So, let’s make a few assumptions here :

- You have a system running at least Ubuntu 10.4 Lucid Lynx, fully updated.
- You want to play World of Warcraft on it.
- You came here for an in-depth guide on how to get that done.

Alrighty, let’s get cracking then !

First thing you need to do is make sure that you have the right drivers installed for your video adapter and that you actually have hardware OpenGL support up and running on your system.

This is usually accomplished by running System->Administration->Hardware and then installing the proprietary hardware drivers for your particular videocard.

OpenGL hardware capabilities are also referred to as DRI, which means Direct Rendering Infrastructure. Without it, OpenGL will heavily lean on your CPU for all it’s calculations instead of your GPU. When you run a rather CPU intensive program, such as a computer game like World of Warcraft, you really do not want to do without it.

To check if you have DRI installed and running, type the following command in a shell :

$ glxinfo | grep rendering
direct rendering: Yes

If you get the response ‘direct rendering : Yes’ or something similar, you are up and running and can go ahead to install Wine.
If you get the response ‘direct rendering : No’, then you need to enable that first. Look for tutorials on how to do that as that is beyond the scope of this post.
If you get a message that the command is unknown, you need to install the mesa utils and try again :

$ sudo apt-get install mesa-utils

So now that you confirmed that you have hardware DRI, let’s go ahead and turn our attention to Wine.

Wine stands for ‘Wine Is Not an Emulator’. So it is not an emulator. It is merely a layer between your windows apps and the linux environment, allowing them to run natively.

So, now we have to install Wine. Do this in a shell or use the program manager to install it :

$ sudo apt-get install Wine

After you installed it, it is time to download the WoW installer or install it from your disks.

Whatever method you choose, use the windows version to install it. You can find the online installer here.

I will assume you downloaded it, and if you did, you need to first give it the executable bit by right-clicking the file and then go to the permissions tab. You can see what i mean on the image here to the left.

When you have set the execute bit, then you can make wine run it by rightclicking again and choosing ‘Open with Wine Windows Program Loader’.

I will assume everything went alright, and you get the installation screen, the EULA which you need to agree on etc. You can basically click through all that and install World of Warcraft like you would on a normal Windows PC.

After the complete installation, you will also have a new icon on your desktop called ‘World of Warcraft.desktop’ and its icon looks like a text file. This is actually the launcher for World of Warcraft, and you need to set the execute bit on it as well, just as you did with the installer executable earlier. When you do, the icon will change into the World of Warcraft icon that you expect.

Another thing you need to do, is get into the properties of the World of Warcraft launcher again and add -opengl to the commandline in it :

env WINEPREFIX="/home/username/.wine" wine "C:\Program Files\World of Warcraft\Launcher.exe" -opengl

This will ensure that World of Warcraft will start in OpenGL mode, which runs a lot better than DirectX mode on Linux.

This is all you need to do ! Happy Gaming !

If you happen to be on the EU realms, visit me on The Sha’Tar sometime. I will be playing as Diaenaeira

So, jumping off the cliff, I installed the 32bit version on my laptop, and to my surprise, the installation was quick and painless and the system booted up smoothly and almost everything just worked !

So instead of giving you a list of what works, here is a list of what does not work :

- Wireless network
- Fingerprint authentication
- Some hotkeys like rotate screen.
- Automatic screen rotation when put in tablet mode.

So this means that the touch screen capabilities of this laptop work out of the box. Well done developers at Ubuntu / Gnome / etc !

So, how do we get the other things to work then ?

First thing i did was to get the wireless up and running. You will need to download a few things so you need another pc with internet connection and a usb stick, or if you have dualboot you can use windows to download it or you can plug in a network cable as that works fine out of the box.

What you can do is just go to System -> Administration -> Hardware Drivers. BUT DON’T !

This will allow you to install the standard proprietary Broadcom STA drivers and you might even be lucky with those. However if you have a system identical to mine, you won’t be. The drivers installed fine, seemed to work, i can actually see my network neighborhood and it looks like it actually tries to authenticate itself to my wireless router.

However, that is basically all it does. It just does not work properly, so I uninstalled the drivers again and went for it manually, which is usually the way i got things working on this laptop/tablet

If you want to follow this guide, please reassure yourself that you are actually having a wireless network card of the type that this guide is intended for, or something really similar : The BCM4328. You can use the lspci command for that, which lists all attached PCI cards on your PC.

$ lspci
03:00.0 Network controller: Broadcom Corporation BCM4328 802.11a/b/g/n (rev 03)

If there is a line in the list that is similar to the one above, you can proceed following this guide. If it’s very different, I think it is better to just google for another guide or maybe just buy some USB wireless adapter that is supported by Linux.

The first thing you ought to do is install the ndiswrapper, and a tool called cabextract :

$ sudo apt-get install ndiswrapper-utils-1.9
$ sudo apt-get install cabextract

When you completed this, it’s time to download the drivers themselves. We will be using the windows drivers for this. (hence the installation of ndiswrapper)

The best you can do is to use HP’s own driver package to make this work. You can download it here :

http://h18000.www1.hp.com/support/files/hpcpqnk/us/download/24001.html

If you look at the list of what this driver is for, you won’t see the tx2000 listed there. I do not known why this is, but this is actually the driver you are looking for.

when you downloaded this, transfer it to your home folder if it is not already there and put it in a seperate folder. I will assume you took the last option as that is the easiest way to do things : put in a network cable.

$ mkdir wldrivers
$ cp SP34152.exe wldrivers
$ cd wldrivers
$ cabextract SP34152.exe

if you list that folder, you will see that the .exe is extracted and the folder now contains a few files including bcmwl5.inf, which is the one we are after.

Next thing you do is install the drivers using ndiswrapper :

$ sudo ndiswrapper -i bcmwl5.inf
$ sudo ndiswrapper -m
$ sudo depmod -a
$ sudo modprobe ndiswrapper

It should be working now already. you can check that by browsing your network neighbourhood by clicking the network icon on top bar on your desktop. If you do not see your own wireless router, unplug the network cable as apparantly it will ignore the wireless SSID that you are already connected to with cable and it doesn’t need to make the same connection twice.

Next thing you want to do is make sure you have wireless available to you the next time you reboot, and that is done by putting the ndiswrapper line in /etc/modules so that at startup the module will automatically be loaded into the kernel.

$ sudo gedit /etc/modules

Just put ‘ndiswrapper’ (without the quotes ofcourse) on a new blank line at the end of the file and you are set.

If it still doesn’t work for you at this point, simply reboot and try again. If it fails after that, you will need another guide than this one as you seem to have another wireless adapter installed. (which hp unfortunatly is really good at, making the same thing in dozens of versions, all with different hardware)

For the other stuff, i still have to figure out the why and how, but as soon as I did, I will add the instructions to those in a new post.

For now, you already should have a perfectly working system including touchscreen capabilities.

However, as always, the weak end of a secure system is always at the user end. Therefore we are going to need to tighten up OpenSSH even more than it already is, which is quite easy to do.

First thing you need to look at is if you are using an up to date version of OpenSSH. At the time of this writing, version 5.3 is the most recent one so when reading this post, it needs to be at least that.

To determine the version, type $ ssh -V

It will produce some output with a version number, most likely with added information about the distribution of linux that you are using.

When you ensured that your version is up to date, it’s time to get to work and tighten the thing down :

Limit allowed users

Another thing you really want to do is limit the users that are allowed to connect to SSH. You can also do that in the sshd_config file :

$ sudo pico /etc/ssh/sshd_config

Scroll down until you find a line that starts with ‘AllowUsers’. If it isn’t in there, add it at the end and put only those users behind it that you want to allow SSH connectivity for.

