A viva typically lasts two hours (but as long as it takes - mine lasted 2h22m), and a common approach is for the examiners to go through the thesis sequentially, asking questions.
Just because they ask a lot of questions doesn't mean you're going to fail.
A Ph D candidate needs to anticipate the questions that are likely to be asked in the viva - the "horrible ordeal where you have to defend your thesis in person before they rip you to shreds." Actually, it's not nearly as bad as it sounds, provided that you enter it having prepared to your utmost.
There are three reasons why Ph D candidates have to have a viva: it is so the examiners can see: It's crucial to get the philosophy of your thesis (as set out in your Chapter 1) absolutely correct, and clear in your mind by the time of the viva, because if the examiners find holes, they'll run rings round you.
Anyone can attend a Ph D viva, but only the examiners and the candidate can participate.
(This means it may be a good idea to attend someone else's viva before your own, though I've never had the balls to gate-crash a viva!
The most important goal in preparing for the viva is to keep the subject alive in your head.
Try to anticipate the questions you'll be asked in your viva and keep working on a file of anticipated questions (both the generic questions listed on this web-page, and questions specific to particular sections of your thesis) and your answers.
They don't give away the result before or during the viva, but you may be asked to wait around for the result at the end (about half an hour), so that they can explain the result to you - particularly if you have to resubmit your thesis (failure without the option of resubmission is very rare, and is not going to happen if you submit anything resembling a sensible thesis).
Tips: Here are some generic viva-questions - you should instantiate each question for your particular thesis, and have a framework for answering it worked out before the viva.