HP D7171A - NetServer - LPr User manual

1
ZKLWHýSDSHU
ZKLWHýSDSHUZKLWHýSDSHU ZKLWHýSDSHU
+3ý1HW6HUYHUý 0DQDJHPHQW
RFWREHUýìäää
3UHð)DLOXUHý:DUUDQW\ý§
0LQLPL]LQJý8QSODQQHGý'RZQWLPH
Executive Summary
The Hewlett-Packard Pre-Failure Warranty1helps you run your business with
less downtime. It extends the advantage of HP’s limited warranty by providing
coverage on critical components, such as hard drives, memory and power
supplies, before they actually fail. It allows the server to monitor the status of
these critical subsystems, notifying the system administrator if the
components appear to be degrading. The Pre-Failure Warranty ensures that
when customers receive notification from system management that a critical
server component may fail, the component is replaced free of charge under
the warranty.
Whether you’re launching a business from the Web, or expanding your
current IT investments, you need to be able to keep your business up and
running 7 days a week, 24 hours a day. The HP Pre-Failure Warranty is
another tool that helps you to do just that.
A Closer Look
The cost of downtime, the time when a system is unavailable to users, can be
prohibitively expensive to a business. System configurations continue to
increase in complexity—server systems require more processors, memory,
and drives—making management and maintenance more difficult. HP,
recognizing this trend, has developed better management tools that improve
system uptime and availability. Tools that focus on problem prediction, real-
time correction and post-error recovery (key elements of a self-healing server)
provide better ways to manage critical errors that can bring down a system.
Predicting Problems
The process of predicting an event requires three different activities:
monitoring, detecting and analyzing. Monitoring entails looking at the
operating parameters of the system components. Detecting means that the
server finds the required events as they occur. Analyzing ensures that the
severity or frequency of the occurrence is understood, and if it indicates an
impending failure, then the system alerts the system administrator.
ýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýý
1This warranty covers those HP NetServers that use HP TopTools for Servers software,
version 4.5 or higher.

ZKLWHýSDSHU
2
+3ý1HW6HUYHUý0DQDJHPHQW
Real-time Correction
Once an event is predicted and an alert is issued, the system prescribes appropriate
corrective action so that downtime is prevented. In some cases, the system can
dynamically reallocate resources, in effect, creating a work-around so that the problem area
is bypassed and downtime is prevented. In other cases, the system administrator receives
a recommendation to take action. For example, the system may recommend adding disk
capacity or replacing a failing disk drive.
Post-failure Recovery
If a failure occurs, there are many hardware and software techniques that automatically
bring a server back on-line, which will minimize downtime. For example, if a hard disk drive
fails in a RAID set with a hot spare drive, then RAID automatically replicates data from the
failed drive onto the hot spare, permitting the system to recover from the failure.
HP Solutions
Hewlett-Packard offers a Pre-Failure Warranty as an additional weapon in a broad arsenal
that is designed to enable administrators to achieve the highest possible system uptime
with their server configurations. The HP Pre-Failure Warranty uses agents and event
reporting as the means to monitor and predict potential problems. The Pre-Failure Warranty
covers the key subsystems and components that are discussed in the following sections.
Event Reporting
Every hardware subsystem in a HP NetServer, such as disk storage, system memory, or
power supplies, has a robust set of management capabilities. The tools that provide access
to this information are HP TopTools for Servers and the TopTools Remote Control Card.
HP NetServer agents are used to report incidents to an event log. An event log is a
collection of entries, automatically recorded by the monitoring software, which reports
failures or potential failure conditions.
HP TopTools for Servers is an easy-to-use, standards-based application that allows
network administrators to manage and maintain servers remotely. It is a graphically intuitive
tool based on the Simple Network Management Protocol (SNMP) and the Desktop
Management Interface (DMI). It provides integrated management of the server
environment. HP TopTools for Servers displays configuration information including
hardware components, operating statistics, operating system (OS) device driver version
numbers, firmware version numbers, an event log for a complete history of server activity
including pre-failure alerts indicating potential system problems. The HP TopTools Remote
control card provides remote management capabilities regardless of the state of the
system. The event log can be accessed using this card even if the operating system or
power is down.

