|
Lecture Notes
Objectives
(Start the Power Point
Presentation here. Present slides 1-7)
After reading the chapter and completing the exercises the student will be
able to identify the following failure points, resulting in loss of a fault
tolerant system:
-
Fans that provide cooling
-
Power, including interruptions, brownouts, blackouts, spikes and sags
-
Network Interface Cards (NICs)
-
Controller cards, Small Computer Systems Interface (SCSI) and Integrated
Device Enhanced IDE devices
-
Viruses, identify entry points, and security
-
Hacking by inside and outside methodologies
-
Technician errors, wrong part, wrong server
-
Data Backups, Disaster Recovery
Fans and Cooling
(Power Point Slides 8-14)
Planning Physical Site Readiness
§
Mention
changes in the physical environment such as temperature and humidity can
adversely affect the lifetime of the server and its components. Other
elements that contribute to the sites physical readiness include floor
space, power availability and weather related events such as fire and flood.
Temperature
1.
Emphasize
to students that servers run hot. Each generation on CPU’s contain more
transistors and other components that make them run hotter than the previous
generation. With today’s gigahertz (GHz) microprocessors, not only do they
require a dedicated CPU fan to help cool them down, but also the motherboard
containing the chip sets that assist the CPU’s have fan(s) attached to
them. Likewise multi-gigabyte SCSI hard disk drives run hot and oftentimes
have cooling fans attached. Other fans are found inside the server’s case
helping to dissipate server-generated heat.
2.
Server
generated heat warms the surrounding air resulting in hotter air entering
the server, compounding the heat issue, which is why server rooms are
air-conditioned. The server room temperature should be between 68 to 72
degrees F (20-22 degrees C). A constant temperature setting is important
because temperature fluctuations cause expansion and contracting of server
components, shortening the server’s life span. Ideally separate
thermostats and air conditioners should be used in the server room in
addition to internal equipment fans.
3.
Using the
server as an example, point out to your class the empty space surrounding it
and why it is important to maintain adequate space in front of, behind, and
around the server for proper airflow.
Air Quality
§
Explain to
students why air quality, keeping the air as clean and free as possible from
dust particles, is another important factor to consider for long server
life. Server rooms often have added filtration systems to reduce the level
of dust in the surrounding air. Dust is an insulator. Server motherboard
chips, and peripherals contains chips such as video and network interface
cards, if allowed to build up layers of dust, will subsequently increase in
temperature, thus destroying or reducing their overall life span.
Humidity
§
Maintaining a proper balance of relative humidity in the range between 40 to
60 percent is important to reduce the occurrence of Electrostatic discharge
(ESD), the static electricity that can shorten, damage or even destroy
several electrical components. Fortunately a side benefit of having an
air-conditioned server room is the removal of excess humidity by the
air-conditioning HVAC system.
Power
(Power Point Slides 15-17)
Availability
1.
Mention
the main power supply outlet for servers should be a grounded, separate
circuit breaker. In many areas of our country these wall outlets are
colored bright orange to designate that the electrical power is connected to
a circuit breaker separate from the building’s main power supply.
2.
In
addition to the servers main power supply, all servers should be connected
to an uninterruptible power supply (UPS) or standby power supply (SPS). Both
supply battery power in the event of a power outage. The difference between
a UPS and a SPS is that a UPS is charged with AC (alternating current) and
always runs on battery power while an SPS runs on AC and switches over to
battery power in the event of a power outage. Most companies use UPS’s,
although many smaller size companies use SPS’s because of their lower
initial cost.
3.
A third
component relating to power availability is obtaining a backup generator
ready to provide power to servers during long extended power outages. Backup
generators, however, can be very expensive and are usually only used by
companies requiring to servers to operate 24/7 (24 hours per day, seven days
per week).
Quality
§
Similar to
clean air, clean power, the absence of surges, spikes, dips, and poor
grounding, extends the life of a server and it related components. Remind
students that both the UPS and SPS also function as surge suppressors,
suppressing occasional power spikes, surges or brownouts while maintaining
the electrical current within a pre-configured voltage range.
