Cleaning Up a Virtualbox XP Install

(Or frankly just cleaning up an XP install. But then XP natively will have a big chunky disc to spread all of its insanity into, whereas a Virtual XP install will probably be quite restricted.)

I have started using VirtualBox to run XP on my Ubuntu workstation just in case I have any compatibility problems with native Ubuntu programs. Like, for instance, the Exchange Server connection in Evolution. This part seems to have been programmed with all the precision and attention to detail of a Jackson Pollock painting. Want to move a large email to a public folder? "The connection to the Exchange Server has been lost". OK maybe that was a tad ambitious. How about just opening a sodding email? "No chance mate, you're getting the raw HTML of the Outlook Web Access login screen instead".

So an XP guest OS with Outlook installed seems like a good idea to cover the rare occasions when Evolution acts like a total dick. Like Tuesdays. Or Thursdays.

I am no expert when it comes to the arcana of Microsoft licensing, but my case has an official sticker on it with a product key, so I think I am safe running XP on it using that key, albeit as a Virtual Machine. Same goes for the Office suite that came with the machine as well. It isn't being used anywhere else since I wiped the disk and stuck Ubuntu on a couple of years ago.

Or Wednesdays, come to think of it.

Well, happily the installation and activation of both XP and Office went swimmingly. My installation CD was only SP1. I installed SP3 over the top of it, but now I have a large collection of crap in the Windows folder to be used in the unlikely event that I ever want to NOT have SP3 installed. Space in general is a bit of an issue. I have a 40Gb SSD drive in the machine, which is more than fine for Ubuntu, but is slightly pushing it for Ubuntu + XP, even as a VM. I gave the virtual machine a 10Gb HDD, which I think is about the bare minimum for an XP install. After installation of Office and a bit of Proprietary Management Software, it is down to less than 3Gb free. Which is not ideal.

So lets fire up Tree Size Free and have a look at where all the space is going.

Biggest Culprit is the Windows folder with 3.5Gb
Next up is the root folder of C: with 1.5Gb
Next is the User folder with about 1Gb in it.
Lastly there is about half a Gb in Program Files.

Yes that adds up to about 6.5Gb. Where has the remaining half gigabyte gone? Two words. System Restore. Great idea in general, no use on a space limited task dedicated machine if you do regular back ups. To remove this, Start -> Right Click on -> My Computer -> Left Click on -> Properties -> System Restore Tab -> Turn Off System Restore Check Box. It is not that hard is it?

OK, first proper clean up - I had left the installation files for the proprietary software lying on the desktop. Delete = 300Mb back. The rest of the User folder was an archive of emails amounting to a couple of hundred Mb which I uploaded to the server, and the local copy of my exchange emails. Can't do anything about them, or indeed the Program Files. Need those to actual run the applications. Still, we've cleared half a Gb. Lets look at the root of C: next.

It has one big file in it: pagefile.sys. It is a whole 1.5Gb itself. This is the famous windows pagefile which the system uses if windows runs out of RAM. Lets run a basic test by opening all the applications installed. Word, Excel, Outlook and the Proprietary software. Done that, and the pagefile usage barely his 250Mb. So, let's trim back the 1.5Gb pagefile to a more sensible 512Mb to give us some headroom.

To do this we go to Start -> Right Click on -> My Computer -> Left Click on -> Properties -> Advanced Tab -> Performance Settings Button -> Advanced (Again) -> Virtual Memory Change Button. I chose a custom size, starting at 512Mb and limited to 1024Mb. Much more sensible for a machine with 1Gb of ram to start with.

Lastly we have the Windows folder. Most of this we DO NOT WANT TO FUCK WITH. However, there are a number of temporary files that we can get rid of. The reason is that XP keeps the files that were replaced by each update or Service Pack install just in case you want to undo the update. As long as it works, I don't. And, as with the System Restore rational, I will be keeping a backup of the install so I don't need the install to back itself up. This is dead easy with virtual disks - you just make a copy of the disk image somewhere. For instance with this command:

sudo apt-get install p7zip-full
7za a XP_10Gb.vdi.lzma XP_10Gb.vdi

So, what can we get rid of in Windows? There are four main folders that I looked at:

C:\Windows\$NtUninstall[blahblah] <- lots of these
C:\Windows\Driver Cache\I386
C:\Windows\ServicePackFiles\I386
C:\Windows\SoftwareDistribution\Download

The first type are just lots of hidden (Explorer Window -> Tools Menu -> Folder Options -> View Tab -> Show Hidden Files and Folders Radio Button & Uncheck Hide Protected Operating System Files) folders storing uninstall data for most windows updates. It is Completely messy of XP to just dump them in the Windows folder.

The next three folders all store downloaded files that have been installed. You MAY need these again if you are installing any new software. What I did was copy all of the foregoing directories into a backup folder, and then stuck that on my NAS [Network Attached Storage]. Bear in mind that you need to install any outstanding updates before moving these files otherwise the update will break and need to be downloaded again.

Also in Windows is a folder called Installer. Do not mess with this. But if you do want to mess with this, google for msicuu2.exe and use that rather than doing it manually.

Anyway, with that clear out, the Windows folder was reduced from 3.5Gb to 1.7Gb, which is ideal. I now have just under 7Gb free, representing a space recovery of 4Gb on a 10Gb drive. Good Stuff.

There is also a temp folder has appeared in the root of C: called MSOCache. This is apparently a copy of the Office install files - presumably in case I want to install any optional items, or parts I chose not to install to begin with. It is only 100Mb or so, and it also appears in the Disk Cleanup program (Start -> My Computer -> Right Click -> C: -> Properties -> Disk Cleanup Button) so I can clean this up easily at a later date if I want.

Why oh fucking why are the stupid Uninstall, Driver Cache, Service Pack, and Downloaded files not also in the Disk Cleanup program? As Advanced options perhaps? As it stands I will have to go through this lengthy process every so often to clean out these unnecessary files. Pain.