Miscellaneous Knowledge

It’s always a horrifying experience to find that the backups you so carefully took previously now don’t work. Unfortunately, i’ve just run in to this problem.

After sending my phone back for repair, I attempted to restore the backups I took beforehand using Titanium Backup. While it recognised the backups, every one said “1 backup. None on chosen platform”. Unfortunately, a quick Google only turned up a handful of other people who were having the same problem, with no resolution.

After a bit of fiddling however, I came across a simple resolution. On the Backup/Restore tab, tap “Click to edit filters” and then “Filter by platform”. In the Filter by ROM build date section, you’ll probably see two or more values: only one of which is checked. Check all of them, then tap the checkmark to accept your new filter. Your backups should now read “1 backup: Latest xx/yy/zzzz 00:00″ and can be restored.

Phew.

RSS is a classic example of a technology which relies on the client pulling data from the server, rather than the server pushing data to the client. While pull is simpler to implement, it requires the client to periodically poll the server for new content to pull. Unless you poll the server continuously, you’ll never get new data in realtime.

This is a common problem for RSS feed readers – how do you balance the requirement for getting new articles in a timely fashion versus not wasting resources communicating when no updates have happened? The answer for most RSS feed readers to to build up information about the frequency of posts to that feed and then pull at a suitable interval. NewsBlur is one site that follows this methodology, the screenshot below shows how often NewsBlur fetches the feed for this blog and why;

NewsBlur’s fetch interval for this blog.

As can be seen, because so few posts are made (I should really resolve that), the feed is only fetched 7 to 9 days. So in the worst case scenario, someone who subscribes to this blog may only be notified of a new article by there reader 9 days after it was originally posted. If there was some way that this blog could push a notification to the RSS feed reader that a new article had been posted though, then the delay between posting a new article and being notified of it could be dramatically reduced.

Enter PubSubHubbub – a protocol which uses the publish/subscribe architecture for notifying feed subscribers when a new article has been published to a feed. The PubSubHubbub Google Code Page has lots of information on implementing both the publish and subscribe sides of the protocol. However, if you’re using a popular blogging platform such as WordPress, then you can easily implement it via plugins which have been already developed.

After installing a PubSubHubbub plugin on this blog, NewsBlur’s fetch interval changes to the following;

NewsBlur’s fetch interval for this blog with PubSubHubbub implemented.

While the feed is still periodically fetched at certain intervals as per NewsBlur’s previous algorithm, it also receives any realtime updates, significantly reducing the timelag between me publishing this post and it appearing in RSS feed readers. Hurrah!

I’ve just come to the end of my 3rd week of my year-long industrial placement. For those not in the know, i’m working in the IT department of a medium sized local company.

People wonder why I like working in IT, my reasoning is that I enjoy playing with systems and making them work, and doing it properly. Sure, doing it a hackish way works, but it’s probably not good in the long run and does run nails down the internal chalkboard of my head! To give an example, i’ve been working on getting Microsoft’s Lync ready for deployment – expanding it beyond the current small test userbase. Lync can be thought of as a corporate version of Skype – enabling instant messaging, voice and video chat. However, it nicely ties in with other systems, such as the user database (Active Directory) and presence information (via Outlook Calendaring), something Skype doesn’t do.

Anyway, one of the problems with Lync is that on the first run it can take anywhere up to 60 minutes to download the address book. Without the address book, users cannot search for other people in the company to communicate with which isn’t really ideal. It turns out there is a registry setting that can adjust this interval to a user specified interval. So, all we need to do is deploy this registry setting to all users. Simple right?

Okay, so we can create a group policy that runs a command to add the setting to the registry on machine start up. This works fine, but it’s somewhat hackish – what if we want to remove the registry entry some point down the line (even though it dosen’t affect anything else)? Removing the policy won’t affect anything since the command has been executed, that registry entry will remain in the registry forever!

Windows Server 2008′s group policy enhancements solved this by adding preferences, registry entries can now be properly defined via group policies. So, we’ll just use preferences. Simple right?

Not quite, preferences are only supported on Windows Vista and higher by default and there are still some XP clients left in the organisation. Thankfully, Microsoft has published an update called Client Side Extensions that allows Windows XP clients to take advantage of the preferences policy. So now it’s just a matter of ensuring that the XP machines have this update. Unfortunately, it’s an optional update so the XP clients won’t get it automatically via Windows Update!

In an ideal world, Microsoft would have made the update available as an MSI and we could deploy it via group policy, but no – it’s available as an EXE only. We could still deploy via group policy by having it run as a command on start up but, like the original registry entry, this is a very hackish solution that isn’t really ideal.