Electromagnetic Interference
1.
Electromagnetic interference (EMI) is a by-product of electricity that
electrical machinery, equipment, motors and fluorescent lights generate.
Make sure your student know the differences between EMI and ESD
2.
If your
classroom has fluorescent lighting, point out that shielded twisted pair (STP)
network cabling that contains an outer foil jacket should be used to run
cabling in the ceiling plenum, reducing EMI giving off from the fluorescent
lights.
3.
Of course,
a certain level of EMI is unavoidable, since even a server will produce it.
Remind students of the importance of keeping server case covers attached to
the server to reduce its EMI.
Network Adaptors
(Power Point Slide 18)
§
Remind
students most Network Interface Cards (NICs) sold today comply with Plug and
Play specifications and are automatically configured by the NOS without user
intervention. Fast Ethernet PCI NICs are often 10/100 Megabytes per second
(MBps) capable and automatically detect and adapt to the proper speed. Web
network servers often employ two or more NIC cards to offer additional
bandwidth and improved performance. Several NIC cards are installed and
configured with one physical TCP/IP address and several virtual or logical
addresses to appear as a single network interface, a process known as
Adaptor Teaming.
Adaptive Fault Tolerance (AFT)
§
Adaptive
Fault Tolerance (AFT) is used to increase the availability of network
servers. Installing two or more network adaptors and configuring AFT
provides an emergency backup connection between the server and the network.
If problems occur with the first NIC, the second NIC automatically takes
over. AFT technology supports up to four adaptive teams, each team
comprised of two to four adaptors.
Adaptive Load Balancing (ALB)
§
Explain to
students ALB (Adaptive Load Balancing) guarantees a consistent level of high
server throughput by implementing multiple NIC’s and balancing the data
transmission load across them. Using Intel components, you can connect four
server adaptors to a switch, configure them for ALB and obtain an aggregate
throughput of 400 MBps. Using ALB, all of the network traffic from the
server is automatically balanced. ALB also delivers the same fault
tolerance as AFT. If one link fails, the others will continue to provide
network connectivity. Using ALB technology,
if a network bottleneck occurs, there is no need to segment the network.
Link Aggregation
§
Link
Aggregation, similar to ALB, is the process of adding multiple adaptors into
a single channeled to provide greater bandwidth. All adaptors share a
single MAC address. Link Aggregation can be implemented using Windows 2000,
Windows NT4, NetWare 4.1 and above, Linux, and Unix SCO UnixWare.
Hard Drive Components and Controllers
(Power Point Slides 19-22)
§
Mention
administrators manage dozens of hard disks. When administrators add or
replace a hard disk drive, it is important to do so with minimal disruption,
to users. Hard disk drives, multiple physical platters inside a sealed,
dust-free housing, are made in clean room environments according to rigid
manufacturing specifications. They have three major physical components:
disk platters, drive heads and the actuator mechanism.
Disk Platters
§
Ask
students to explain the phase “hard disk crash”. The disk platter is a
rigid disk inside the sealed hard disk enclosure. In the past it was made
from aluminum/magnesium alloys, coated with syrup containing iron-oxide
particles. Current manufacturing uses a glass platter that is actually a
glass-ceramic composite. Glass-ceramic composite is more thermally stable
than metal making it more difficult to “crash” the media. Glass-ceramic
media also increases the rotational speed of the platters, resulting in
faster hard disk drive access.
Drive Heads
§
Explain
the drive head reads and writes data to the magnetic media on the disk
platter. Each platter contains two heads, one that reads media on the top;
the other reads media on the bottom. The drive heads are ganged together
into a single actuator mechanism and each head moves in unison with other
heads. The actuator mechanism uses the servo mechanism to position the
drive heads.