AllowUsers yourself user2 user3

Save the file and restart SSH by issuing the restart command.

$ sudo /etc/init.d/ssh restart

The command there might vary a bit depending on your distribution.

This will diminish the chance for scriptkiddies to guess username/password combinations.

Turn off .rhosts usage

You really don’t need this as SSH can emulate the behaviour of the rsh command. rsh became obsolete anyway.

$ sudo pico /etc/ssh/sshd_config

Scroll down until you find the command IgnoreRhosts. If it isnt there, add it at the end :

IgnoreRhosts yes

Save and restart your SSH daemon.

Turn off host based authentication

Probably the worst idea someone can have is authenticate based on the IP adress that is connecting. If someone successfully spoofs that particular IP adress, you are in trouble as you lost your last line of defense.

$ sudo pico /etc/ssh/sshd_config

Scroll down until you find the line that says HostbasedAuthentication. If there isnt any, add it at the end :

HostbasedAuthentication no

Save and restart your SSH daemon.

Disable root login via SSH

Ok ok, this is 2010, SSH is secure, there is no way of sniffing the password out of your packets, but we have the su and sudo (recommended) command. Why not play it safe and deny root login anyway. With su and sudo at least it gets logged who does what with root privileges. So I am still all for denying root logins.

$ sudo pico /etc/ssh/sshd_config

Scroll down until you find the line that says PermitRootLogin. If there isn’t any, add it at the end :

PermitRootLogin no

Save and restart your SSH daemon.

Change the port SSH listens on

By default, SSH listens on port 22. So this is also the port that brute force scripts target, while you can save yourself some resource load by letting the SSH port listen somewhere else, like, 300 or 400.

Before you do this, make sure you opened up that port in your firewall configuration first or you will lose SSH functionality when you restart the daemon and then can not log in again on the other port because the firewall will deny you access.

$ sudo pico /etc/ssh/sshd_config.

Look for a line that starts with Port. Change it to :

Port 300

Save and restart the SSH daemon.

I know this list is by all means not complete, but I welcome any additions anyone is willing to make.

The current list however ought to be enough to be discouraging enough for most attackers.

Most modern server distributions already have a more hardened stack, but most of the time, you can even tighten it up further, which is generally a good idea in my humble opinion.

This list is by no means complete and I welcome more suggestions for further tightening down.

Decrease the ARP cache cleanup interval

First thing you can do is set the interval on your ARP to reduce the risk of ARP attacks. Google if you want to know more about that as explaining how that works is beyond this scope and would take an article on its own.

If you are using linux, ‘# man 7 arp’ command will give you extensive information on what ARP is and how to manipulate it. What we are looking for is ‘gc_interval’ and ‘gc_stale_time’. If you are using a recent version of linux, gc_interval and gc_stale_time is defaulted to 30 and 60 seconds respectively. If that is the case, you don’t need to change anything. You can check the values by issueing the following commands :

# sysctl -a | grep gc_interval
# sysctl -a | grep gc_stale_time

(yes i realize there is also a more neat way to do this : sysctl -n for example, but we want to get all settings from all interfaces as you can have more than 1)

It should produce output similar to this :

net.ipv4.neigh.default.gc_stale_time = 60
net.ipv4.neigh.lo.gc_stale_time = 60
net.ipv4.neigh.eth0.gc_stale_time = 60

net.ipv4.route.gc_interval = 30

If these values are different, we are going to change them with the sysctl command again :

# sysctl -w net.ipv4.neigh.eth0.gc_stale_time=60
# sysctl -w net.ipv4.route.gc_interval=30

If you want to make that permanent, put those commands in your /etc/rc.sysinit or /etc/rc.local.

Disable ICMP broadcast echo

To prevent your server being used for something called a smurf attack you need to disable your ICMP broadcasting.

if you want to check the current setting, do

# sysctl -a | grep icmp_echo_ignore_broadcasts

if that produces anything like the following result, you have to turn it off.

net.ipv4.icmp_echo_ignore_broadcasts = 1

turn this off by issuing the following command and putting that in your /etc/rc.sysinit or /etc/rc.local.

# sysctl -w net.ipv4.icmp_echo_ignore_broadcasts=0

Disable ICMP routing redirects

There is something called ICMP routing redirect attacks. Google that if you want to know more about it.

to find out if your system is allowing ICMP routing redirects, issue the following command :

# sysctl -a | grep _redirects

If this produces the following result (or similar) you have your work cut out for you :

net.ipv4.conf.all.accept_redirects = 1
net.ipv4.conf.all.send_redirects = 1

# sysctl -w net.ipv4.conf.all.accept_redirects=0
# sysctl -w net.ipv4.conf.all.send_redirects=0

Disable IP source routing

Another way to hack your box is by spoofing IP adresses that you might trust as internal hosts, and thus allowing a spoofer access to services that you would normally deny them based on IP.

To find out if your box allows source routing, issue the following commands :

# sysctl -a | grep accept_source_route
# sysctl -a | grep forwarding
# sysctl -a | grep mc_forwarding

if your box is not acting like a router as well as a server, if any output on the NIC’s you are targeting is set to 1, disable them with similar commands again.

# sysctl -w net.ipv4.conf.all.accept_source_route = 0
# sysctl -w net.ipv4.conf.default.accept_source_route = 0
# sysctl -w net.ipv4.conf.eth0.accept_source_route = 0
# sysctl -w net.ipv4.conf.all.forwarding = 0
# sysctl -w net.ipv4.conf.eth0.forwarding = 0
# sysctl -w net.ipv4.conf.all.mc_forwarding = 0
# sysctl -w net.ipv4.conf.eth0.mc_forwarding = 0

Enforce packet sanity checking

If your server receives a packet where the source and/or destination field in the header don’t make much sense, it’s safe to assume you don’t want it around or do anything with it.

To check if this filter is on or off, issue the following command :

# sysctl -a | grep rp_filter

The result should be something like this :

net.ipv4.conf.all.rp_filter = 1
net.ipv4.conf.eth0.rp_filter = 1

If it is not set to 1, we will make it so :

# sysctl -w net.ipv4.conf.all.rp_filter=1
# sysctl -w net.ipv4.conf.eth0.rp_filter=1

Again add this to your /etc/rc.sysinit or /etc/rc.local if need be.

Log and drop ‘martian’ packets

Martian packets are packets that the host has no valid route back to it’s source. So it doesn’t know how to reply to them. These may be issued by attackers trying to obfuscate their location. Anyone that actually wants to use your services has no desire to obfuscate their location and prevent your box from contacting them back. So it’s safe to assume you want to drop these packets and log them for your review.

To check the state of this setting, issue the following command :

# sysctl -a | grep log_martians

net.ipv4.conf.all.log_martians = 0
net.ipv4.conf.default.log_martians = 0
net.ipv4.conf.eth0.log_martians = 0

These values ofcourse should not be 0, but 1.

# sysctl -w net.ipv4.conf.all.log_martians=1
# sysctl -w net.ipv4.conf.eth0.log_martians=1

Make your system more resistant to (SYN) flood attacks

The first and primary thing you can do to harden your system against a flood attack is simple : put in more RAM.
Each SYN packet has the purpose of establishing a connection to your server, which requires resources. Therefore having more RAM available to allocate those resources will make it more resilient against this form of attack.

