deKay's Blog

We’re currently in the process of migrating servers from Server 2003 to Server 2008, which is proving a right old pain for various reasons. This morning’s headache was reinstalling a load of network printers on the new 2008 install that worked perfectly well on the 2003 install.

64-bit 2008, by default, only installs printers with the 64-bit 2008 and (for some printers, it seems, but not others) the 32 and 64-bit Vista/Windows 7 “User Mode” drivers. The old-style 2000/XP/2003 drivers don’t get a look in. There doesn’t seem to be a way to add them on the server either.

Thankfully, someone else has already found a way of sorting it. Basically, you navigate on an XP client to the Printers and Faxes share of the 2008 server, then click File > Server Properties > Drivers > Add, and point it at the XP (and other, if necessary) drivers for the printer.

There are a few issues, particularly if the 2008 and XP drivers have different names, but these can be resolved or worked around.

Here’s the full set of instructions – read the comments there for how to fix the additional problems!

If, like us, you have WebMarshal, and it’s tied to an SQL Server Express database, you may find it sometimes just… breaks. You’ll get eventlog messages like “3003:  Failed to write session to database” and “PRIMARY filegroup is full” and “Could not allocate space for object ‘dbo.DomainFileLog’”. Chances are, your database is full.

The Express version of SQL Server 2005 is nice because it’s free, but it does limit database sizes to just 4GB. This may sound huge, but for us, that’s only around 4 months worth of internet logs, and we need to clear some space every so often. I always forget how to do it, so this post is a reminder to me as much as anything.

Before you start, to make things easier, get yourself a copy of Microsoft SQL Server Management Studio Express. It’s free, and makes managing your SQL databases much easier. Once installed on your SQL Server machine (and connected to it), you can set about fixing WebMarshal.

Firstly, you need to purge a load of old records from the log. Chances are you don’t need to go back more than a few months anyway. If you do, you might want to backup your database, name it as a “snapshot” of today’s date, and then restore it if you need to have a look at old entries. How you purge depends on the version of WebMarshal you have, but for either 3.x or 6.x you start my clicking open the WebMarshal database from the object explorer on the left.

Now, on the pane on the right, you’ll have an empty command window. In here, you need to paste one of the following commands:

For WebMarshal 3.x:

exec dbo.PurgeLogData @NumDays = 'n'

Where n is the number of days to keep. So, if you want to get rid of anything more 60 days old, replace n with 60.

For WebMarshal 6.x:

exec dbo.PurgeLogData @PurgeDate = 'yyyymmdd', @MaxRecords = 0

Where yyyymmdd is the date from which everything older is deleted. So, if you wanted to get rid of everything from before 10th April 2010, enter yyyymmdd as 20100410.

Whichever command you use, after typing it in, click the “! Execute” button to perform the task, which may take some time.

Unfortunately, all this does is remove the data. It doesn’t shrink the data files, so your 4GB database will still be 4GB. Thankfully, there’s an easy way of doing this.

Back on the object explorer pane, right-click the WebMarshal database and choose Tasks > Shrink > Database, then click OK. Again, this may take a while. Once finished, for good measure, choose Tasks > Shrink > Files, then click OK.

Now your database should be much smaller! To check, right-click the database once more and choose Properties. In the General section, there’s an entry for Database Size. If this is way below 4GB, you’re done. If not, you might want to do a bit more purging!

Apparently, there’s some sort of football tournament on at the moment. Quite a big one too. I wouldn’t know.

However, the fact that it’s on, and some of the games are on during the working day, causes a few problems in that staff at work want to watch them. And they don’t have teevees. Me to the rescue then, with this motley crew of software and devices:

• An Elgato EyeTV 410
• A PowerMac G5
• VLC
• The VLC Plugin for EyeTV

Of course, this method should work for pretty much any supported EyeTV product and Mac. You’ll also need to know the IP address of the Mac you’re streaming from.

Firstly, you need to install the plugin. You can find this here. Once done, open EyeTV then open VLC. Make sure you can access EyeTV from VLC by choosing File > Open Capture Device and choose EyeTV from the selector. You may find you have to choose Screen then EyeTV again as it sometimes doesn’t “see” EyeTV right away. Pick a channel from the channel selector (you can change this later) and click OK. Don’t choose anything to do with streaming here.

You should now have an entry like “/eyetv-channel=3″ in your VLC playlists. Play this, and you’ll get your EyeTV channel in VLC!

Next, you need to so the streaming. Click File > Streaming/Exporting Wizard.

Choose Stream to Network, click Next, then choose the /eyetv playlist from the “Existing playlist item” list, and click Next again. For Streaming method choose HTTP, and leave Destination blank, then Next again. Leave all the Transcode options unticked, click Next, and choose “MPEG TS” for the encapsulation format. Click Next once more. In the next window, click Finish. You now have a ready-to-watch stream!

Now, on another machine on the network, open VLC. Choose File (or Media, depending which version you have) and Open Network Stream. For Protocol, choose HTTP, and in the address box type ipofmac:8080 (where “ipofmac” is the IP address of the Mac that’s providing the stream). Now Play!

All things being well, you should now be watching your TV stream across the network. If you need to change channel, you can do that on the “server” by simply flicking channel on the EyeTV controls.

If you too were shocked and horrified at the awful Bingification of Google today, here’s how to fix it. Which you’ll want to do. I mean, just look at it:

First, click “Add your own background image now”, and upload this suitably sized blank image. Set that as your background, Bing and You’re Done(TM).

Phew, eh?

