Is it advisable to have a master or PhD thesis topic in mind before entering graduate school?

There may be advisers, departments, universities, countries (academic cultures) or fields where you need to come up with your own idea. I would not think this is the norm, however. In many cases Phd positions are financed by project funding so that the project is largely defined. This does not mean that the entire PhD is staked out in advance but the direction is. When you start a PhD you need an adviser which would imply that the field and direction of research of that adviser will determine the direction of your research. In many academic systems you directly apply to a PhD project which is pre-defined. I could probably come up with more cases that point away from come up with your own thesis topic.

That said, however, it is not inconceivable that someone could enter a system with an own idea but since coming up with great ideas commonly involves having a deep understanding of a field, and that in itself being one of the goals of a PhD, it would be a very rare case.

So depending where you are or where you are heading in the academic world, you do not need to know the thesis topic in advance. You will be looking for topics that may interest you and once finding positions announced decide if they fit your interests. It is rare that you find exactly what you dream of so many go for positions that are "close enough". Another point here is that you would probably not select a topic only, you would also consider the academic milieu and if you think it would be good for you and your endeavour into research.

This is a very local issue, depending (as Peter Jansson says) on the country, university, field etc. For instance, at least in the arts and humanities, it is usual in the UK to apply to PhD programmes with a thesis proposal that explains pretty much what your thesis will be on. Of course, there's nothing to stop you changing what you work on once you're accepted into the programme, have discussed it with your supervisor etc.

As to whether or not it's advisable, again that depends on local conditions. If your PhD has coursework then you have plenty of time to think about good thesis topics while you do that. On the other hand, for a purely research-based PhD like those in the UK, it's good to know the topic going in because you're supposed to get on with research straight away.

I've seen mixed results with this approach, to be honest. I ended up sticking with exactly what I proposed, but I suspect that's less common: research often doesn't turn out how one expects, some topics seem like a good idea at the start but as one learns more one's focus shifts and what appeared to be an interesting and tractable question turns out to be tedious or impossible to make progress on. Many of my fellow PhD students changed topic partway through, although often this was more a change of emphasis than a complete change of topic—that's less common, and correspondingly more difficult since ars longa, vita brevis (in particular, PhD students in the UK are expected to complete within four years).

This will very greatly by the discipline. My PhD was in mathematics, and I do not know of any of my colleagues that entered grad school with an idea of the problem that they would solve for their dissertation. In fact, most entered with no more than a notion of the area of mathematics that interested them (algebra, analysis, topology, applied math, logic) and at least half ended up in a different area from their notion on entry.

In mathematics it seems not feasible to me to know the area upon entry. The frontiers of mathematics are just too far away ...