Other things you can do on a linux server is enforcing the use of something called TCP SYN Cookies. This will prevent the server from really allocating resources for TCP buffers unless the server gets an ACK back on it’s own ACK/SYN reply to the client. If it does get an ACK back it means the client is legit only then will it allocate TCP buffers and use resources for it.

To check your current setting for this, issue the following command :

# sysctl -a | grep tcp_syncookies

net.ipv4.tcp_syncookies = 0

If you get this response, with the setting as 0, you will want to fix it.

# sysctl -w net.ipv4.tcp_syncookies=1

This is currently all that i can think of at this time that you need to do for hardening your TCP/IP stack on a linux server.

However, if you as a reader have more suggestions for me, feel free to drop me a line in the comment box.

The base OS of this solution will be the Ubuntu distributions, which can be found here.

First thing you need to do, is of course, set up the platform that will run the virtual servers we are going to build.

My choice for that was the 64bit desktop version of Ubuntu Lucid Lynx, which is the 10.4 LTS distribution. LTS simply means Long Term Support. Which will ensure that the coming few years your distribution will still be supported.

Why did I choose for a desktop version to run my virtual servers ? Well, basicly, linux is linux, and the type of distribution doesn’t really matter if you build a server out of it or a desktop. I chose for the desktop version because it comes with a fully pre-configured X/Gnome installation, and doesn’t contain anything I don’t need on the base platform. The only thing i needed to do was install OpenSSH to ensure that I have remote access to the machine.

However if you want to go hardcore and install a non-gui distribution like Ubuntu Server on it, that’s a perfectly valid choice as well. Although administrating the thing might be slightly harder to do, especially when it comes to installing the virtual servers. So i highly recommend the desktop version to get it all running, and then later you can just switch runlevel to 3 instead of 5 to get rid of the GUI overhead. Although I have to add that the GUI doesn’t give you much overhead apart from like 100MB extra memory load to run X/Gnome and like a half % cpu load in the worst case for X/Gnome just sitting idle.

For the virtual servers I chose Ubuntu Lucid Lynx 10.4 LTS Server. This can be either 32bit or 64bit, depending on your hardware and what VirtualBox will support on your platform. I will get back to VirtualBox later.

So now that we sorted out what OS’es we are going to use, let’s look at what we need to get the server base platform running and ready to accommodate virtual servers :

Next to an Operating System you will of course need some software that is capable of running a virtual machine inside your host machine. There are many virtualisation platforms to choose from, but I won’t cover them all, as some are really far from being viable to run a production server on, and others are too complex and beyond the scope of the audience I intend to have here, which are people looking for help on getting something done and broadening their understanding and knowledge.

That said, there are basicly 2 products to choose from, which is VMWare and Sun’s VirtualBox.

I chose VirtualBox for a very simple reason : I like it more than VMWare.

Next to my personal preference, I have a lot of experience with VMWare and VirtualBox, and although VMWare is a good product, i find it’s performance at times wanting in comparison to VirtualBox. Next to that, with VMWare you are required to register and jump through hoops before even having the hope of obtaining the free version of VMWare being VMWare Player, in contrast to VirtualBox which can be downloaded directly, is open source AND doesn’t require registration and questions you really don’t want to answer because of the risk of having some marketing dude harassing you during work hours about their great product.

That said, VirtualBox is the platform of my choice, because it’s simple to set up, fast and above all, flexible in that it can run your virtual machines in ‘Headless’ mode, in contrast (for as far as I know) to VMWare Player. ‘Headless’ in this case means you will not have a window constantly on your desktop showing you the screen output of your virtual machine, which in the case of a linux server really isn’t that interesting to look at. Instead it will run as a background service. Also VirtualBox’ Headless option makes VirtualBox viable for non-GUI installations on the host ! Definatly worth looking at and keeping in mind as an option. With VMWare Player you are pretty much stuck with a GUI on your host platform. (For as far as I know)

Headless mode on VirtualBox will still provide you with a means of looking at the screen output of that virtual machine. VirtualBox will provide you with a RDP style service that you can connect to, and will also allow you to configure the ports that it uses. (hence you can put the RDP server of server 1 on say, 4001, and server 2 on 4002.) And the good thing about the desktop version of Ubuntu 10.4 LTS is that it comes with an RDP client pre-installed. Joy !

So the first thing you will do is install the OS of your choice onto the server, on which you decide for yourself which services you want to keep running etc. I would go for as bare as possible on running daemons and possibly limit to OpenSSH and X/Gnome only depending on your desires or goals to limit the resources the host platform will use. the GUI is only there to make your life a bit easier for the rare moments you need to log on and do something for administration.

I won’t go into detail on how to install Ubuntu and such things, as there are so much great guides on that subject out there already and it really is not in the scope of these articles. But from this point on, I assume you have a bare desktop version of Ubuntu installed on your server, and are ready to give it a spin.

First of all, I highly recommend tightening security before doing anything else like firewall etc. This is also beyond the scope, but in the future i will create some articles about iptables. Once that is done, browse to the VirtualBox website and download the package that suits your distribution.

On Ubuntu, installing the package is easy : Just doubleclick it and install it and after providing your password and a quick sigh, it is installed and ready to use.

So now that we have our host up and running, tightened it down and got rid everything you really don’t need, it’s time to start thinking about what we really need to build such a wonderful high availability server with fail-over solutions and how we are going to achieve that goal.

That will be done in the next part of this series.

The first thing i need to share my thoughts about is actually the crash of one of my servers.

The machine was running a native installation of RedHat, but it was a bit old already and it desperately needed replacement. So, up to a point, the raid controller decided to go bonkers and crash my server several times until i figured out that it was actually the controller at fault instead of the disks. And that is when you are actually facing a complete reinstall of your system.

Now to make lemonade from the lemons i got, I decided to go fancy and actually make it a virtual server implementation.

So how did I go about it ? Well, I am not going to provide a complete cookbook here, but I will tell you what i am running, and how I came to that decision. I would love to hear some of your thoughts on this as well as I virtually (pun intended) have no system administrators in my direct vicinity to reflect upon.

So let’s get started : I have an AMD Opteron 165 based server. Not the newest fancy high end technology, but it is a rocksolid platform that is still very competitive compared to todays hardware and servers. That said, this proves that even with some older hardware, you can build a perfect virtual server platform that performs well and possibly even run another virtual server next to it as a fail-over solution, all on the same box !

Some of you might think that doing this will actually make no sense at all to have a fail-over solution on the same box, but I have to disagree. Short from the server physically burning down to the ground, fail-over solutions on the same physical server do make sense if you run virtual servers. One virtual server can still for some reason get damaged by a rogue program or a mistake on your end, or a bug somewhere in a program. Or even having to take down the main virtual server for maintenance like upgrading the OS counts as a possible reason. If stuff like that happens, you will still have that other virtual server taking over in a heartbeat. (pun intended)

So in short, apart from the server’s physical state degrading, there are several situations where a virtual server can still die and needs to be taken over, maintaining high availability.

Whatever kind of hardware you choose, if you go for a production platform, make sure you get at least a RAID1 array in it, at least 2 processorts, and all the RAM you can get your hands on. And make sure it’s a 64bit platform otherise you have no way of adressing all that RAM. For a medium load server I would recommend having 8GB RAM, but 4 would do also if you are into tuning instead of running out of the box.

This blog series will go into detail on how to create this quickly and easily with high performance using available free and opensource software solutions. If you want to read only this entire thread, please click on the Virtual Servers category.

I hope you will all enjoy the read and be able to learn something here that you can use for your job or even at home.

