I found this article interesting, it basically says that with 2TB hard drives or larger sizes, you are more likely to encounter an unrecoverable read error. But is this just another Y2K doomsday? Don't HDD's have enough advanced hardware ECC error and read recovery to prevent this from happening?
I'm almost tempted to build a 3 x 2TB RAID 5 array and pull one drive out and rebuild to see what happens. I do agree with larger sizes, you have a more likely risk of failure and uncoverable RAID 5 arrays though.