ZKLWHýSDSHU
3
+3ý1HW6HUYHUý0DQDJHPHQW
HP NetServer Agents
HP NetServer Agents unlock the monitoring capabilities designed into HP NetServers.
These Agents monitor the current state of server subsystems by measuring certain
operating parameters, counting the occurrence of particular events (for example, the
number of read operations performed on a disk drive), and report on the state of critical
functions (such as whether the system cooling fan is operating).
The Agents provide information to management applications and can generate alarm
notifications for the system administrator if significant changes occur in the fault or
performance aspects of system operation. For example, when an Agent detects that the
hard disk drive exceeded a preset threshold, and is likely to fail, then the Agent issues an
alert through SNMP. HP TopTools for Servers receives the alert and immediately issues an
alarm to the system administrator. The process of monitoring and sending alerts is called
the Pre-Failure Monitor Service.
Alerts and Thresholds
The HP NetServer Agents detect the health of the server (and server components) by
measuring and collecting server operating parameters and comparing the measured values
to a set of pre-determined and user-established values called thresholds. A threshold
defines the minimum, maximum, or specific value for a given fault or performance
parameter. If a value of a server parameter falls outside of its threshold, the Agents
immediately generate an alarm that is displayed and can be forwarded to a pager or email
so that the administrator is promptly notified.
All alerts are recorded in the Event Log. The Event Log within the Status Tab displays a
summary of events received by the management system, which can be viewed, deleted,
printed or saved to a text file.
See Figure 1 for an example of the information provided when a memory module is likely to
fail.

ZKLWHýSDSHU
4
+3ý1HW6HUYHUý0DQDJHPHQW
Figure 1. HP TopTools information provided during a memory module failure
Subsystems and Components Covered by the Pre-Failure Warranty
The HP Pre-Failure Warranty covers all HP NetServers using HP TopTools for Servers,
during the normal warranty period.2The Pre-Failure Warranty specifically applies to hard
disk drives, memory, and power supplies.3(See Table 1 for information on coverage by
platform.) Since the memory, power supplies and hard disk drives represent the three
major building blocks of the system, the failure of any one may be catastrophic.
The HP Pre-Failure Warranty covers only HP-made or HP-qualified components. Other
third-party vendors’ products are not warranted.
Table 1. HP Pre-Failure Warranty Coverage by Platform
Product Hard Disk Drives Memory Power Supplies
HP NetServer E Series X
HP NetServer L Series (except below) X X
HP NetServer LXr 8000 & LXr 8500 X X X
ýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýýý
2Pre-failure alerts for memory require HP TopTools 4.5 or higher. Pre-failure alerts for disk drives and power
supplies require HP TopTools 3.11 or higher.
3Pre-failure warranty on power supplies is available on the HP NetServer LXr 8000, LXr 8500, and above.

ZKLWHýSDSHU
5
+3ý1HW6HUYHUý0DQDJHPHQW
Hard Disk Drives
All hard disk drives shipped with HP NetServers or available as accessories are now
SMART (Self-Monitoring and Reporting Technology)-enabled.4These 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
4Only SMART-enabled hard disk drives will issue pre-failure alerts.

ZKLWHýSDSHU
6
+3ý1HW6HUYHUý0DQDJHPHQW
commonplace in DRAM cells, cause data to be stored incorrectly in a memory location, but
subsequent accesses can correctly store data into that cell. With properly tuned hardware
and software, occurrences of soft errors should be months or possibly years apart. In
reality, the environment and other hardware and software conflicts can cause noise or
contention that may precipitate more frequent random memory errors.
To protect the server from data loss associated with these errors, HP offers ECC memory
scrubbing as a diagnostic and repair tool. Basic ECC memory detects and corrects single-
bit memory errors, however, the errors themselves still remain in RAM. Each time the bad
data bit is accessed, it must be re-corrected by the memory controller. Two versions of
memory scrubbing are available—proactive and reactive. Reactive memory scrubbing
improves on basic ECC memory by writing the corrected data to RAM, after the error
occurs. Proactive memory scrubbing scrubs memory at pre-scheduled times or upon
demand. In either case, when a memory error is detected, the corrected data is written
back to RAM. This action ensures that all ECC memory errors will be logged and the bad
data bits are “scrubbed” from RAM.
The Pre-Failure Warranty covers the replacement of Single In-Line or Dual In-Line Memory
Modules (SIMMs or DIMMs) used in the server’s main memory when the predefined
thresholds for correctable (soft) errors have been exceeded. Soft errors are recorded in the
Event log and a pre-failure alert is issued through HP TopTools for Servers.
Third-Party Memory
HP recognizes that purchasing or upgrading a mission-critical server requires confidence
that the server possesses high-quality memory compatible with the other components and
subsystems, so HP subjects its memory modules to a tough testing and qualification
process. HP quality and compatibility testing identifies the most reliable memory for use in
each HP NetServer. HP is confident of the quality and testing performed on its
components; therefore, the company warrants all products manufactured or distributed
by HP. HP warranties
do not cover
third-party memory, nor the damage that third-party
memory might cause.
Power Subsystem and Voltage Regulator Module
An uninterruptible power supply (UPS) is the most cost-effective step in increasing overall
system availability. Once you have protected and conditioned the power reaching the
system itself, the next component to consider is the power supply. Because power supplies
have a lower meantime between failure (MTBF) than digital electronic circuits, it is a good
idea to have redundancy built into the system. HP NetServers have redundant power
supplies included throughout the L Series.
The HP Pre-Failure Warranty for the power supplies provides an additional measure of
protection. The HP NetServer monitors the power supply voltages (+5v, +12v, -12v and
3.3v). Additionally, since a Voltage Regulator Modulator (VRM) powers each processor in a
HP NetServer, the system also monitors the voltage of each VRM.