Now that we have built a PC from our components and closed it so that we can power it up, we are ready to install an OS on it. Well… Almost ready. Since we do not have a CDROM installed into the NAS, we need to figure out a way how to get the OS onto our PC so that we can create a NAS out of it. Of course you need a temporary monitor, keyboard and mouse for it to get Linux installed so hook those up too. If you do have a CDROM in it you can of course skip the whole USB key process.

The way I did it was take a USB memory stick of at least 1GB, and turn that into a bootable drive containing a linux server edition. As I stated in the previous article, I chose for Ubuntu 8.10 Intrepid Ibex Server Edition. Some may argue that for stability i better choose the 8.04-2 LTS Hardy Heron Server Edition and in my defense I can say that I actually tried that, but it would not install from the stick, so I had to choose for 8.10 since that one worked flawlessly. The reason why 8.04 wouldn’t install will be explained later on.

First we need to build ourselves a bootable USB stick with Ubuntu 8.10 installation media on it. The way I did it was just use the Create Bootable USB Stick on my laptop with Ubuntu 8.10 on it and using the Server Edition ISO. It all went by itself. However, i will assume you will not have linux or not the 8.10 version of Ubuntu, and thus do not have access to this neato tool. We will tackle that with unetbootin !

First you need to download the ISO containing Ubuntu Linux 8.10 Intrepid Ibex Server Edition ISO, or possibly a later version if you can get it. You can do that here. Store the .ISO on a location on your hard drive, and then you need to download another tool so that you can create a nice bootable USB stick from it. **Warning ** : put the contents of your USB stick onto your hard drive as the drive will be formatted and you will lose your data.

Now we need another utility that will enable you to create the bootable USB stick. That program is called ‘unetbootin’, is opensource, free of charge and even comes with a handy manual on the same page where you can download it. Check it out here !

Now I will assume you have built your bootable Ubuntu Linux Server Edition version 8.10 Intrepid Ibex memory stick. So with your loaded stick in hand, walk over to your NAS-to-be and plunge the memory stick into one of the USB ports, power on the PC for the very first time (or the second time if you tested it already) and go straight to the BIOS by hitting the DEL key when it provides that option. (or any other key that takes you to your BIOS setup, depending on what PC you decided to spend on the NAS)

Once in the BIOS, the first thing you need to do is make sure the system is able to run in ‘headless’ mode. This simply means that it needs to be set so that it can boot without a monitor, mouse and keyboard plugged into it. This typically is the setting that says Halt on [and here you can choose on which errors it should halt like keyboard error, mouse error, whatever more it can detect], which you need to disable entirely. On the Foxconn however it was the other way around, i had to tell it that it could run headless by turning on the feature Ignore Errors and then enabling the keyboard and mouse. So depending on the system you use, this can differ quite a bit.

The next thing you should do is make sure the system boots from USB disk first by setting the order by which it should boot, and put the USB disk on top. In my case however with the Foxconn hardware, i had to go to Integrated Peripherals -> USB and had to force my USB stick to be seen as an FDD / ZIP drive, otherwise the thing just wouldn’t boot and just gave me a ‘boot error’ notification. Took me a while to figure that one out.

[Now here is your explanation why I did not go for the LTS version of Ubunutu Server. The 8.04 LTS version booted just fine, and then went to search for a CDROM and could not accept the fact that my memory stick was marked FDD by the BIOS and i got stuck in the installation routine : In order to install 8.04 you need to satisfy it's demand for a CDROM unit, and quite possibly your memory stick is seen as a CDROM unit on other hardware, but with me it gave that 'boot error' notification so i had to put it on FDD, which completes the circle because that made the installer want to see a CDROM with Ubuntu on it. -sigh- luckily 8.10 Intrepid Ibex did not give me that problem]

Now on to the actual installation :

After the standard language selection (on which you preferably choose a language that you can actually understand) you are provided with several options, of which the first (Install Ubuntu Server) is the one you need. So just hit Enter when you see this and the installation sequence will be initiated.
Because I am dutch (the .nl extension should have been a dead giveaway of this fact) and I could not take (decent) screenshots while installing Ubuntu, I decided to mimic the whole procedure in VMWare in English so that you all can actually read and understand what is said on the screenshots.
However, because space is limited on my lappie and this is just an example of how the process is done, I’m working with 40GB disks instead of the real 1TB i have in my NAS.

screen 1

Now, after you pushed enter on the Install Ubuntu Server option, you are presented with a language selection screen again allowing you to choose the language of the installation. In this case I selected English again. (So it is possible to choose English for the installation procedure and end up with a french or Italian server. nice !)

After that you can choose your location and determine your keyboard layout. Since this is different for everyone, I will not cover this in this manual, you should be able to handle this on your own though.

When this is done, you will see it detecting hardware, cdrom etc and retrieving a load of packages. After that you are asked for a host name. Pick something sensible like ‘nas’ or whatever you like to call it. You will be the one attaching your other PCs to this NAS and you can make it as hard on yourself as you want to

screen 2

After that you will see some things flashing etc and you will end up at a screen that says ‘Partition Disks’. Now this module is the one we need to set up our RAID !

Do pay attention now as we will be setting up a full RAID1 mirroring system.

The reason why I chose to go for RAID1 mirroring is that I need availability and data security. The whole purpose of the NAS is that I have a place where I can put my valuable data and that even when a hard disk dies on me, I do not end up losing any of it. However, if you want to go for RAID0, be my guest although i do believe that would be defeating it’s very purpose. Consider this well before you choose either. In any case, choose the ‘Manual’ option and go from there.

As a remark I just want to say that I am aware that a hardware RAID solution is better when it comes to strain on the system resources as Linux RAID tends to use the CPU for it’s synchronising etc. In my defense I say that for a home server, Linux RAID is excellent. The Atom CPU that this system has, is a dual core with hyperthreading, thus enabling 4 threads to be ran simultaneously at 1.6Ghz ! That ought to be enough to serve 25 simultaneous users on that system while it’s not even breaking a sweat. So why buy something that we don’t actually need ? Plus, with a hardware RAID solution you need to buy that extra disk that you have to keep at hand because hardware RAID will not accept different size disks, while Linux RAID only looks at partition size and not actual disk size and thus is the more economical solution without sacrificing anything on any level. Linux RAID is perfect for our needs in a home server that is on a very tight budget.

Ok, following are 25 or so screenshots for your referral and I will walk through the whole RAID setup step by step so it will be easy to follow and as accurate as possible.

screen 3

When you used brandnew disks, you will see a screen similar to the screen 3, showing two disks with their respective size. In my case, VMWare virtual disks.

If you have not used new disks, you probably have some partition information on them. Get rid of those so that the screen resembles it.

screen 4

When you did that, you can select the first disk by pressing the Enter key and it will ask you to create a partition table on it since that is still missing like in screen 4.

Just select YES and repeat the same process for the other disk.

screen 5

Now if you did it right, you are presented with a screen much like screen 5, displaying 2 drives and their respective free space.

screen 6

The first thing we are going to do is reserve space for the swap drive. In my case I wanted 3GB of swap space to complete my 4GB memory space desire. (3GB virtual + 1GB physical)

screen 7

You will be asked if that partition should be a Primary or a Logical partition. Choose Primary.

screen 8

Next you will be asked where the SWAP partition should be placed, at the beginning or the end of the drive. I just put it at the end of the drive since I like my boot partition on the beginning of the drive. If however for some reason you need it at the beginning of the drive, go right ahead.

screen 9

Now you will be presented with screen 9 (or similar) and you should select ‘Use As’ which by default is set to Ext3 Journaling System. We do not want this to be Ext, we want to use this partition in RAID.

screen 10

When you press enter on the ‘Use As’ option, you get a new screen (screen 10) and you have to select ‘Physical volume for RAID’. Press enter on that selection.

screen 11

You are returned to the screen 11 and you can see some things has changed. The bootable flag should remain off, which it is by default. This is our swap partition, we won’t be able to boot from it anyway.

