Applies To: Windows Server 2008, Windows Server 2008 R2
Hardware failures occur most frequently in moving mechanical parts, such as fans and disk drives. Hardware failures may not cause the system to fail immediately. A single, small hardware failure may often go undetected until it leads to more serious failures. For instance, a burned-out cooling fan can cause a processor to overheat, resulting in increased errors and reduced performance for a long time before the hardware component fails completely.
Using some common hardware fault tolerance techniques, you can often isolate and minimize the effects of small hardware failures:
Multiple power supplies
Error correcting RAM
Hot-swappable disk drives
Multiple network adapters
You can also configure Simple Network Management Protocol (SNMP) to notify you of any detectable abnormal conditions in your hardware.