually be changing some files to alter that.

The /proc filesystem will need to be mounted in order to allow some facilities to work properly:

root@vmserver:/# mount -t proc proc /proc
root@vmserver:/# 

Now we are to change the hostname that the VM reports:

root@vmserver:/# echo "vmname" > /etc/hostname
root@vmserver:/# 

And set the root password:

root@vmserver:/# passwd
Enter new UNIX password: 
Retype new UNIX password: 
passwd: password updated successfully
root@vmserver:/# 

Read from /etc/apt/sources.list and establish our package database (ends up being in /var/lib/dpkg):

root@vmserver:/# aptitude update
Get:1 http://mirror squeeze Release.gpg [835 B]
...
Current status: 584 new [+584].
root@vmserver:/# 

There should not be any updates to apply, since we just recently installed, but it never hurts to check:

root@vmserver:/# aptitude upgrade
No packages will be installed, upgraded, or removed.
0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
Need to get 0 B of archives. After unpacking 0 B will be used.
...
root@vmserver:/# 

If we're running a fairly standard VM (ie nothing experimental that needs some special kernel), here are the default packages we'd likely want to install:

root@vmserver:/# aptitude install ssh linux-image-2.6.32-5-xen-amd64 locales console-data
The following NEW packages will be installed:
  busybox{a} console-common{a} console-data firmware-linux-free{a} initramfs-tools{a} 
  kbd{a} klibc-utils{a} libbsd0{a} libedit2{a} libgssapi-krb5-2{a} libk5crypto3{a} 
  libkeyutils1{a} libklibc{a} libkrb5-3{a} libkrb5support0{a} libuuid-perl{a} 
  libwrap0{a} libx11-6{a} libx11-data{a} libxau6{a} libxcb1{a} libxdmcp6{a} 
  libxext6{a} libxmuu1{a} linux-base{a} linux-image-2.6.32-5-xen-amd64 locales 
  openssh-blacklist{a} openssh-blacklist-extra{a} openssh-client{a} openssh-server{a} 
  ssh tcpd{a} xauth{a} 
0 packages upgraded, 34 newly installed, 0 to remove and 0 not upgraded.
Need to get 43.7 MB of archives. After unpacking 143 MB will be used.
Do you want to continue? [Y/n/?] y
Get:1 http://mirror/debian/ squeeze/main console-data all 2:1.10-5 [1180 kB]
Get:2 http://mirror/debian/ squeeze/main kbd amd64 1.15.2-1 [404 kB]
Get:3 http://mirror/debian/ squeeze/main console-common all 0.7.85 [128 kB]
...

You've probably been seeing these messages appear:

locale: Cannot set LC_CTYPE to default locale: No such file or directory
locale: Cannot set LC_MESSAGES to default locale: No such file or directory
locale: Cannot set LC_ALL to default locale: No such file or directory

What is happening is that the locale environment variables are not properly set… time to get that taken care of:

root@vmserver:/# dpkg-reconfigure locales console-data

And be sure to check:

[*] en_US.UTF-8 UTF-8

And make the “Default locale for the system environment”: en_US

Usually there's no need to touch the keymap, so when it asks: “Don't touch keymap”

Also important is the timezone, which can be configured as follows:

root@vmserver:/# ln -sf /usr/share/zoneinfo/US/Eastern /etc/localtime
root@vmserver:/# 

Before we wrap up, we should ensure we will be able to access the VM's console; instead of /dev/tty1, Xen has us using /dev/hvc0, so be sure to change that in /etc/inittab. Specifically, we want to look for this block:

1:2345:respawn:/sbin/getty 38400 tty1
2:23:respawn:/sbin/getty 38400 tty2
3:23:respawn:/sbin/getty 38400 tty3
4:23:respawn:/sbin/getty 38400 tty4
5:23:respawn:/sbin/getty 38400 tty5
6:23:respawn:/sbin/getty 38400 tty6

And change it to this:

1:2345:respawn:/sbin/getty 38400 hvc0
#2:23:respawn:/sbin/getty 38400 tty2
#3:23:respawn:/sbin/getty 38400 tty3
#4:23:respawn:/sbin/getty 38400 tty4
#5:23:respawn:/sbin/getty 38400 tty5
#6:23:respawn:/sbin/getty 38400 tty6

Hey now, it looks like we're on set on VM configuration, so let's umount and exit from our chroot environment:

root@vmserver:/# umount /proc
root@vmserver:/# exit
exit
vmserver:/xen/images# umount /mnt
vmserver:/xen/images# 

Now it is time to create our Xen VM configuration file in /xen/conf… each config has unique characteristics in it, here is a sample file that can be altered:

##########################################################################
# LAIR Xen VM configuration file
##########################################################################

#################################################
#  Kernel + memory size
#
kernel  = '/xen/boot/vmlinuz-2.6.32-5-xen-amd64'
ramdisk = '/xen/boot/initrd.img-2.6.32-5-xen-amd64'
memory  = '192'
vcpus   = '1'
#extra  = 'clocksource=jiffies'

#################################################
#  Disk device(s).
#
root    = '/dev/xvda1 ro'
disk    = [ 'file:/xen/images/newvm.disk,xvda1,w',
            'file:/xen/images/newvm.swap,xvda2,w'  ]

#################################################
#  Hostname
#
name    = 'newvm'

#################################################
#  Networking
#
dhcp    = 'dhcp'
vif     = [ 'mac=00:16:3E:AA:BB:CC' ]

#################################################
#  Behaviour
#
on_poweroff = 'destroy'
on_reboot   = 'restart'
on_crash    = 'restart'

Look through the file and make the pertinent changes (memory, disk names/location, VM name, MAC address).

If this is to be a regular service provided (or even if it is just an experiment), efforts should be documented so others know the purpose of the VM, and the originator has a record to recall when they've forgotten some of the specifics of their activities.

If desired, you can:

• go to the sample documentation page
• view the wiki page source
• copy it (to your clipboard)
• create a new page under the system namespace (named for the fully qualified hostname)
• edit its wiki source, and paste
• fill out that page as appropriate

If all our ducks are in a row, let's fire up our VM:

vmserver:~# xm create -c /xen/conf/newvm.cfg
Using config file "/xen/conf/newvm.cfg".
...