Now when you are done, select ‘Done setting up the partition’ and you will be returned to screen 5 again, displaying your drives with respective partition table, this time with the partition we just created. Select the FREE SPACE on the same disk you just made the partition on, and create a new partition with the size equal to the remaining size on the disk. Make it a primary partition again and when you get to the screen above, also put this one on ‘Physical volume for RAID’ as shown in screen 9 and 10.

screen 12

This time however, you should put the bootable flag to ‘on’ as shown in screen 12 since we can and will be booting from this partition. When you are done, select ‘Done setting up the partition’ which will again take you back to screen 5. Repeat this process for the other disk using equal values, and you should be presented with screen 13 when you are done.

screen 13

As you can see in screen 13, we now also have the option ‘Configure software RAID’, which we should select so that we can set up both disks in a RAID1 setup.

screen 14

You will most likely be confronted with this screen, telling you about the changes that need to be written first before you can setup RAID. Just say YES and go to screen 15.

screen 15

Now you are presented with screen 15, in which we select ‘Create MD device’. If you should wonder what MD stands for, it’s simply shorthand for ‘Multiple Disk’, nothing fancy about it.

screen 16

Now you should have ended up in the selection screen on which type of RAID you want to setup.

Now here is the breakdown of our options : We do not have enough disks for RAID5, so forget about that option. RAID1 is mirrored and RAID0 is striped. For a thorough explanation on the differences between them i refer you to wikipedia.

In short I can say that RAID1 is the better choice when it comes to both (reading) performance and data integrity. If in a RAID1 setup one disk decides to trip and die, the server will happily keep running on the other disk and you will not lose a single byte of your precious data, and putting in a new disk will mean easy resynchronisation between the two and be secure again. The downside however is that with 2x 1TB drives, you end up with only 1TB usable. The upside is that reading is distributed over both drives and thus speed increase is great.

With RAID0, you concatenate both drives, ending up with the full 2TB when using 2x 1TB drives, and the data when both reading and writing is spread over both drives, enhancing read and write speeds. The downside however is that if one disk in a RAID0 configuration gets hiccups and dies on you, and you have valuable data on it, you really screwed yourself over because you will be lucky to salvage half your data, and some dude named Murphy predicts that the most important data will be on the drive that just died. If you want a NAS for safe storage, do -not- go for RAID0 just because you can use the full size of both drives. In effect you are not better off than having a USB hard drive, only USB hard drives are a lot cheaper than our NAS. Be safe ! Go for RAID1. (if after reading all of this still want RAID0, well.. *shrug* I warned you )

screen 17

When you select RAID1 you are asked how many active devices you want to put into your RAID1 array. Standard this is filled out with 2. Keep it that way.

screen 18

When you pressed enter on screen 16, you get screen 17, asking us how many spare devices we have available for the RAID1 array. (this has to do with a possible RAID0+1 or whatever setup) We do not have a spare device. So we leave it at 0 and hit Enter.

screen 19

Now we are going to select which partitions on which drives we want to combine into a RAID1 disk. Select the second partitions from both disks and hit continue. (we created SWAP first, so that is partition 1 and that is why we take partition 2 first. That way, the normal disk will be called MD0 and the swap partition MD1. It’s neater that way, but if you do not have an obsessive compulsive disorder to keep things orderly and neat on your machines like I do, any order is just fine. It’s just a name)

screen 20

Now, when you created the first RAID1 disk, you will be presented with screen 15 again, and repeat the process for the remaining 2 partitions, thus creating MD1, after which you are taken back to screen 15 again, and this time you select ‘FINISH’. This will take you all the way back to screen 21 below.

screen 21

Now we are ready to determine which partition is going to do what. As you can see in the image above, i did not reverse the partition order in that screenshot. On your screen it should be the other way around now if you followed my instructions to the letter, or the same if you didn’t give a crap in what order they show up.

Now select the partition on the RAID device that is going to hold your data and OS. That should be the big one that is not going to be your swap area.

screen 22

Change the screen as shown in screen 22, using the partition as Ext3 Journaling file system and setting the mount point to ‘/’ being the absolute root of your file system. When you are done, go to ‘Done setting up the partition’  and you will be taken back to screen 21.

From there select the swap partition on the RAID device, and be taken to screen 23.

screen 23

Change the ‘Use as’ option to ’swap area’ and the screen should look like screen 23 above. when you are done, select ‘Done setting up the partition’ and get taken to screen 24.

screen 24

Your screen should look like screen 24 above now and you can select ‘Finish partitioning and write changes to disk’.

Now some of you readers might argue that putting the swap area also in RAID is going to do nothing for your performance, and you are absolutely right. It is hardly going to have a negative impact on your performance either and if the system really is using the swap area a lot, you might want to consider putting in some extra physical RAM instead of letting it utilize a much slower medium.

The reason why I am also putting swap into RAID1 is simply availability through redundancy. If one disk is to die, chances are 50% that it will be the one containing the swap area and thus taking down the whole system with it. When it’s in RAID1, it will not matter to the system as it will happily keep running on the remaining good drive. (And besides, what else can I do with the leftover 3GB on one drive that is better than having it fully redundant)

screen 25

When you are asked to save the changes to disk like in screen 25, of course answer yes and the system will complain to you with screen 26.

screen 26

This simply means that the system is not aware that you built a RAID setup yet and that you should reboot. However, we will not reboot ! I tried this 2 times with my real NAS and with VMWare, and with VMWare it turned out that it happily used the mirror regardless of this warning and everything worked just as it should, and with the real NAS when installation was completed and the system rebooted, it noticed the disks were out of sync and it happily rebuilt the RAID setup automatically by syncing the disks, just the way it was supposed to do. So do not reboot. It’s a waste of time and you end up getting the same warning again anyway.

If someone can enlighten me why this warning is displayed at all, or why it says we have to reboot, please do, as I still don’t get it.

This warning will display for any RAID disk you created, so in this case, it will be twice.

Just hit continue and the installation will continue taking you to screen 27.

screen 27

Now the installation wants to create it’s first user. Just enter your full name here and continue.

screen 28

Then it wants you to pick a user name. By default it will display the first name you used in screen 27. Change it to your liking, or leave it as it is. It’s up to you. After that you will be asked for a password and to confirm that password again as you should have expected.

screen 29

After this it will ask you if you want to setup an encrypted private folder for you. I had no real use for this so I chose NO, but if you want it, just run though it’s setup until you get to screen 30. Again, it is up to you if you want this or not.

screen 30

Here you are asked if you have a proxy running on your network that this server should use. Fill in the proxy adress or if you don’t have one (most likely) just leave it empty and hit Enter.

screen 31

The next question you will get is if you want automatic updates or use landscape to manage your server. Landscape seems to be something from Canonical and it’s probably a great thing if you have multiple servers to run in a corporate environment, so I bet it’s useless to us common people. Choose ‘Install security updates automatically’ if you want to have your system get fed the critical security updates automatically, which in my opinion is a wise thing to do. However if you don’t want any updates to be ran and do it all by hand, that’s fine too, it’s up to you.

screen 32

Now you are able to choose which services you run. For an as generic as possible setup i suggest you mark the following :

- LAMP Server
Linux Apache MySQL PHP in one package.
- OpenSSH Server
Secure Shell access to your server. A must have, regardless of what you may think of it
- Print Server
If you intend to connect a printer to this machine and share it on the network. If not, install it anyway, saves you a headache later and it does not sit in the way or anything.
- Samba Server
This is a must have. Without it your NAS is nothing more than a fancy 300 dollar paperweight. Samba is the whole center of our NAS and is the service with which we provide our RAID1 disk space to our network.