There is one further way of deploying this update and that is to centrally manage updates rather than having the clients contact Windows Update themselves. This can be achieved via Windows Server Update Services (WSUS). Of course, this isn’t availble within the organisation at the moment, it’s on the neverending todo list (of which i’ve been assigned lots of tasks, which is good for giving me a range of experience). So, Windows Server Update Services needs to be installed in order to ensure we can force this optional update. Since you never go straight into production, we need to test it first, so a new test server is being built to run WSUS. If it works fine, then we can make another server in production.

So to sum up, in order to have the Lync address book’s initial sync perform much faster we need to set up a new test server and if all goes well, a new production server, taking several days.

It’s the hilariously inefficient/crazy things like this that make me love working in IT. It might take a while, but it’ll be done the right way – and won’t suddenly cause problems 1/2/n years down the line.

iptables is a common linux software firewall. However, for many new linux users who are unfamiliar with the command line (or google!) some simple tasks such as opening a port on the firewall in order to allow a web-facing service to be reachable by everyone can be daunting.

However, it’s actually quite simple. Just issue the following command;

iptables -I INPUT -p tcp –dport PORT -j ACCEPT

Where PORT is the port number you wish to open, e.g 80. Note that there is a double dash before dport.

The syntax of this command is quite simple. -I INPUT indicates you are altering the input table -p tcp indicates that the protocol is TCP, –dport PORT indicates the destination port and -j accept indicates the action to be taken on matching the rule – in this case to accept traffic. Note that the command requires superuser privileges to execute.

Don’t forget that this will be in vain if iptables isn’t actually running. Under CentOS, you can start the iptables service by issuing the command service iptables start. You’ll also need to save the iptables rules if you want them to persist across a reboot, under CentOS you can issue the command /sbin/service iptables save to achieve this.

One and a half years after originally obtaining my Nokia E71, my contract has come up for renewal – new phone time! Before I talk about my new phone though, a few words on how the E71 has performed for me while i’ve had it.

Overall, it’s been really good. I’ve taken the phone pretty much everywhere with me every day and used it to keep track of emails, add appointments, browse the web, watch YouTube, prevent myself from getting lost and listen to music. As expected, the phone’s initial battery life has decreased over time especially considering the heavy usage i’ve put it through – it now lasts for about a day with light usage. If I bought a new battery however i’m sure it would be restored to it’s former glory. Of course, having used it for a year and a half there are also a few long-term gripes i’ve had with the device;

Music Player

If you add a new song to your phone the music player application takes a long time to rescan your music collection if you have more than about 250 songs. With an 8GB memory card, you can fit quite a few songs on the device so this ended up being a major problem when I added anything new.

Video Player (‘RealPlayer’)

The included RealPlayer video player is just as pathetic as the desktop version, it supports extremely few formats and those formats that it does support to be encoded in a special way in order to play. This became annoying when trying to transcode video to play on the device.

Unicode Fonts

By default, my European version of the phone only had support for western-European characters. A lot of my songs have non-English titles and attempting to distinguish between different artists, albums and songs by the number of squares became quite an art. However, I eventually found a way to load Unicode fonts onto the phone (via the Memory Card, a quick google should find the method) and I was able to see the proper titles of my music tracks.

Internet Connection

The phone has two ways to connect to the internet, via the mobile network and via WiFi. Unfortunately, there is no way to say “Use WiFi if available, if not fall back to the mobile network”. Consequently, all of my access was via the mobile network which, while functional, isn’t as fast as what i’d get over WiFi and eats some of my data allowance.

Email Client

A lack of a native HTML email client made reading some emails awkward. Nokia now offers it’s Ovi Mail service, but i’d rather connect to my exchange server directly rather than through them. It’s reasonably good for POP/IMAP accounts though.

Headphone Socket

The headphone socket takes a 2.5mm connection, so you can either use the terrible earbuds that came with the phone or get a cheap adaptor off eBay, so you can use your 3.5mm headphones. I opted for the latter solution, but the only downside is that the adaptors are cheap and nasty – they break every few months and consequently I have to order a new one.

Charging

Once the device was fully charged, it would switch to battery power, which would deplete slightly and trigger charging again. Aside from the annoyance of having the screen activate every 15-30 minutes to inform me that it was charging, I don’t think this did much good for the longevity of my battery.

