One way of doing this is to strip out the hard drives and destroy them. New hard drives can be put in, and Windows reinstalled. However, that is a little wasteful, costly and takes a not-insignificant block of time.

Instead, I looked for some way to shred the files on the hard drives, and I discovered an excellent Open Source tool for doing this – Eraser.

Now, it is worth describing what the actual problem is here, because it is not obvious. When a file is deleted on Windows, it gets moved to the recycle bin. That means it does not actually get deleted, but instead just gets moved out of the way. At any time you can restore those deleted files.

The recycle bin can be cleaned out or emptied and that will delete the files completely. However, it does not erase the data in the files – the data is still there, and anybody with the technical know-how can recover that data from those files. How can that be?

Files stored on a hard disk consist of two parts: the data and a name in a folder. The name points to the data like an index card in [an older] library points to where a book can be found. Just as you would be able to rip up the index card without the book disappearing from the library, you can remove the name of a file without its data disappearing. All that happens when you delete a file is that the name is removed from its folder, but the data is still there.

The hard disk area where the data is stored is marked by Windows as “free”, so it will eventually get over-written, but there is no guarantee for how quickly this will happen.

So what can be done? This is where Eraser comes in. What Eraser will do is shred the file contents by overwriting it with rubbish – random characters. The idea is that once this is done, there will be no evidence left of the file contents. You can tell Eraser to erase files and folders that currently exist, or you can tell it to go through all the hard disk that is marked as empty, and shred every byte of that disk.

That is what I am doing with the dentist’s machines, and I am now happy that client’s personal details will not get out in the wrong hands.

This subject does go a lot deeper than this when you look at forensic security. For example, data that has sat on a hard disk for a long time, can “sink” deeper into the disk surface and be recoverable by some organisations, even after it has been over-written. However, considering the specialist equipment and skills needed to do that, I can’t see dental records being worth that much effort.

If you do use Eraser, do make a donation. Even a couple of pounds or dollars will help to support this great project.

Essentially the data provides 1.7 million UK postcodes, with their British National Grid Reference, local authority and ward listed for each. The most interesting data is the grid reference, supplied as a full Northing and Easting value. These can be used to calculate distances between postcodes, as well as being able to convert them into Latitude and Longitude data for displaying on maps such as Google’s offering.

The data does lack Northern Ireland and the Channel Island data at present, which can be a problem if you wish to use the data to cover the whole of the UK. The Northern Ireland grid system has a different origin and projection, so care must be taken if those postcodes and data are merged in from other sources. See here for details of The Irish Grid system.

There is a wiki that covers this data, but it is incomplete and has not been updated in nearly half a year, which is a bit strange. I suspect updates are being lost somewhere, because I cannot believe nobody wants to update it (there are technical inaccuracies in the description of the data that someone should have noticed by now). I have made a few additions to the wiki page, so we will see what happens next. If it sticks, I would encourage others to go in and add what you know and discover in the data.

So, what does this data look like? In short, taking into account the areas it does not cover, pretty much complete. I have loaded the current dataset into MySQL, then plotted each postcode as a pixel in an image. Some pixels contain more than one closely positioned postcodes, and a more sophisticated plot would show this, but a simple plot of pixels gives us an idea of what it looks like.

OpenData Postcodes - Great Britain - Click for larger image

Click the image to look at a much larger version with lots of detail. I particularly love how you can see the valleys in South Wales.

This simple PHP script generated the above image. All the table needs to contain are the Eastings and Northings from the Ordnance Survey data. It is surprising how it manages to get through 1.7 million postcodes in about ten seconds:

<?php
$width = 3200;
$height = $width * 1.5;

$db_name = 'opendata_postcodes';
$db_user = 'username';
$db_pass = 'password';

$easting_max = 655448;
$northing_max = 1213660;

$max_points = 2000000;

$db = new PDO('mysql:host=localhost;dbname='.$db_name.';charset=UTF-8', $db_user, $db_pass);

$sql = 'SELECT easting, northing FROM postcode LIMIT :limit';

$stmt = $db->prepare($sql);
$stmt->bindParam(':limit', $max_points, PDO::PARAM_INT);
$stmt->execute();

$img = imagecreate($width, $height);
$background = imagecolorallocate($img, 230, 230, 255);

while ($row = $stmt->fetch(PDO::FETCH_ASSOC, PDO::FETCH_ORI_NEXT)) {
  $x = round(($width*$row['easting'])/$easting_max);
  $y = $height - round(($height*$row['northing'])/$northing_max);
  imagesetpixel($img, $x, $y, 300);
}

header("Content-type: image/png");
imagepng($img);
imagedestroy($img);

Here is the same image with some crude colours introduced to show the density of postcodes.

