HP D7171A - NetServer - LPr Manual page 5

ZKLWHýSDSHU
+ 3 ý 1 H W 6 H U Y H U ý 0 D Q D J H P H Q W
Hard Disk Drives
All hard disk drives shipped with HP NetServers or available as accessories are now
4
SMART (Self-Monitoring and Reporting Technology)-enabled. These drives have been
designed with the capability to inform the host (system administrator) when a drive is
experiencing abnormal operation that is likely to lead to drive failure. Indicators (attributes)
that correlate to imminent drive failure have been determined through research. Some of
these indicators include start times, short/average/long seek times and recoverable and
unrecoverable data error rates. SMART drives constantly take internal measurements
of these indicators to determine performance or operation deterioration. If a SMART
attribute crosses a predefined threshold, then an error is reported to the system
administrator through HP TopTools for Servers. By sending this warning in advance of the
drive failure, HP alerts the administrator to schedule drive replacement prior to drive failure.
HP recommends that, in most cases, the drive should be replaced within 24 hours of a
SMART report.
Of course, no system can predict all possible failure mechanisms. Server components,
including hard disk drives, are simply too complex. The HP method of alerting the system
administrator through the Pre-Failure Warranty provides reasonable protection against a
significant number of possible failure scenarios.
Third-Party Drives
Before HP ships any drive model, it must pass a stringent battery of qualification tests. The
HP testing process has been customized over many years of experience with disk drives.
HP continues to improve the testing process so that even the most obscure problems may
be detected. Because third-party drives have not undergone the HP qualification process,
the HP Pre-Failure Warranty does not cover them. In fact, HP customers who have
purchased certain hot-swappable drives that are advertised as 100%-compatible, but that
have not been qualified by HP, may receive false pre-failure notifications or even no
notification at all.
Memory
Memory system integrity is essential to keeping complex networking systems up and
available. A memory subsystem failure can significantly affect a company's productivity.
Basic fault-management components, such as Error Correction Code (ECC) memory and
the Pre-Failure Warranty, are necessary to support memory-intensive applications such as
databases. As both the memory configurations and the user population increase in size,
the potential increases that a memory failure will negatively impact a great number of
users.
Two types of memory errors can occur. A hard error is a physical failure within a Dynamic
Random Access Memory (DRAM) cell that prevents data from being stored reliably in one
or more locations. Failure of a single DRAM cell can abruptly halt a system. Soft errors
often result from a temporary loss of charge in a DRAM cell. These errors, which are
ýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýý
4
Only SMART-enabled hard disk drives will issue pre-failure alerts.
5