As I mentioned earlier though, I have been very pleased with the device overall. So if you’re looking at getting a cheap 2nd hand Nokia E71 off eBay, go for it. You may need to buy a new battery though! The newer E72 fixes a lot of the hardware issues I noted above as well as shipping with a newer version of Symbian. If you’re able to get a good deal one one, then give it a go. Otherwise, i’d highly reccomend getting an Android-based device.

In my original post I noted that at home I received hardly any signal, but it was perfect everywhere else. Since then, i’ve gone to uni (for over a year now) and the signal all over Bath and the uni campus has been perfect, even inside buildings. It seems 3 is the only mobile network with decent coverage on campus too! Anyway, when returning home i’ve noticed that the signal is now full, 3 appear to have increased coverage in the area. Now, the only time I get no signal is when i’m travelling on the tube.

And for anyone who is after my old E71, sorry, you can’t have it. It’s been passed on to a family member – recycle!

I recently bought an Asus Xonar DX as I was getting fed up of the interference on my onboard sound card. However, as I use headphones I often need to use the soundcard’s functions to boost the bass a bit if I want to hear it. While most applications would go via the Xonar control panel and boost the bass, ffdshow would bypass it and not boost the bass. Hence, any movies I played where distinctly lacking bass.

To solve this, simply set Audio Channel to something other than “2 Channels” on the main section of the control panel. Don’t worry, the Xonar card automatically handles downmixing back to 2 channels (or whatever your Analogue Out is set as) for the actual output. This is simply the interface it presents to programs such as ffdshow. Now, ffdshow should go via the Xonar control panel and apply all the effects you’ve set, such as bass boosting.

Ever felt like trying out the KDE desktop interface, but didn’t fancy downloading Kubuntu and installing from scratch? Good news! You can easily install KDE from your current Gnome-based Ubuntu install and it’s dead simple.

Open up Synaptic Package Manager and search for “kubuntu” in the quick search, in the list of results select kubuntu-desktop and then click apply. Synaptic will then download and install all the files for KDE.

However, how do you switch to KDE from Gnome once it’s installed? Simply log out and then click your username as usual, at the bottom of your screen click Session and select KDE and then log in as normal. This time, instead of Gnome you’ll get the KDE desktop interface. If you want to switch back to Gnome just log out and select Gnome under session.

Now you can try out a new desktop interface without going through a lot of hassle.

Recently, my phone has been complaining that it is out of memory, and hence disables automatic retreival of email amongst other things. Try as I might, I just couldn’t free up any extra space – I install programs to the memory card and keep all of my files on there. I couldn’t see anything on the phone I could delete.

However, all it seems all messages (Text and Email) are stored on the phone’s main memory, for someone who has a lot of emails and never deletes text this amounts to a surprising amount of data.

Thankfully, it’s possible to move the messages from the phone on to the memory card. Simply go Menu > Communication > Messaging > Options > Settings > Other and set “Memory in use” to Memory Card. You’ll then be asked if you want to copy over the messages, select Yes and your messages will be copied to your memory card and all future messages will be stored there. Doing this freed up a whopping 55MB of space on the phone memory!

A few months back, I picked up a new phone and contact. Before this, I used a SE k800i on Orange PAYG. While this worked fine for calls and texts (albeit in low volume) internet was a no go. Therefore, when I saw 3′s offer of a free phone with 1GB (“unlimited”) internet for £20/mo, I snapped it up straight away – despite all the horror stories you hear about 3.

The first phone I received was faulty, it wouldn’t charge. I took it back to the 3 shop the next day and got a replacement there and then with out any problems, good customer service there. Living in London, I get a good signal pretty much everywhere except at home. Everywhere else i’ve been has given me a medium or strong signal. While mildly annoying at times, I often use the WLAN instead of 3G at home instead.

Having not traveled outside of London since getting the phone, i’ve not had a chance to see what coverage is like. However, judging by the recently released maps from Ofcom, 3 has the best 3G coverage of any network. I’ll update this after my holiday in Mid-July where I will be on the canals of North Wales.

Recently, I also got £30 of credit on my bill for referring my Dad, which was very welcome!

As for the phone itself, i’ve got to admit I am very impressed – it does everything I want. It syncs my contacts, calendar, email etc with my Exchange server, the inbuilt web browser is fast and renders pages well, the music player is standard but good enough. In addition to this, there are many other applications (There are so many applications for the Symbian platform). The only prolonged issue I had with the phone was occasional lockups. The phone came with the first ever software revision, however since it was 3-branded I was unable to upgrade to the latest version. A simple change of the product code however meant that I was able to upgrade the phone to the latest software which gave a noticeable speed boost to the phone, increased battery life, removed 3 branding and also enabled the SIP client. Using the SIP client I am now able to have a second internet phone number on the phone which is extremely useful. However, while it works over 3G it uses the radio continually (even when the phone isn’t in use) and will drain battery life fast, operating over WLAN there are no issues.