The other options, I don’t deem them necessary but if you have a use for them, go right ahead and install them as well. Just press space to select an option and hit Enter when you are done and the installation will continue, installing the options you chose.

screen 33

If you chose to install LAMP server as well, you will be presented with a dialog asking you for the root password for MySQL. Now this isn’t essentially the same as the user ‘root’ on your system but is the user ‘root’ on the MySQL service. So fill in a password that preferably is different of what you are going to choose for the root user of the system. (if you’ll do that, as it isn’t necessary because you can sudo whatever you need)

When you entered the password it will ask for confirmation and then the installation continues.

After the whole installation is done, it will tell you to remove your media (USB stick, CD, whatever you used) so that it can reboot without starting up the installation again. When you get to this point, congratulations ! Well done. Go grab a beer or any other beverage that is your favorite because you have just laid the base of a wonderful and flexible NAS solution based on Ubuntu Server, for a mere 300 bucks.

I planned a part 3 of this series, but i never got around to it unfortunatly.

However, there is plenty resources on the internet on how to use the services we just installed.

Cheers !

Something entirely different this time, i know. Recently I felt the need to have ’secure’, easy accessible and reliable storage capacity for my data. In the past I have used mobile drives, 2,5Inch laptop drives in a case with a USB interface. While that worked just fine in the ’secure’ and ‘easy accessible’ area, it is of course not really reliable. In practice you tend to lug those things around with you, bumping them and occasionally accidentally dropping them.

A week ago i connected my 500GB portable drive containing my business data to my laptop at home, and went to get a cuppa, hooking my finger behind the USB cable and pulling the thing off the table, resulting into a nasty drop where my only luck was that the USB cable was quite short and did not reach the ground. As you can imagine, the drive started it’s gravity induced journey and right before it hit the ground, there was no more USB cable left and the drive went into a nasty spin slightly upwards while disconnecting from the USB cable, to finally land flat on it’s vibration absorbing pads on the bottom.

This time, I really lucked out. The drive and the data did not sustain any damage from the sudden disconnect and drop to the floor. (Actually it isn’t the falling that is a problem, it’s the sudden stop)

That is when i decided that I had to have a NAS. For those among you that need a reminder on what NAS meant and those that have no clue what a NAS is : NAS stands for Network Attached Storage.

Having looked around, I found out that ready-made NAS boxes are immensely expensive (600+) and lack flexibility. (ie you buy a NAS and that is all you are going to get) So I decided to build my own.

Now that the introduction is over, many people use the term NAS and SAN like they are the same thing, and as the naming suggests, they actually are not. SAN stands for Storage Area Network while NAS stands for Network Attached Storage. As the naming suggests, a NAS is actually a device that you couple onto your network and can access as a server offering disk shares with a high level protocol (NFS and/or CIFS), while a SAN is actually a cluster of storage devices bound into a high-speed network (in the ye olde days with a SCSI solution, these days fibre optics are being used) offering it’s storage capacity at a low level protocol (iSCSI) to servers that in their turn can be accessed by clients. Typically a NAS is suitable and convenient for homes and small businesses, while a SAN is typically useful for enterprises. (enterprises as in 500+ employee businesses, banks, hospitals etc) You really do not need a fibre optic storage network in your home

Now I can get all technical about a SAN and it’s true inner workings and compare that to a NAS, but that would defeat the purpose. I hope I did not lose you halfway already. Let’s just say that you can build a NAS for close to 300 bucks, and you have to fork over at least 200 Bennies (20k) for a decent SAN solution. Now THAT is a difference.

So, what do we need for our own home-brew NAS ?
- 1 Personal Computer
Now this PC will be doing absolutely nothing most of the time so consider your choices. It should provide you with remote storage, not support NASA by guiding Expedition 20 to Mars and back. Buying something really speedy is just going to give you a high electricity bill, while the files you push and pull over your 100mb (or 1Gb) just will not go any faster than what your network allows. That is really all it has to do. Besides, we want to stay close to that 300 bucks, not 3000 bucks.
- 2 Hard-disks
Now for the hard-disks you can settle for something not super-modern and speedy as well. Like I said, all data is going through our home network and we do not have 500 simultaneous users so we can settle for 5400rpm disks. They are less noisy and do not generate tons of heat and are more energy efficient while providing enough speed for our needs.
- 1 stick of memory
512MB ought to be enough, but since you can buy 1G for only one or two bucks more, i suggest you get on the safe side and go for 1G.
- 1 OS
For the OS, we of course look at the OpenSource sector, and more specifically linux type. Now I know there is loads of out-of-the-box solutions here to provide for all your NAS needs, like freeNAS and such, and of course I tried most of the non-commercial solutions and they all provide at least the basics of what I wanted and are great solutions in their own right, but i craved more flexibility, more dynamic setup options and more up-to-date’ness. With the NAS on a stick solutions out there you are kind of depending on their (handful if not only 1) creator(s) to provide new releases and such, while with a standard linux server distribution you have a large community almost guaranteeing security updates and more flexible ways of even updating the entire OS.

And, that is about all we need.

“But Uptown, yo man, what about CDROM, keyboard, mouse, monitor ? don’t we need more ?”

In fact, yes we do. We need a screen obviously for our setup, network cards etc, but most of these things you already have and since a NAS is supposed to run ‘headless’, it should be able to boot without anything connected to it except your router (and of course your power cord. Yeah, i better take the wind out of your sails before i get some smartasses here pointing that out). So in fact, yes that is all we need for the NAS to run standalone. For the installation however, it is a whole nother story, but I will get back to that later.

First things first : Buying a PC.
Yes I am aware that you might have an old PC lying around that you can use for this project, but it’s probably not what we need. The choice of PC for building your home server is based on a whole set of other criteria than the choice of PC for let’s say, your average company with a server room. This one probably has to sit in your living room, close to your router as it will be UTP connected. (yes, wireless NAS is a bad idea, too much variables making it potentially unreliable) So first of all, the thing has to be visually pleasing, secondly it has to be silent, thirdly it has to be energy efficient and lastly, you don’t want it to generate tons of heat. That’s nice in winter, but in summertime you are going to curse it. Obviously none of these things matter in a server room, but in your living room, they do. You do not want to justify having a noisy ugly NAS to your spouse by telling them it’s the new micro climate control

Another criteria is that it has to be really cheap and has as much of the standard components like video, network etc. already built in so you do not have to worry about that. Our solution to all these criteria are called Barebone PC’s. These are mostly cheap visually pleasing small energy efficient complete PC’s that in most cases only require drives, memory and a CPU to be built in as custom components to make it a complete working system. Plus, they don’t come with an overpriced license for a useless OS that we do not want. (yes, I am talking about Windows)

If you take a good look on webshops that sell barebones, you will see that they come in all sizes and shapes and different architectures. What we are aiming for is energy efficiency while providing moderate speed when needed. The platform I can recommend is either Intel’s Atom or AMD’s rival (if there is such a thing, i don’t know really). The Atom ensures low power consumption and thus low heat generation and there are already a few Barebone manufacturers that utilize this platform.