Colours to show the density of postcodes

As usual, we start with the basic recipe, then try a few alterations each year. The idea is not for any single ingredient to be a prominent part of the taste, but for all to work together to create something new. This time the tweaks are:

• Add orange and lemon. I usually try one or the other, but I think putting half an orange and half a lemon into the pan with a bottle of wine kind of widens the spectrum of the flavours.
• Vanilla – a short length of vanilla pod. This adds a little something to the aroma.
• Juniper berries and allspice – just a couple of each.
• A couple of green cardamom pods, broken open.

A shot of vodka also adds a bit of warmth. We really must not let this become too much of a habit, but it’s cold outside, and it’s Christmas, so we can be forgiven.

The basic gist of the story is that my broadband IP address got black-listed as a SPAM generator, and that meant I could no longer send out email through the broadband mail server. It was frustrating, because it seemed that there was nothing I could do about it, but I have learnt a few tricks now to get over the problem next time.

It all started while repairing a laptop. The laptop had caught a virus or two and needed scanning and updating. Stupidly I plugged it into my home network to get various updates.

Now, on my broadband I run an SME Server. All traffic goes through that server. When I send emails from any PC in the house, the transparent proxy catches those emails, queues them up, then sends them all on via my cable provider’s SMTP server – smtp.blueyonder.co.uk

Normally the server scans all outgoing mail for spam and viruses, and it is usually pretty good at it. Unfortunately, what this laptop was spitting out consisted of extremely simple SPAM emails – just a single sentence and a short URL going through bit.ly. That slipped right through the SPAM scanner and in no time at all, 60,000 SPAM messages were queued up on the server and streaming out to Virgin broadband. Not that I knew this was happening, mind.

At some point, some part of the network noticed this, shook its head slowly, then put my IP address on a blacklist. This blacklist was then used by my ISP to basically cut me off from sending any email through its servers.

The result is that smtp.blueyonder.co.uk simply gave me a message telling me to get lost, then dropped the connection each time I tried to send out mail. This could be seen at the lowest level by connecting using telnet, pretending to be a mail client:

# telnet smtp.blueyonder.co.uk 25
Connected to smtp.blueyonder.co.uk.
Escape character is '^]'.
550-Administrative prohibition. This IP address is a source of spam. Please
550 Contact abuse@blueyonder.co.uk for details of how to fix this.
Connection closed by foreign host.

Switching off the transparent mail proxy of the server and sending direct to smtp.bluyonder.co.uk using Outlook Express or Thunderbird also gave this same error. Using telnet like this though, just removes the possibility that the email client is not set up correctly.

Okay, so need to tell my ISP – the IP address owner – to remove me from the blacklist. To cut a long story short, that proved impossible. There was simply no-one at the ISP with the authority to do so, or even the understanding of what I was asking for.

Next approach: the IP address is poisoned, so let’s get a new one. Easier said than done. I left the broadband modem off all night, I rebooted it many times, I reboot my server many times, but I just could not shake off that IP address. This almost permanent IP address is kind of nice to have normally, but now I  wanted to change it, it was proving difficult.

After several calls to the ISP, I finally managed to talk to someone who knew about these things. “Try bypassing your server and connect your PC direct to the broadband modem”, was his suggestion. I tried that, and behold, a new IP address was allocated. Cool. Just switch back to my server and…hmm, back to the old IP address.

So the Motorola Surfboard modem – or perhaps the cable box at the bottom of the street – was remembering the hardware and trying to allocate the same IP address to the same hardware every time. And it could remember the hardware for at least a full night.

So I guess I would have to change the network card in my server, which was a bit of a bind. Then I struck on this little gem: although the MAC address of the network card is coded into it, the operating system can override (or perhaps reprogram) that MAC address. By changing the MAX address, the network card in the server would look to the cable modem like a different piece of hardware.

I recorded the current MAC address, then added one to it, and issued this command:

# ifconfig eth1 hw ether 00:50:FF:E6:B9:4B

The network card was “eth1″ and the ethernet (“ether”) hardware address was being set to “00:50:FF:E6:B9:4B” (it was “00:50:FF:E6:B9:4A” originally). If you are having this problem with a Windows machine rather than a Linux server, then digging into the driver settings can often give you a box where the MAC address can be edited. Do use this option with care, since MAC addresses are supposed to be unique globally and two machines with identical MAC addresses nearby can play havoc with your network.

One reboot later and I was allocated a new IP address, and email was back online!