Overall, i’m extremely pleased with the both the operator and phone, i’ll be sure to update this if it changes.

I run several Left 4 Dead servers which I play on regularly, these are all public servers with unrestricted access. Unlike previous Valve games, Left 4 Dead includes a built in vote kick system, which is sensible since due to the server sizes, it’s unlikely you will have an admin player playing, hence the voting system is used for removing any troublemakers.

However, it is frequently abused, players will randomly votekick another player without warning because they want a friend to come on, without even asking. In many cases, they’ve even votekicked me, the owner and admin.

While they soon realise this is a huge mistake on their behalf (as they become permanently banned from all my servers), it is somewhat annoying to have your gameplay interruped by a votekick.

Thankfully, due to the power of SourceMod, someone has written a plugin that obeys the admin immunity system built into sourcemod – normally used to prevent lower-level admins from abusing higher-level admins. By default, all players have an immunity of 0, as a “Full Admin” I have given myself an immunity of 100. Hence, when anyone tries to votekick me, the vote instantly fails. Not only this, it shows me who it is, for all votes.

By default, the plugin only grants full admins veto ability and displays who is votekicking who. I recommend you edit the /left4dead/cfg/sourcemod/sm_plugin_votemanager2.cfg file as follows; (if it does not exist, create it)

l4d_vote_kick_immunity 1
l4d_vote_log 1

This will turn on the admin immunity system and also log all votekick attempts to the SourceMod log.

Players beware, never kick the admin.

Today, I received the following email from Google regarding their checkout service, which allows businesses like my own to receive credit card payments into our bank accounts:

Hello,

We’re writing to let you know that on 5 May 2009, Google Checkout’s transaction processing fees will be changing. We will be transitioning from our 1.5% plus £0.15 per transaction rate to a new tiered fee structure, where the rates will vary depending on the amount of your monthly sales processed through Checkout. The rate you will be charged beginning on 5 May 2009 will be based on your sales processed through Checkout during the month of April 2009. Each month thereafter, we’ll continue to use the prior month’s sales volume to determine your transaction processing rate. For more details about the new rates, please visit https://checkout.google.com/seller/fees.html?hl=en&gl=GB

We will also be discontinuing the AdWords free transaction processing promotion on 5 May 2009. Any AdWords transaction processing credits accrued during April 2009 will be applied towards transactions that occur on 1-4 May 2009.

Fees are the same for all payment types (Visa, MasterCard, UK Visa Electron, UK Maestro, Solo) and there are still no monthly, setup, or gateway fees. For cross-border transactions, there will be an additional 1% fee assessed per transaction. To learn more about Google Checkout fees, please visit https://checkout.google.com/seller/fees.html?hl=en&gl=GB

So in short, the fees that Google are charging small sellers like myself who don’t process more than £1,500 in sales with them per month have gone from 1.5% + 15p to 3.4% + 20p, which is exactly the same as PayPal charges. To add insult to injury, they are also withdrawing the AdWords free credit (for every £1 you spent on adwords, £10 worth of transactions where processed free) and they are adding an extra 1% for any cross-border transactions. As we do a lot of sales outside the UK, this makes Google Checkout even more expensive.

Many users are in outrage at the change, and this will most likely cost them dearly. The only winners here are the large companies processing over £55,000 a month with Google Checkout. I am voting with my wallet, and will now no longer suggest Google Checkout as our preferred payment option.

So if you’re looking for a payment solution now, I highly reccomend PayPal, lower fees, telephone support and much better brand recognition.

For the past two years, my main desktop which is designed for high performance gaming (and occasionally, work) has had hard drives running in a RAID0 array. A RAID0 array takes two hard drives and combines them to form one larger drive, it also splits the data evenly across them

RAID0 Data Striping

Image from http://en.wikipedia.org/wiki/File:RAID_0.svg

However, by using RAID0 with two drives you automatically double the chance of failure. In a simple RAID0 array, failure of one drive leaves you with exactly half your data, pieces A2, A4, A6 and A8 in the diagram. You’ve got half the data, but it’s not usable in any shape or form (imagine reading a book with every other page missing).

Now after two years, one drive has finally failed – without warning. No smart errors, nothing. For a few days before, I had been receiving I/O related bluescreens however and I should have taken this chance to backup the entire drive.

