Having spent the last 24 hours trying to get Proxmox to play nice with the new VRack 1.5, it looks like it works perfectly, including online live migration of venet based OpenVZ containers, which didn’t work in VRack 1.0.

The configuration makes eth1 the default card for traffic from vmbr0, however allows eth0 to function alongside it so that you don’t loose out on monitoring features. We also route IPv6 traffic through the VRack on vmbr0 and add additional IP ranges for your VM use to vmbr0.

All the configuration that’s needed is done in: /etc/network/interfaces.

Here is my resulting configuration:

# 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

# for Routing
auto vmbr1
iface vmbr1 inet manual
    post-up /etc/pve/kvm-networking.sh
    bridge_ports dummy0
    bridge_stp off
    bridge_fd 0

# vmbr0: Bridging. Make sure to use only MAC adresses that were assigned to you.
auto vmbr0
iface vmbr0 inet static
    address 94.23.XXX.10
    network 94.23.XXX.0
    broadcast 94.23.XXX.255
    gateway 94.23.XXX.254
    bridge_ports eth1
    bridge_stp off
    bridge_fd 0
# A secondary IP subnet used for VMs
    up /sbin/ip route add 178.XXX.YYY.128/26 dev vmbr0
        up /sbin/ip route flush cache

#VRack IPv6
iface vmbr0 inet6 static
        address 2001:41d0:XXXX:6810::10
        netmask 56
        post-up /sbin/ip -f inet6 route add 2001:41d0:XXXX:68ff:ffff:ffff:ffff:ff7f dev vmbr0
        post-up /sbin/ip -f inet6 route add default via 2001:41d0:XXXX:68ff:ffff:ffff:ffff:ff7f
        pre-down /sbin/ip -f inet6 route del default via 2001:41d0:XXXX:68ff:ffff:ffff:ffff:ff7f
        pre-down /sbin/ip -f inet6 route del 2001:41d0:XXXX:68ff:ffff:ffff:ffff:ff7f dev vmbr0

auto eth0
iface eth0 inet static
    address 5.XXX.YYY.25
    broadcast 5.XXX.YYY.255
    #Setting up the routing
    up /sbin/ip route flush table 80
    up /sbin/ip route add table 80 to 5.XXX.YYY.0/24 dev eth0
    up /sbin/ip route add table 80 to default via 5.XXX.YYY.254 dev eth0
    up /sbin/ip rule add from 5.XXX.YYY.0/24 table 80 priority 80
    up /sbin/ip route flush cache
    post-down /sbin/ip route flush table 80
Read more

I’ve been spending this morning optimizing the Flosoft.biz website in terms of load times in Browsers, and one key element of that is sending the correct expires headers to allow Browsers to cache the data.

Now, as of Plesk 11.5, you can edit nginx settings via the Control Panel, but this isn’t always straight forward, so I thought I’d write a small tutorial.

In the Control Panel:

  1. Select your Domain
  2. Click Web Server Settings
  3. Scroll down to nginx settings
  4. If you have “Serve static files directly by nginx” checked (which I recommend), you’ll need to remove the file extensions you’re going to use below, such as jpg,gif,…
  5. In the text box “Additional nginx directives” copy / paste the following configuration:

location ~* \.(js|css|png|jpg|jpeg|gif|ico)$ {
expires 30d;
add_header Pragma public;
add_header Cache-Control “public”;
try_files $uri @fallback;

That’s it. Just hit OK and enjoy a website that sends the correct headers for your static images and CSS.


Read more

One of the legacy systems we still use at Flosoft.biz is Plesk. Over the last few years it has slowly gotten better (don’t worry, it still completely breaks on every version upgrade) and nowadays comes with nginx.

However, I noticed that for some obscure reason, it doesn’t enable GZip compression for the webpages it serves? This is quite odd, having myself worked a lot with nginx over the last few years, it’s a default configuration!

Don’t worry, it’s quite easy to enable it though:

Just edit the following file as root: /etc/nginx/conf.d/gzip.conf

gzip on;
gzip_proxied any;
gzip_types text/plain text/xml text/css application/x-javascript;
gzip_vary on;
gzip_disable “msie6”;

Then run nginx -t to test the configuration and if that’s all ok, restart nginx by running /etc/init.d/nginx restart.

That’s it. Your webserver will now be serving your pages with GZip compression.

Read more

This just arrived from our ISP.

Good morning,

The consumption of our network has exceeded 100Gbps over the whole 280Gbps of total capacity that we have between us and the Internet.

We are the largest hoster in Europe in terms of use and of network capacity. (soon it will be the largest in hoster in terms of the number of servers hosted).

We are currently working to establish new connections, who will allow us to pass from 280Gbps to 350Gbps of network capacity with new peering points (Espanix, Tix, MIX, NIX, PL-IX, WIX, VIX).

Today and tomorrow we will install our infrastructure in 5 London
datacentres then increase the capacity of 1Gbps at Linx to 2x10G.
Next week we install Brussels at 2 datacentres and Amsterdam at 3 datacentres. Then Warsaw, Prague, Vienna, Milan, Zurick, Frankfurt and Madrid.

Also, we are interconnecting new capacity at Teleglobe, Global Crossing and Level3, Paris, London and Amsterdam.

We will interconnect any operator in 11 capitals to any peering point in the 11 capitals for the simple price structure of 100Mbps = 100Euro. And so if an AS lies in Frankfurt and wants 100Mbps on Espanix, the realization of the circuit takes 3 minutes (+ time to move the cables between our facilities and the customer and our facilities and peering point).

