Wrote this a couple weeks ago during some free time with a thought of helping others understand the
3 main focus - heat, shock/vibration and power. Taking care of these can help determine the useful lifespan of your hard drives.
Now in case you are wondering why I have left out water as an obvious hard drive killer, I am making the assumption that you KNOW that operating ANY electrical equipment in the presence water will fry the equipment and possibly you along with it too. So let's get on with the first part.
A) HEATIf you have started with a healthy hard drive free from manufacturing defect etc,
high temperature is the No 1 killer for your hard drive.
Google's
"Failure Trends in a Large Disk Drive Population" http://research.google.com/archive/disk_failures.pdf is a good study on general trends of hard disks in data centers. Although other studies such as Backblaze does not find any co-relations between failure rates and temperatures (Backblaze's study covers up to 30°C), Google's study tracked temperatures from 15°C to 50°C.

Google's findings showed higher failure rates for:
- Hard disk temperatures above 45°C
- Temperatures below 25°C
- Aging hard disk drives (3 years and older) with average temperatures 40°C and above
According to a similar study
"Datacenter Scale Evaluation of the Impact of Temperature on Hard Disk Drive Failures" http://www.cs.virginia.edu/~gurumurthi/papers/acmtos13.pdf released by Microsoft and the University of Virginia, when temperatures reached 50°C, average failure rate increased 79% when compared to hard disks operating in 40°C environment.
QUOTE
Temperature Management in Data Centers: Why Some (Might) Like It Hot"
http://www.cs.toronto.edu/~bianca/papers/temperature_cam.pdf concluded that there is a co-relation between hard drive temperature and Latent Sector Errors (bad sectors). Exponential increase for temperatures above 50C.
Latent sector errors (LSEs) are a common failure mode, where individual sectors on a disk become inaccessible, and the data stored on them is lost... increase in error rates tends to be linear, rather than exponential, except for very high temperatures (above 50C).
From the studies above, we can conclude that temperature above 50°C can kill your hard drive and create a lot of bad sectors in the process. Also the ideal target temperature for hard drive operation seems to be around 30°C - 45°C.
So what you can do to keep your hard disks running in the target temperature?1) Ensure good airflow in chassis. A cramped up chassis that impedes airflow will only trap heat. If you have dust-filters installed on your PC, remember to clean them regularly.
2) Make sure that you have good airflow in the room that you are running the PC. The airflow in your PC can only run as cool as the room that it is operating in so it is a great idea to turn on the air-conditioning.
3) Monitor your PC's temperature especially the harddisks more so if you tend to overclock your PC. There are plenty of software available in Windows platform - Hard Disk Sentinel, HD Tune, Speedfan, HWmonitor, CrystalDiskInfo etc. For Linux - hddtemp, lm-sensors, smartctl or the basic smartd etc. Ubuntu and Mint comes with built-in sensors app that you can turn on.
4) Hard disks spinning at 7200rpm tend to run hotter than 5k drives, so you may want to increase the rate of airflow if you have upgraded from a 5k drive to higher spin rate.
5) Use "Green" hard drives. They generally have power saving features that allows them to sip power compares to the always-on enterprise HDDs. The idea is less power being used = less heat being dissipated into the chassis.
6) Idle your hard disks when not in use. Idle drives will dissipate less heat.
7) Lastly, run an energy efficient PSU. An Inefficient PSUs will introduce heat to your chassis through dissipation as it converts electrical power. More about this later.