No, I'm not trying to sell you anything.
There are all sorts of reasons to do a PhD, all of them insane. I haven't finished mine yet, so I may not be the best person to give advice on the topic of how to get the thesis out the door, but here are a few things that I wish people had said to me when I started. Hopefully it might help to make your journey a little bit easier.
You'll read lots of stuff. So that you don't read lots of stuff, forget lots of stuff, read it again with a vaguely familiar feeling, forget it again, then design an experiment that seems perfect only to find you've redone an experiment you read about two years ago, summarise all the papers you read, even the crap ones. This will help clarify your thoughts on the papers and provide pointers to them later. You still may end up reading the same paper a few times but hopefully summarising them will let you get more out of them each time.
You could choose to implement a specialised system for keeping track of the papers that you've read, but just putting them in the literature review chapter of your thesis template works pretty well. At the same time, you should enter all the required bibliographic information in whatever bibliographic information management software you're going to use.
Keep track of your citations
Did I mention that you should use bibliographic information management software? You must, whether it's BibTeX, EndNote or something else. The last thing you want to do is go back to that paper you read that's been sitting on your desk for a year because it's just completely on point and then spend a week trying to find the goddamn citation details because you didn't write down where you got it from.
In fact, don't just write about the papers you've read. Write about anything and everything that pops into your head that's even vaguely related to your research. Maybe even stuff that's not related. It will improve your written expression and it will clarify your thoughts. Some people keep a paper journal, but I recommend blogging, partly because your words will be searchable and viewable in different ways--by date or by tag, for example--but also because I type faster, and more legibly, than I write.
You can choose to keep your super special secret sauce under wraps so that you don't get scooped. I keep two blogs, one private, one public (that would be this one), but in hindsight I probably should have written more about my research publicly. An additional benefit of a public blog is that you can hand out the address to those people who come up to you at conferences. Sure, they may be completely bonkers to be interested in what you're doing, but insane recognition is better than none.
Get stuff in the template
If you haven't already, download your thesis template now. Right now. Become familiar with the layout. While you're there, find a couple of theses in related areas and have a look at them too. It's fun to read what your supervisor wrote nearly 20 years ago. Try and work out what needs to go where. Put in some headings and regenerate the table of contents. That's your plan, see?
You don't have to constrain yourself to only filling in the literature review, write down a paragraph or two under each heading. With all the guff in the template it's probably 20 pages long already. Not far to go, hey?
This is a bit of a tricky one if you're doing a PhD in a discipline where versioning is not common, but I highly recommend it. If you're programming at all, you have no choice, set up a repository using your favourite versioning software. If you don't have a favourite, ask someone clever. I use subversion, and I'm happy.
Just to be clear, I'm not just talking about code. I'm talking about versioning your writing output. In research, that's your thesis and any papers you're writing. Submit a paper to a conference? Tag the release. Got accepted? Well done. Reviewers give you good feedback? Fix the paper and tag the final release.
Getting work done
If you're anything like me, you're likely to find the world an immensely interesting place. Hugely more interesting than your research topic, in fact. The one rule that helped me start making some real progress was that whenever I sat down at the computer, the first thing I did had to be research related. Go to the toilet, get something to eat, get some exercise--highly recommended--go to bed, but when you come back to the computer, research is the first priority. It doesn't really matter what it is, it's just got to be something to do with your research. What is important is what it's not. It's not checking your email, or catching up on the news headlines, or reading the 612 unread items in your feed reader, or playing Nethack. Yes, I know, I'm linking to Nethack. Don't go there.
A final word ...
Don't procrastinate by writing crap on your blog though.
Good luck with your study!