The IDE Interface
Inform students the IDE
(Integrated Drive Electronics Drive), because it plugs into a 16-bit bus
known as AT bus, is more appropriately referred to as ATA (AT Attachment).
The IDE hard disk drives’ controller is built into the circuitry or “host
adaptor” attached to the drive. The host adaptor is the physical interface
between the hard drive and the ISA computer bus that can easily process data
from two of the fastest IDE drives attached to it.
§
Compare
other IDE devices such as CD-ROMS, ZIP drives and tape backup drives have a
variation in the ATA specification known as APAPI (ATA Packet Interface).
This chapter’s focus is on the ATA interfaces.
SCSI Interface
1.
Mention to
students the SCSI (Small Computer System Interface), like ATA, has several
standards of specifications. Similar to IDE, SCSI drives include controller
circuitry on the hard drive. Unlike IDE that communicates directly on the
system bus, the SCSI device communicates with the CPU through the host
adaptor.
2.
SCSI
standards range from SCSI-1 to SCSI-3. Note that each version of the SCSI
standards is backwards compatible with previous versions. Although you are
unlikely to find many SCSI-1 implementations today, you need to know the
basics of SCSI-1 for the CompTIA Server+ exam. Administrators must possess
knowledge of all the major SCSI standards
Troubleshooting Viruses and Security
(Power Point Slides 23-26)
§
Point out
to students that preventing viruses involves a proactive solution. An
Anti-virus utility like Norton Antivirus or McAfee Virus Scan should be
installed and configured and set to auto protect. Auto-protection
constantly scans hard drives, downloaded files from the Internet or an
inserted floppy drive for the presence of a virus. If a virus is found, the
antivirus utility alerts you and subsequently cleans or quarantines the
virus for permanent deletion.
§
Point out
that virus protection is only as good as the definition updates. All
definitions should be updated once a week or more often if updates are
released sooner.
Planning a Secure Location
§
The
administrator is responsible for ensuring that unauthorized personnel do not
access the server room.
§
Emphasize
to students that anybody, even a member of the IT staff, can damage or steal
company assets. It is important to implement sound security measures to
reduce or remove such opportunities. Remember to always log off or lock the
servers before leaving the server. Also it is prudent to log onto a server
using a user account with typical rights and permissions and to rename the
administrative account along with using a complex password for it.
Restricting Access to the Server Room
§
The server
room door should be a solid, heavy, secure door with a heavy-duty lock.
Although a simple keyed lock is probably sufficient for smaller
organizations, larger companies should consider better controlled access
like using a keypad, smart card scanner, or even emerging technology like
bio-recognition devices that use physical identifies like fingerprints or
retinal scans.
Monitoring Access to the Server Room
§
Mention to
students despite using the best security efforts, unauthorized persons may
still find a way into the server room. Monitoring access is another
important security measure. A sign-in method, video surveillance or logs
are commonly used. And do not forget to lock the server itself along with
cabinets containing hardware and software.
Should Users Be Able to Access the Server Directly?
1.
Explain to
students that some servers are not designed for direct user access. A
Dynamic Host configuration Protocol (DHCP) server, for example, requires no
direct access on the user’s part. Upon boot up, the user’s workstation
queries the network for an IP address. The DHCP server responds to the
client query, the client responds with an acknowledgement (ACK), and the
DHCP assigns the client an available IP address to use for a specified
number of days. The client computer is done with the DHCP server until it
is time to renew its lease.
2.
Emphasize
to students user’s should only access the server from their workstation.
Only administrators are allowed to access the server locally.
Technician Errors:
Right Parts, Right Server
(Power Point Slide 27)
Planning for Failure
§
Murphy’s
Law prevails! Always plan for unanticipated problems, conflicts and
incompatibilities. Adequate failsafe measures like backing up the server
can help up to quickly recover from an upgrade failure. Other failsafe
measures include clustering for mission-critical applications and having a
“hot spare”, an extra component available to perform the same functions.
