I have a few machines which run 24/7. As a general rule, I replace disks when they reach about 30,000 runtime hours (~ 4 years) if they were not already replaced in the meantime by faster and bigger drives. Starting at this time, they become spare parts and often get an easier second life as external/testing/backup-only disks until they fail. What I mean is: do not worry about HDD life expectancy as long as you check their health and anticipate failures with S.M.A.R.T. monitoring tools.
Main desktop: Debian GNU/Linux Wheezy 64bit w/custom 3.14 longterm kernel - MATE 1.8.1
(i5 email@example.comGHz - 16GB DDR3 - HD6770 w/radeon driver - SSD+RAID1)