Later this year, we’ll be migrating Microsoft Exchange 2003 over to 2007. Apparently, this isn’t technically a migration but is instead a transition. Something to do with moving and upgrading the same product, rather than changing from another product (like Lotus Notes or something, presumably). Semantics, eh? Tch.

This task is one fraught with peril at every turn, so requires a bit of planning and testing. The first issue is that Exchange 2007 only runs on a 64-bit version of Windows Server. Naturally, we only have 32-bit 2003. So we need to upgrade the server from Windows 2003 (32-bit) to Windows 2008 (64-bit), and there’s no upgrade path. We need to build a new server, but we don’t have the hardware and the current Exchange server is easily up to the task anyway, so after the transition it would be redundant.

Virtualisation to the rescue!

The intention, then, is to virtualise the current Exchange Server, plonk the virtual machine temporarily on another server, then wipe the original to put Windows 2008 and Exchange 2007 on. After that, there’s the task of actually doing the transition.

Today I started on a trial run of the process. I’ve already got a running 2008 server with Exchange 2007 on it ready, and I’ve found a way of making a virtual machine out of a running server: disk2vhd. Once you run it (and you don’t need to take your machine down to do so), you get this:

Yeah, the server is called Rupert.

Then we wait. For a very, very long time. A day, in fact.

After that, it should be as simple as plugging the new VHD into something like Virtual PC or Virtual Box. But it seems that wasn’t the case. I hadn’t taken account of the fact that the machine with Virtual Box installed on it was a single Quad Core processor, and the Exchange Server was running on a Dual Quad Core processor. So it wouldn’t boot, as it didn’t have enough CPUs.

Looks like I’m going to have to put it all on hold for a bit then, until I get some more hardware, or find a way of reducing the CPUs

Horror of horrors this morning: the wireless on my MacBook disappeared. It didn’t get physically removed, it just… vanished.

The icon I have on the menu bar just kept saying “No Airport Card installed”. I thought two things: either it’s physically fried, or my daughter’s “tinterneting” (that is, banging the keyboard) has made it pop loose. Both would require a dismantle, and it isn’t a job for the impatient.

Thankfully, there was another, less intrusive, thing to try first – zapping the PRAM.

Turn off the MacBook, then turn it back on and hold Command-Option-P-R. Keep them held down until you get three startup chimes, then let go.

As if by magic, the card “reappeared” and all is well with the world once more!

We all know that when you want to do conditional formatting in Office, you use Excel. Right? Or maybe Access. But what if you want to merge some data from Excel into Word, then conditionally format some of the fields? Word doesn’t do conditional formatting!

Except, of course, it can.

For this I’m going to assume you have some data in Excel. Two columns headed “Data Number” and “Letter”. The Data Number is just 1 to 10, and the Letter is a, b or c.

Now in Word, I want to merge this data in. Only I want the letter to appear formatted with a red background (if it’s the letter a) and a green background (if it’s not). We create a new Word document, and set it up as a Mail Merge letter. Point Word at the Excel file as the data.

You can insert the fields as normal with the button on the toolbar, and get this:

But how to do the formatting? You need to manually enter the field and make some changes to it. To enter a field yourself, press CTRL-F9 and you’ll get a pair of braces { }. These aren’t normal ones, as you can tell by their formatting. Inside these, you’ll need an “if” command. The format for “if” in Word is:

{ if [CONDITION] [DO THIS IF TRUE] [DO THIS IF FALSE] }

so we can do the following:

Which will show the Letter field if Letter = “a”, and the Letter field if Letter != “a”. Which doesn’t do anything we couldn’t do already. However – you can format individual bits of that command, like this:

Can you guess what this does? Try it and see! Of course, you can also change the fonts and weights for the true and false. I’ve attached a file with the data and template I’ve used. Click here to get it!

I’ve created a helpsheet for everyone who has ever wanted to create digital dress-up dolls! This shows how to import images to use as fills for clothing, so you can scan in some material and use if as a dress or something!

(Download the PDF here: Creating Dress-Up Mannequins)

I realise this may seem like a bit of a random thing for me to produce, but it was something I was asked to do for work. So that’s my excuse. And nothing to do with my desire to dress up dolls. Er, or something.

I tried to format a Buffalo 1TB external hard drive on my Mac today, to HFS+ (journalled). But it failed, over and over again with the error “Volume Erase Failed with the error: The underlying task reported failure on exit”.

It was a brand new drive, already formatted as FAT, so I assumed it was faulty. Turns out, however, there’s a simple reason for it not working: you can’t format HFS+ partitions on certain external hard drives if the boot sector is set to Windows-style “Master Boot Record”. In Disk Utility, when you choose your partition setup, there’s an options button – under which there’s a setting for boot sector. If you choose GUID or Apple Partition Map, you can format without any problems!

I had an unusual query this week. A member of staff wanted to create a line graph that showed the chart position of a song over a number of weeks.¬¨‚Ć It’s an easy thing to do – two columns in the spreadsheet, one for “Week” and one for “Position”.

Problem is, in a chart, the low numbers are at the top and the high numbers are at the bottom, so the graph comes out upside-down. You can’t just make the chart position numbers negative either, as the x-axis would then be on the top of the graph.

Thankfully, there’s still an option: right click the y-axis on the graph and choose “Format Axis”. Tick the boxes for “Values in reverse order” and “Maximum axis value”. While you’re here, you might want to change the Minimum value to 1 from 0, as there’s no 0 position in your average pop charts!

See:

You can also download the Excel file, if you like.