Verifying System Resources
§
Administrators must verify that a free PCI slot is available prior to the
upgrade. Today’s motherboards typically include five PCI slots that quickly
fill up. Typical servers contain two network cards, a SCSI controller for
slow devices, a fast SCSI controller or two for hard drives. That leaves
one or two PCI slots free, at most. Other devices like Zip drives, hard
drives and CD-ROM’s need to have a drive bay available. Remember to check
power supply requirements to ensure that your present power supply can
handle the new load placed on it with the upgraded components. Consider
adding additional cooling fans for extra hard drives.
Making an Inventory
§
Explain
that administrators need to keep an inventory of all parts of value in each
server for proper asset tracking, budgets, and warranty service. Many
medium to large size organizations purchase inventory tracking software like
Microsoft’s System Management Server (SMS) that scans your entire network to
inventory all computers and their installed software.
Data Backups, Disaster Recovery
(Power Point Slides 28-29)
§
Discuss
with students disaster planning and the various tools used. Network
failures happen. The administrator’s primary goal is to return users network
resources as quickly as possible. One of the most important elements of
disaster planning is a tape backup system. Having mission critical data
available for restoration is a must in planning for disaster recovery.
Types of Backup Hardware
§
Ask
students to list factors to consider when purchasing a backup tape device.
First, consider the value of your data. Include the amount of data that
needs backing up, budget, NOS compatibility, and the speed of backing up and
restoring data.
§
Mention to
students there are several types of backup devices to choose from. Optical
disks such as CD-R and CD-RW are inexpensive but offer slow performance.
Tape devices are known for their durability, relatively inexpensive media,
and fast transfer rates but can be expensive.
Most backup products offer SCSI and EIDE interfaces. Small organizations
that have less than 10 GB of data to backup up should consider purchasing an
EIDE tape drive unit. EIDE tape drive units are inexpensive and offer
reasonable performance. Many EIDE tape drives use an economical Travan tape
cartridge. SCSI tape drive units are the choice for most companies because
of their high performance and reliability.
Backup Types
Review with student the
major type of backups:
1.
A
Normal
backup copies all selected files and clears the archive bit on each file
backed up. The archive bit identifies the file as having been backed up.
The next time this file is modified, the archive resets. Normal backups
take the most time but are the easiest type backup to restore, needing only
the most recent backup to restore all the selected files.
2.
An
Incremental backup copies only the files that have been created or
changed since the last Normal or Incremental backup. The Incremental backup
also marks the files as having been backed up by clearing the file archive
bit. Make sure to do a complete backup before doing the Incremental
backup. A restore involves having the last Normal
backup and all of the Incremental backup sets since the last Normal backup. The
advantage of the Incremental backup is speed. It is faster than both the
Normal and Differential backup. The disadvantage of the Incremental backup
is the longer restore time.
3.
A
Differential backup copies files that have been created or changed since
the last Normal or Incremental backup. The Differential backup does not
mark the files as having been backed up. Make sure to do a complete backup
before doing the Differential backup. To restore data, restore the last Normal
tape backup and the last Differential backup. Differential backup is faster
than the normal backup and only requires two tapes for restore. The
disadvantage is the Differential backup takes more time to back up the data.
Developing a Backup Strategy
§
Stress to
students careful planning, reliable equipment, and well-chosen tape backup
software make file recovery quicker and easier. Remind students that
testing the backup data – a test restore- should be periodically performed.
Although there are several backup strategies the most common used is the GFS
(Grandfather-Father-Son) strategy.
The Grandfather-Father-Son Strategy
Inform students the Grandfather-Father-Son strategy uses three sets of
tapes, twelve in total for daily, weekly and monthly backup sets and is
implemented as follows:
1.
Back up
the Son –
Label fours tapes Monday thru Thursday and use them for differential
backups.
2.
Back up
the Father-
Label five tapes Week 1 through Week 5. These tapes are used for weekly
Normal backups on Friday.
Back up the Grandfather – A monthly tape used for Normal backup on
the last business day of
the month. |