My choice went to the Foxconn L10-S3, which is based on the Intel Atom Dualcore platform. It’s dimensions are 283 x 95 x 282mm, which satisfies my size requirements, has a shiny silver/black finish on the front end which satisfies my visual requirements, is silent because it has low rpm fans and the Atom is even passively cooled, which satisfies my noise requirements, plus that it is energy efficient. The Atom that is already built into this machine has a tdp of 8 watts max and the power-supply can deliver a max of 170 Watts. More than enough to support our energy efficient hard-disks.

As I said, this thing already comes with a built-in Atom CPU so the only thing that it’s lacking is a strip of memory and one or two hard-drives. The Barebone costs 129 Euro’s when i got it a week ago, so that is actually a whole lot of PC for your buck. (And that satisfies my requirement that it has to be cheap)

The system sports a hard-drive bay, floppy bay and a 5.25 Inch fullsize CD bay. But obviously, I did not use the CD bay for a CDROM, but instead I bought extenders so that I could place a second 3.5Inch hard-disk into that place.

As for memory, I just bought a stick of Kingston 1G DDR-2 533Mhz RAM, as that is the speed that the mainboard requires. (It’s based on Intel’s 945 chipset) And as you can see, those things come at 10 or so bucks a piece. You can even go cheaper than that, but I would recommend spending a buck or 2 extra for reliability. But if you think you can do better for less, knock yourself out

Now we can turn to the hard-drives, which will actually be the expensive parts of the PC for a change.

Since the Foxconn has the Intel ICH7 SATA controller on board, you are able to buy standard SATA drives for the system. Again, we want energy efficient, silent drives that don’t heat up too much.

My choice went to the Western Digital Caviar Green series. The 1TB version to be exact, of which I bought 2. (I wanted RAID1 so i needed 2) These drives come at a price of around 8 cents per MB, so they will set you back 79 bucks a piece at the time of this writing. 158 bucks in total, which makes them more expensive than the rest of the system all together including the memory bank we had to slap in !

Ofcourse you can put in smaller drives for less money if you want, this is just a choice that i made. You have to adapt your choices to your needs.

So this is the result of my labor. As you can clearly see, the CPU is being passively cooled, and the 2nd hard-disk occupies the CDROM bay. Behind the cpu you can see a glimpse of the 1GB memory bank and the power-supply is on the right. The 1st hard-disk is sitting under the CDROM bay where it is supposed to installed. We can not see it on this picture.

Optionally there is room for one PCI based upgrade, should you decide to make a media center of this machine. The system should be able to handle that neatly!

So what do we have at the moment ?

One Foxconn L10-S3 for 129,-
One 1GB memory bank for 10,-
Two WD 1TB hard-disks for 158,-

This totals EUR 297,- for a complete working PC system. (Swap out one hard-disk, put in a CDROM, buy a keyboard and mouse and you have a nice internet PC for the same 300 bucks!)

And yes I know the title says 300 dollars while i currently talk about Euros. I guess I will justify that by saying that when it comes to computers, 1 dollar in the US equals 1 euro in Europe. Maybe you will have to fork over 10 dollars more than i did with euros, but you will stick around the 300 usd with buying these quality components.

This concludes the first of the series I will post about this system. The next post will be about installing and configuring the OS and making a true NAS out of this system.

As for benchmarking goes : here is some output from something called tiobench/tiotest on this very system.

No size specified, using 1792 MB

Unit information
================
File size = megabytes
Blk Size  = bytes
Rate      = megabytes per second
CPU%      = percentage of CPU used during the test
Latency   = milliseconds
Lat%      = percent of requests that took longer than X seconds
CPU Eff   = Rate divided by CPU% - throughput per cpu load

Sequential Reads
                              File  Blk   Num                   Avg      Maximum      Lat%     Lat%    CPU
Identifier                    Size  Size  Thr   Rate  (CPU%)  Latency    Latency      >2s      >10s    Eff
---------------------------- ------ ----- ---  ------ ------ --------- -----------  -------- -------- -----
2.6.27-11-server              1792  4096    1   65.66 22.60%     0.058      909.64   0.00000  0.00000   290
2.6.27-11-server              1792  4096    2  105.68 79.08%     0.072      114.69   0.00000  0.00000   134
2.6.27-11-server              1792  4096    4   54.61 85.23%     0.255      811.17   0.00000  0.00000    64
2.6.27-11-server              1792  4096    8   47.04 130.0%     0.568      880.40   0.00000  0.00000    36

Random Reads
                              File  Blk   Num                   Avg      Maximum      Lat%     Lat%    CPU
Identifier                    Size  Size  Thr   Rate  (CPU%)  Latency    Latency      >2s      >10s    Eff
---------------------------- ------ ----- ---  ------ ------ --------- -----------  -------- -------- -----
2.6.27-11-server              1792  4096    1    0.75 0.958%     5.212       32.05   0.00000  0.00000    78
2.6.27-11-server              1792  4096    2    0.83 1.535%     9.399      159.65   0.00000  0.00000    54
2.6.27-11-server              1792  4096    4    0.78 2.134%    17.324      133.51   0.00000  0.00000    36
2.6.27-11-server              1792  4096    8    0.87 4.233%    29.435      344.79   0.00000  0.00000    21

Sequential Writes
                              File  Blk   Num                   Avg      Maximum      Lat%     Lat%    CPU
Identifier                    Size  Size  Thr   Rate  (CPU%)  Latency    Latency      >2s      >10s    Eff
---------------------------- ------ ----- ---  ------ ------ --------- -----------  -------- -------- -----
2.6.27-11-server              1792  4096    1   70.25 76.87%     0.051      399.90   0.00000  0.00000    91
2.6.27-11-server              1792  4096    2   67.18 178.2%     0.107      927.14   0.00000  0.00000    38
2.6.27-11-server              1792  4096    4   68.73 446.8%     0.206     1065.27   0.00000  0.00000    15
2.6.27-11-server              1792  4096    8   70.03 1017.%     0.379     4954.40   0.00022  0.00000     7

Random Writes
                              File  Blk   Num                   Avg      Maximum      Lat%     Lat%    CPU
Identifier                    Size  Size  Thr   Rate  (CPU%)  Latency    Latency      >2s      >10s    Eff
---------------------------- ------ ----- ---  ------ ------ --------- -----------  -------- -------- -----
2.6.27-11-server              1792  4096    1    1.42 1.002%     0.016        0.07   0.00000  0.00000   142
2.6.27-11-server              1792  4096    2    1.32 2.370%     0.019        0.12   0.00000  0.00000    56
2.6.27-11-server              1792  4096    4    1.34 5.246%     0.024        0.12   0.00000  0.00000    26
2.6.27-11-server              1792  4096    8    1.24 8.098%     0.036       21.10   0.00000  0.00000    15

Cheers !

First thing you need to do, is make sure everything works with the sound system. That means, speakers are configured right, but also confirming that the (built-in if you use a laptop) microphone works as well.

To get this done, you might have to take a look at this page, but obviously i will not cover getting your sound cards to work on every system possible. This entry is about getting Ventrilo to work, and hopefully a comprehensive guide at that. This also means that this guide assumes that you have Wine installed and properly working ! If you haven’t done that yet, get it done first.

So, now that you have your sound card setup and working properly, it’s time to start working on Ventrilo. If you have not done so, download the latest version of Ventrilo for Windows. (you can click the image below to get to the download page)

Alright, now that we have the setup file sitting on the hard disk, it’s time to fire it up and install it. Again, proper installation procedure is not covered by this guide. You should be able to click ‘Next’ until you see the ‘Finish’ button

Click to view fullsize

Now that we have Ventrilo properly installed under Wine, we have to make a few configuration changes. First thing you need to do is make sure that Wine uses ALSA as sound system. Some of you might have also checked OSS as second system to initialize, but on my system that made the sound choppy under some circumstances, so it might work for you or it might not to also have OSS on.