At the same time, we will start the tender transit 100Mbps = 100Euro also in the 11 capitals.

These services are oriented to small networks that now pay high prices for IP services precisely because they are small. We believe that when you’re little, you should pay little.

For our transiting, we use for Cisco MPLS layer and Force10 for the switching layer. It runs 10G. For transport between capitals we use our own network based on a Infinera 160x10G capacity. For transport in South and Eastern Europe  we use wavelengths 10G-based network Interoute. We expect to build links to own as soon as we need more 2x10G.

Cool ey?

Read more

It’s been a hell of a weekend until now. Travelling all across Belgium, and ofcourse the pre-launch of the Cold War Crisis mod.

Flosoft.biz provided 2 mirror servers with the release on it, so I thought I might post a small summary of the damage caused 🙂

The total data transferred was 184GB from both mirrors. This means about 600 people downloaded the mod in the first 24 hours.

Below you can find a few graphs of the mirrors. Please note that the scale changes.

If you haven’t had a chance to grab it, get it on Monday, when it’s officially released!

Read more


As you may have expected, Flosoft.biz has been working over the last few months on their new VPS offers. After nearly 2 months of work, they are ready, and have launched their offer. Their plan to revolutionize the VPS market has succeeded.

Their VPS08 beats the competition on all levels. For comparison we will use Rimu Hosting, VPS Link (Spry) and Media Temple.


RAM is often the main problem with VPS. The VPS08 offers 256MB Ram in it’s base model, and it can be extended to up to 8GB at a price of 4€ / 256MB. The 256MB model costs 9,99€.

The competition is a lot more expensive: Rimu Hosting’s machine Miro VPS2 has 256MB Ram too, and costs 29,95$; VPS Link’s Link 3 costs 24,95$; Media Temple, at 50$ for (dv) Base, is the most expensive of the 4 VPS options.


Harddisk storage is also often very limited on the VPS. The VPS08 comes with 20GB by default, which can be extended to up to 1TB at the price of 2,50€ / 5GB. Again, the machine costs 9,99€.

Rimu Hosting offers up to 16GB of storage on it’s Miro VPS4 at the price of 49,95$. VPS Link’s Link 4 comes with a 20GB Harddisk too, but at the price of 39,95$. Last, but definitely not least, Media Temple, it’s (dv) Base comes with 20GB at the price of 50,00$. Again, a clear advantage for the VPS08 of Flosoft.biz.

Connection & allowed traffic:

Don’t you hate when your machine gets shut down because you exceeded your traffic? Or that when you get your monthly invoice, you see that you have to pay a lot more than you thought because you exceeded your traffic limit? Well, Flosoft.biz doesn’t believe in them, and thus offers it’s VPS without any traffic limit. Again we are talking about 9,99€. This is not some fair use policy, or anything else, but real unlimited traffic.

Do we still have to look at the competition? Quickly said, no other ISP offers unlimited traffic. If you want more than 1TB, you are paying at least 150$ a month.


The VPS08 is a lot cheaper than any offers from Flosoft.biz’s competition. Get one today while they’re still available!


Read more

This howto will explain to you how to install a frontend on your server which is accessible via NX. This tutorial is for Debian based systems, and has been tested on an Flosoft.biz FlexServ (RPS).

1. Check the basic Debian setup.

We need to modify the sources.list

vi /etc/apt/sources.list

Add the following 2 lines:

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

Close and save (:wq) and run this command to update:

apt-get update

Then check if your system is up to date, and if necessary install updates.

apt-get dist-upgrade

2. Installing the X window manager xorg.

Just type in:

apt-get install xserver-xorg-core xorg

There will be a few questions at the end, for now just go with the defaults.

3. The Login Manager

Now you have 3 options. You can install any of the following Login Manager’s. Your options are:

  1. KDM
    KDM is probably the best if you want to use KDE
  2. GDM
    GDM is probably the best if you want to use Gnome
  3. XDM
    XDM is probably the best if you want to use Fluxbox or XFCE

Once you have chosen one of the GUIs run one of the following three commands:

apt-get install kdm
apt-get install gdm
apt-get install xdm

4. The GUI

Now again, you have a choice of different Graphical User Interfaces.

  1. KDE
    Personally my favourite on Debian
  2. Gnome
    My favourite on Ubuntu
  3. Fluxbox
    Never used it
  4. XFCE4
    Never used it

Once you have chosen one of the GUIs run one of the following three commands:

apt-get install kde
apt-get install gnome
apt-get install xfce4
apt-get install fluxbox

Thats all for the base setup.

5. Reboot

You should reboot to make sure the X server starts.

shutdown -r now

6. Create your user

Once your server has rebooted, and you have relogged in, you should create a user which you will use for the GUI.

adduser mynewusername

6. Getting the NX packages

Now we need to setup the NX server, so that you are able to connect to the server from your home. So you need to download the NX server pacakges:


7. Installing the NX packages

As you have the packages now in your directory, you need to install them via dpkg.

dpkg -i nxclient_3.1.0-2_i386.deb
dpkg -i nxnode_3.1.0-3_i386.deb
dpkg -i nxserver_3.1.0-2_i386.deb

8. The Services

Now we need to make sure the services are running.

/etc/init.d/ssh restart
/etc/init.d/nxserver restart

9. The Browser

Last, but not least… well actually least … Firefox! You will need a nice Browser, so Firefox is the way to go.

apt-get install firefox

Now you’re system is setup and you’re ready to use it. Simply setup your NX Client and have fun!

If you have any questions, don’t hesitate to ask me or just leave a comment.

Read more