ZKLWHýSDSHU
7
+3ý1HW6HUYHUý0DQDJHPHQW
Each voltage has an upper and lower threshold for which an alert can be issued. If any
voltage is out of range, an error is logged in the Event log. For an extreme voltage
deviation from range, the system issues shutdown commands to prevent hardware damage
and protect customer data from unstable circuitry that can result from fluctuating voltage
levels. (Thresholds can be set so that the system will shut down upon over-temperature
conditions as well.)
In addition, on certain HP NetServers, the power supplies will generate a pre-failure
condition. Again, this pre-failure condition is monitored by the system and alerts are
generated and logged in the Event log, recording the specific power supply that has
encountered the pre-failure condition, which will be replaced under the system warranty.
Processors
Servers with multiple CPU processors are commonplace. Typically these servers are
designed for high-performance applications that require more CPU power than is currently
available from a single processor. CPU faults are usually unrecoverable errors, meaning
that neither the operating system nor the hardware can manage the fault conditions. If a
server encounters a CPU fault, the operating system halts and users lose their network
connections. When this happens, the CPU needs to be taken offline for reconfiguration,
after which the servers can be restarted (sometimes automatically) and users can
reconnect to the server.
Some server manufacturers have chosen to monitor certain aspects of processor functions,
such as the L2 cache error corrections. HP has studied this failure mode and has
determined that its frequency of occurrence is so low that monitoring it for a Pre-Failure
Warranty is not appropriate. HP has instead implemented fault-resilient boot, so that a
failed processor is removed from the configuration during POST.
How to Take Advantage of the HP Pre-Failure Warranty
This section details how system administrators are warned of potential Pre-Failure
Warranty events. Because HP TopTools is the standard monitoring mechanism, the
procedures are described in terms of what happens with HP TopTools.
HP TopTools identifies pre-failure conditions by displaying a yellow status indicator to show
that a component is degraded. Along with the yellow color code, HP TopTools also displays
messages recommending replacement of the component in a pre-failure condition.
Upon receiving a HP TopTools alert that a component has degraded, the system
administrator follows the on-screen instructions or contacts an authorized HP service
provider to determine procedures and eligibility for replacement of the affected component
under warranty.

ZKLWHýSDSHU
8
+3ý1HW6HUYHUý0DQDJHPHQW
If any component fails or indicates a pre-failure condition, HP Support will require the
following information:
•HP NetServer Agent version level
•HP TopTools for Servers version level
•Product serial number(s)
•Product model name(s) and number(s)
•Applicable error messages or pre-failure alert
•Add-on boards or hardware
•Third-party hardware or software
•OS/NOS type and revision level
Conclusion
Hewlett-Packard offers the Pre-Failure Warranty as another tool that helps you efficiently
manage your servers and associated service and support, ensuring maximum availability
and performance. Pre-Failure Warranty extends the advantage of the HP industry-leading
three-year limited warranty by applying it to critical server components before they actually
fail.
HP provides coverage on components with the highest potential for failure, based on
studies and historical service rates. HP continues to monitor new technologies and analyze
system failure trends so that HP NetServers will include the appropriate components in the
Pre-Failure Warranty program.
For additional program details, contact your Authorized HP Reseller or the local
HP Customer Care Center.

ZKLWHýSDSHU
9
+3ý1HW6HUYHUý0DQDJHPHQW
For More Information
For more information on HP NetServers, see:
http://www.hp.com/go/netserver
For a complete list of current white papers and technology briefs, see:
http://www.netserver.hp.com/netserver/papers
Intel is a registered trademark of Intel Corporation. Microsoft andWindows are registered trademarks of, and Windows NT is a trademark of, Microsoft
Corp. All other brand names are trademarks of their respective owners.
Technical information in this document is subject to change without notice.
© Copyright Hewlett-Packard Company 1999. All Rights Reserved.