Now on the bottom you can see that i set Hardware Acceleration to Emulation and that i also checked the Driver Emulation check-box. You should do this. This gave me the best results anyway.

As you might have observed, this screen uses a language other than english. It’s in dutch. Doesn’t matter much, I just tell you this in case you wonder what the gibberish is about.

Now, anyone can get this far without too much trouble i guess, but after that is where the trouble started for me. Whenever i started up Ventrilo, it complained it could not find a codec called MSGSM 6.10. After some searching this appeared to be a proprietary codec. If you have a Windows installation, you can grab it from the Windows\System32 folder and put it in the Windows/System32 folder of your Wine installation, or alternatively, you can grab it from Filefront.

After you have done that, you have to edit a file called system.ini, which is in the Windows folder. (if you have installed standard Wine, this should be in ~/.wine/drive_c/windows, or if you use Crossover Games, it should be in ~/.cxgames/[yourbottle]/drive_c/windows). it should look like something like this :

[mci]
MPEGVideo=mciqtz.drv
MPEGVideo2=mciqtz.drv
avivideo=mciavi32.dll
cdaudio=mcicda.dll
sequencer=mciseq.dll
vcr=mcivisca.drv
waveaudio=mciwave.dll

[drivers32]
msacm.imaadpcm=imaadp32.acm
msacm.msadpcm=msadp32.acm
msacm.msg711=msg711.acm
msacm.winemp3=winemp3.acm
vidc.mrle=msrle32.dll
vidc.msvc=msvidc32.dll
vidc.cvid=iccvid.dll

Now the trick here is to let Wine and thus Ventrilo know where it can find that new msgsm.acm file we just downloaded, and we do that by adding a line to the [drivers32] section :

MSACM.msgsm610=msgsm32.acm

Now, for me this worked just fine, but i also read reports of people that claimed Ventrilo still was not able to find the msgsm 6.10 codec. This can be solved by adding the full path to where to find the codec :

MSACM.msgsm610=C:\Windows\System32\msgsm32.acm

After this, it’s time to make sure Ventrilo is able to use your microphone. You being able to hear yourself through the mic when you un-mute the Microphone channel is no guarantee that Ventrilo can use your mic too. I found that out the hard way. So to be sure, and if you use Gnome (KDE uses a similar process) you need to open up the sound configuration panel (double-click the speaker icon in the system tray top left of your screen).

When you have it open, go to the Settings screen and check the boxes for Recording and any variation on that. That should get you an extra Tab in the configuration panel called Recording or something similar if you didn’t have it already.

There you should enable the microphone and mute the channel. When you did this, the sound that the microphone records is also routed through Wine and Ventrilo. (this was my biggest issue, it took me a good week to figure this one out. I’m such a n00b!)

Anyway, after you did this, fire up Ventrilo and go to the settings screen. Once there, you check all the boxes next to ‘Use DirectSound’ on the tabs ‘Voice’, ‘Binds’ and ‘Speech’.

Furthermore just leave the hardware mixer section empty as it’s optional and we do not want to get Wine confused over anything. After that, just put the coded on GSM 6.10 and set the frequency to something suitable. (higher means better quality, but also more cpu intensive).

Alright, now you should be able to use Ventrilo like any windows user can

If you have comments or questions, please do let me know in the comments.

Happy Gaming !

VMWare has chosen a new path however, and one that is much much easier too. It’s VMWare’s newest bundle of joy : The bundle installation file. (sorry about the pun, really)

VMware did good i think, but looking at my previous post on this issue and seeing how I had to deal with things then and now, there is a world of difference between them. VMWare has completely made the vmware-config.pl procedure obsolete and has automated that process fully ! Go VMWare !

So if you are running Ubuntu, you really do not want to download the RPM or TGZ, just go straight for the .BUNDLE package and follow my guide below to get it working.

First of all, download the file at the vmware website and pick either the Workstation version or the Player version. Either is fine with me since they install practically the same. For this tutorial however i will be using the workstation version as that is what i have available to me. Please note that the .bundle extension is not registered to a package manager or anything. I am not sure exactly what it is either but it seems to be very similar to a Windows Installer format for Linux. (Yay, just what we need, -another- ’standard’ on linux. But in this case i’ll let it fly as it seems to work really good)

Now, we have to execute this file as root and we can not do that easily from the GUI. (or at least I haven’t found out how to do that easily) This effectively means that we better place the downloaded file (either player or workstation) at our home folder’s root for our own convenience and then fire it up with the shell. So place that file you just downloaded into your home folder and open the shell after which you punch in the following :

$ sudo ./VMware-Workstation-6.5.1-126130.i386.bundle

click to view full size

This should give you a nice installation wizard which you should follow to install, but if you toyed around with any version of VMWare before, it will first get rid of any old grit and stuff that still resides on your drive that can interfere with proper functioning of VMWare.

So basically if you royally screwed up for any reason at all (for example by following my now very obsolete first guide) you don’t have to worry about any debris getting in the way as the new VMWare installer takes care of all that for you.

Now this is something i desperately would have wanted a few years ago when VMWare went bonkers and i had to tinker my way through it to figure out what was wrong with it.

click to view full size

But, VMWare has come a long way since and they really improved this.

So ok, after the optional uninstall of old crap lying around, VMWare will show you an install wizard, ask you where eclipse is if you have that and that was it ! It will completely install and configure itself, including autobuild modules for your kernel etc etc etc. So there is no more need to run any additional scrips or modules. Instead, it will just be sitting there, ready for you to fire it up in the System Tools menu.

You will have to agree with me, this is an installation procedure, how an installation procedure should be, except for that shell thing getting in the way, but hey, this is linux and the shell is a very important tool, if not the most important tool in our arsenal of running linux and that means we will be seeing a lot more of it so that is not really a notable disadvantage.

So there we have it : A running VMWare and no headache !

Well wouldn’t that be nice now if that were true. As fate would have it, I recently upgraded to 8.10 Intrepid Ibex and had not done anything serious with VMWare since this last upgrade to it.

So, I went back to my jolly programming, quickly finding out that many keys utterly mandatory in almost any programming language, being the ‘, “, +, =, -, pgup, pgdown, home, end etc etc simply were not working or doing something entirely different than what they were supposed to do in my VMWare guest OS.

So I set up a search and i actually found several solutions, of which the most comprehensive is this one here which is called ‘VMWare and the FUBAR keyboard‘. If you are unlucky enough to be in the same spot as what that page describes, then you might find a real solution in there. But as it went for me, it partially solved the problem, but it got to the point of being silly and i felt like i was working around the symptoms and not actually fixing the problem. It seemed to have something to do with a bug in some handling routine of the keys and thus confusing VMWare.

Now I was not all that satisfied with Intrepid Ibex and wanted to go back to the stability and speed of Hardy Heron (8.04) again. This meant that I actually had to downgrade, which I gladly did at this point as i really had some problems to take care of with the help of VMWare.

So after switching back to 8.04, and reinstalling VMWare, the problem simply went away, so my suspicians that VMWare was at fault was completely erronous. The problem is with the Intrepid Ibex version of Ubuntu. So there you have it : If you ran into this problem and could not fix it with the help of that website i just mentioned 2 paragraphs back, you really should consider going back to 8.04 Hardy Heron.

Good luck virtualizing !