Of course, while I am celebrating, someone else will be given my old IP address, be labelled a SPAM creator, and will now have their own problems with their email. I have reported it to the ISP and explained what I think will be happening, and there is not a lot more I can do about it. I feel kind of bad about that, but I have learned never to plug a virus-laden machine into my network again, at least not without some kind of firewall and a virtual private network (VPN) to keep the prying eyes of any virus away from other machines on the network.

This MP3 is just a couple of minutes long, and contains a mix of scary Hallowe’en sounds. We used it on an iPod set on auto-repeat, with a small speaker by the front door, just to add a bit of spooky atmosphere.

The sounds came from various Creative Commons sources, and was put together using the excellent Open Source Audacity application. It is not broadcast quality (not that Audacity is not capable of that) but did the job.

halloween-background.mp3 (5Mbyte)

This is the power of Open Source tools and Creative Commons source: the important thing is not that the stuff is essentially free, it is the fact you are given the opportunity to be creative with the tools and media and to customise it without restriction. The right to do so, given to you through the licensing terms, is where the real power lies.

Ideally a zone should have a number of domain name servers located in different places. A zone is basically a domain and all its sub-domains. The zone record tells other servers where to find everything for that domain – where the websites are, the main servers etc.

Plesk run on a server or virtual server (VPS) generally provides two name servers. They will be configured automatically when websites are set up on the server, and are very convenient to use. Plesk also allows some customisation of the domain records so that, for example, a sub-domain can point to another server, or the mail server points to your office server.

The problem comes when the server running Plesk is down or inaccessible. A mail server trying to send an email to your domain will suddenly find there is no record anywhere of that domain – it has basically dissappeared.

The way around this is to use external name servers. By setting up the external name servers as a slave servers, they can be made to automatically keep themselves updated from your master name servers on your hosting machine.

For our domain, I used www.zoneedit.com to set up a slave for one of our zones. The slave was pointed at our hosting server (using its IP address) and that was it for that end – the zoneedit slave just fetches updates when it is appropriate and updates its own records, putting them onto two geographically-separated name servers of its own.

Back to the domain registration settings, I then needed to add the zoneedit name servers to the domain, giving me four in total (the two on our Plesk hosting and the two on zoneedit). There is no priority assigned to any of these name servers, so a server looking for our website or mail server could go to any of these sources. It is therefore very important to ensure they are all synchronised identically.

One last step was to tell Plesk to allow these name servers to fetch a copy of the zone records. This is not so obvious but is in Plesk under the option:

Server -> DNS Settings -> Common ACL

In here I added the IP addresses of the two zoneedit name servers assigned to my zone, and then the updates simply went ahead.

You can test this works using nslookup command from another server. Create a new sub-domain, then check it can be seen on your Plesk server:

nslookup new.sub-domain.example.com ns.plask-server.example.net

Then try the same command but pointed at the external name server. You will find the external name server cannot see the sub-domain immediately. Try again in an hour – and it should be there. If not, check your setting and TTL etc.

Hopefully that will give you more resilience; even if your website is down, other severs that your zone points to are still locatable.

One thing that I am still looking for a solution to, is domain aliases. A domain alias for a domain set up in Plesk will not get transferred to the slave name server. Even if the alias is a sub-domain of the main domain, it does not get transferred. So the hosting server will correctly return the IP address for example.com for the alias myalias.example.com, but the slave name servers will not. This is not just a problem for when your hosting server goes down, because there is no priority on the name servers – all and any could be used at any time for a server to locate your alias domain, and depending on which name server it hits, it is pot luck as to whether it finds it or not. I’m looking for a solution to this, so any hints appreciated.

I’m also looking for a decent and cheap slave name server service. zoneedit is good and reliable, and easy to use, but could get costly when we have scores of domains to sync up. Any suggestions?

Basically, there is a key combination – somewhere around the “A” on my keyboard, probably involving the control and perhaps the shift keys, and is so easy to accidentally press when not being careful, that locks up all my keyboard input.

When I say “locks up”, I mean that the keyboard becomes completely unresponsive. Nothing gets it back, except a full reboot, and if you know how long my PC takes to boot, you would realise it is not very convenient (I do manage to make a cup of tea while it boots, which is a bonus, kind of).

I have tried unplugging and plugging back in the keyboard. I have tried undocking the PC (I use a Toshiba Tecra laptop in a docking bay). I have tried logging off and on again. Yes, I have tried logging off and on again. None of this works; only a full reboot does the job.

It locks up both the external Logitech keyboard and the laptop’s built-in keyboard, so there appears to be no workaround.