Thankfully, most of my data is stored on a local storage server, running FreeNAS. The drives in here are thankfully RAID1 – where the data is equal on both disks so if one drive fails you have an identical set of data on the second. Therefore, the only thing i’ve really lost is installed programs, easy to replace.

I’ve now ordered a new single drive (750GB Western Digital Caviar Black). I’m done with RAID0!

The Network Time Protocol is, as the name implies, a standard for syncronising time across networks. This is extremely useful as it allows all computer clocks to be syncronised to a common, correct time rather than being “around” a certain time. It’s especially useful on servers, where co-ordinating events or tracing failures needs precision and accuracy.

So how can you sync your computer with a NTP time server? It’s actually quite simple;

• Windows – Windows XP and higher already sync out of the box to time.windows.com, however time.windows.com may be very far from you and hence there is a delay in the time making it less accurate. If you want to change the server, double click the clock (XP, on Vista double click and select “Change Time and Date Settings”), then click the Internet time tab, enter the server (XP, on Vista click Change Settings) and click Update Now to sync.
• Linux – Use the command ntpdate followed by the NTP server you want to sync with

So, how do you go about choosing a better NTP source for windows and a source for linux? Enter the NTP Pool.

There are quite a few “Stratum 1″ NTP servers, which get their time directly from cesium clocks or GPS satelits, these are the most accurate sources of time available. However, if we all synced with a stratum 1 server, they would quickly be overloaded. Most people don’t need that kind of precision, so enter “Stratum 2″ servers. Stratum 2 servers NTP sync with “Stratum 1″ servers, so while they are not as exact as they are, they are only different by a few milliseconds – hadrly a problem. Hence, you should sync with a “Stratum 2″ server. But how do you find one? The NTP pool is a pool, a collection of NTP servers intended for the masses to sync with. Either use 0.pool.ntp.org as your NTP sync source, or if you want extra accuracy select your continent, then country from the right to get a server closer to you. Now you can enjoy accurate time!

However, the demands upon the NTP pool grow daily, but not it’s supply of pooled servers. If you have a server with some spare bandwidth and a static ip address that is available 24/7, donate a server to the pool. You can adjust the amount of NTP traffic you get by setting your bandwidth in the control panel, even home users can participate as long as they meet the requirements!

OpenTTD, as it’s name implies, is an open-source remake of 1995 game Transport Tycoon Deluxe. Back when I had my first computer, TTD was one of my favourite games. So what was Transport Tycoon all about? Basically, you have to build a transport network, using Trains, Planes, Trucks, Buses, Boats etc and earn as much money as possible, sounds simple but when you start building huge networks, all that can go out of the window!

OpenTTD has taken the original game, made for DOS and hence incompatible with 2000, XP and so on and re-written it completely in C. However, it dosen’t stop there – as it’s open source many enhancements have been made which hugely enhance the game experience. For example, trains are the major network tool, able to carry large amounts of cargo at high speeds. However, with large networks and large numbers of trains comes problems – signalling, platforming and so on. The original TTD only allowed for a maximun of 4 platforms at a station – hardly a huge amount when dealing with large towns or industry, signals where very basic and had little logic, no custom graphics could be added and so on.

One thing I always found lacking, even in OpenTTD however was passenger destinations – passengers would just turn up at stations and you could deliver them absolutely anywhere, giving you no incentive to link up every town. This is in stark constrast to the real world (I presume!) where people have all different destinations, and may need to take several modes of transport to reach their final destination. Again however, the open source nature of OpenTTD trumphs again – a Passengers & Cargo destinations version of OpenTTD is in developement, and once complete will be merged into the main version of the game. This adds another dimension to the game, ensuring that certain key routes are not overloaded which can often be an interesting problem to solve.

If you’re looking for an interesting game, which runs on pretty much anything and will keep you entertained and thinking then go give OpenTTD a try. The only restriction is that you will require the graphics from the original TTD, which can’t be distributed for legal reasons, however they are easily found with a bit of searching.

On my cPanel server, I noticed that users where being given the incorrect FQDN (Fully Qualified Domain Name, e.g example.com) for mail.root as the SSL address for the mail server. This is odd, since there is no where to actually configure the SSL host address. After a bit of poking around, it seems it sets the address based upon the address of your SSL cert. In my case, I was using a Wildcard SSL certificate from CACert (free SSL certificates). However, as the certificate was a wildcard certificate, it had no idea what the actual FQDN you wanted was! After purchasing a cheap RapidSSL cert for £10 and installing it, the SSL Host FQDN displayed correctly for users. Plus they won’t get any certificate warnings any more!