So – what could this be? I have a feeling it is some kind of “feature” that is misbehaving, but cannot think what it could be. I have not even been able to find the key combination purposefully to give me any more clues. It just happens, and when it does I know my fingers were pressing some keys around the “A” end of the keyboard.

I have a feeling I will never find a solution to this. Perhaps I need a new laptop.

http://magento.consil.co.uk/themes.php

This script is installed in the root directory of our 1.4 test installation.

What it gives you is a layered approach. First you select the interface, then the theme within that interface (try default->modern for example).

The script will then list all the layout files that make up that theme, and show you where those layout files come from, i.e. which parent themes they are inherited from. A complex theme may consist mostly of its own layout files (the iPhone theme for example). A simpler theme may override just a few layout files, choosing to fall back to the default theme, or the new “base” theme in Magento 1.4.

For each layout file, the handles are listed. Each handle provides some unit of functionality in the theme – usually a block or a widget – some simple, some very complex. The next step will be to document what all these handles do, and how they get called up. Some are invoked from the Magento core, some from modules, and some from other templates in the theme, so it can all get very complex very quickly, leaving the average themer lost when they only wanted to change a few aspects of a theme.

It is the content of the handles that are the really important part of a theme. The layout files just provide a convenient way to package together a number of handles in one place. The same handle can appear in multiple layout files, and they are simply chained together in alphabetic order (of the name of the layout file).

What the layout files do is allow a module to keep all its handle-based theme functionality in one place. A theme developer could also use it to keep all their theme additions in one place too – there is no point spreading that functionality all over a custom theme by overriding dozens of layout files if all you want to do is tweak menus and blocks.

By clicking on the “handle” link in the script, the table gets inverted – now you get to see all the layout files that each handle is included in.

Selecting a handle will list the contents of each instance of those handles – the XML instructions to the theme engine.

To take this script further, it would make sense to interpret the content of the handles. Ultimately it should then be possible to do searches on those interpreted instructions to find, say, where a particular block or image is displayed from.

Feedback is welcomed – is this any use? Should I publish the script? Does anyone want to expand on it, perhaps packaging it into a module to help theme developers?

Download the script here: themes.zip (contains theme.php, to be copied to your Magento root directory).

I took a look at free or cheap online fax sending services in the UK, and many offer lots of promise in the Google search results, but few deliver once you go through their long registration processes – usually “free” means “subscribe monthly” by the time you get to the site.

I did come across FreePopFax though that seems to work a treat. Just upload your Word for PDF document, enter the recipient’s fax number and your e-mail address, and off it goes. You then get an e-mail when the fax is sent.

So far as I can see, FreePopFax is an affiliate of the monthly subscription service PopFax, but adverts are inserted to pay for the service. Bear that in mind – the recipient may receive adverts.

I would add a disclaimer: I have no idea who is reading these faxes en-route, so don’t use it for anything that contains credit card or other personal details that could result in identity theft, but for the odd fax, give it a go.

The Plesk web management interface is pretty cool. There are lots of sophisticated things you may want to do on a web server, but 99% of those tasks are encapsulated into the admin screens provided by Plesk; it simply gets the job done of managing web hosting.

Every now and then you will probably want to reorganise the domains, clients and databases within a server. One task includes moving a database from one domain to another.

At the server level there is no concept of a database being owned by a domain – all databases are global to the server. Within Plesk, however, each domain has its own associated databases which can only be managed from within the account that owns that domain.

To move a database to another domain, you can export it, delete the old database, create a new one in the new domain, then reimport the database. There is an easier way though.

Log into the server using ssh then issue this command to get into the MySQL command line tool:

$ mysql -u admin -p -D psa

You will need to enter your Plesk admin password.

Next take a look at the current database list:

mysql> select * from data_bases;

+----+-------------------+-------+--------+--------------+-----------------+
| id | name              | type  | dom_id | db_server_id | default_user_id |
+----+-------------------+-------+--------+--------------+-----------------+
|  1 | consilience_db    | mysql |      1 |            1 |               1 |
|  2 | client1_db        | mysql |     18 |            1 |               2 |
| 29 | client2_db        | mysql |      6 |            1 |               4 |
+----+-------------------+-------+--------+--------------+-----------------+
3 rows in set (0.00 sec)

Each database will be listed with their name and “dom_id” (domain ID). Find the ID of the database you want to move, and find the dom_id of the domain you want to move it to (the dom_id will appear in the URLs used to manage the domain in the Plesk admin screens). Then update the dom_id, for example to set database ID 29 to be owned by domain 18:

mysql> update data_bases set dom_id = 18 where id = 29;

That’s it – the database will now appear under the new domain in Plesk.

This will only work on a single server. If you are migrating between servers then there are other tools for the job.