Transcript Chapter 11 Advanced NOS Administration

Chapter 11
Advanced NOS Administration
11.1 Backups
11.2 Drive Mapping
11.3 Partition and Processes Management
11.4 Monitoring Resources
11.5 Analyzing and Optimizing Network
Performance
Backups
Overview of Backup Methods
•
•
•
The backup process involves
copying data from one computer
to some other reliable storage
medium for safekeeping.
Once the data has been archived,
the system administrator can then
restore data to the system from
any previously recorded backup.
Considerations that are relevant
for storage devices:
– Cost
– Size
– Manageability
– Reliability
Overview of Backup Methods
• There are four types of backup procedures that
define how the backup will take place:
– Full - will backup everything on the hard drive at the
scheduled point in the day
– Partial - backs up selected files
– Incremental - only the files that have changed since
the last backup will be selected for back up
– Differential - backs up files created or changed since
the last normal or incremental backup
Installing Backup Software
• With the release of
Windows 2000, the backup
utility that existed with
Windows NT was greatly
enhanced.
• This Windows NT tool has
replaced RDISK, and it is
now used to back up the
Registry.
• It is no longer confined to
only tapes as output media,
and it offers a number of
other possibilities as well.
Installing Backup Software
• The Linux OS includes two utilities that can perform
backups.
• They are tar and cpio. The tar utility, which is a tape
archiever, combines multiple files into a single file
that can be copied to the media.
• The cpio utility is used to copy in/out.
• It can perform three basic actions, one of which must
be specified:
– -i To extract from an archive
– -o To make a new archive
– -p To print/pass (through the files)
Backup Hardware
• The most common backup hardware device used
before and now is some form of magnetic tape drive.
• Tape devices are known for their long-lasting
performance, which is partly due to the tape drive
mechanics that some systems include.
• There are a variety of tape devices that use different
tape formats for storing data.
• Many tape drives can also compress the data before
it is stored on the tape.
• In most cases the compression ratio is 2:1.
Backup Hardware
• Quarter Inch Cartridge
(QIC, pronounced “quick”),
is a tape standard.
• Travan tape drives have a
higher storage capacity
than the older QIC tape
drives. The most recent
standard implemented on
Travan tape drives was
hardware compression.
Backup Hardware
• 8mm tape technology uses a tape
similar to 8mm videotape and the
same helical scan system used by
a VCR.
• Mammoth 8mm tape technologies
are an improvement on the original
8mm tape technologies with higher
storage capacities and faster
transfer speeds.
• AIT technology uses 8mm tapes
that use the helical scan recording
hardware, which is much like a
VCR and have memory in the tape
cartridge, known as Memory-InCassette (MIC).
Backup Hardware
• The Digital Audio Tape (DAT) tape
standard uses 4mm digital
audiotapes to store data in the Digital
Data Storage (DDS) format.
• Digital Linear Tape (DLT) technology
offers high capacity and relatively
high-speed tape backup capabilities.
DLT tapes record information on the
tape in a linear format.
• Linear Tape-Open (LTO) technology
comes in two distinct forms.
– One is designed for high storage
capacity (Ultrium), and the other is
designed for fast access (Accelis).
Backup Strategies
• There are several accepted
backup strategies, and
many of them are based on
this popular GrandfatherFather-Son (also known as
Child-Parent-Grandparent)
method.
• This backup strategy uses
three sets of tapes for daily,
weekly, and monthly
backups.
Automating Backups
• System backups should be performed at regular intervals
to provide consistent data protection.
• Automation of such regularly scheduled events is an
important backup consideration.
• Automation not only increases backup consistency but
also gives system administrators more time to address
other important network issues.
• The most common automated backup procedure consists
of a system administrator inserting a tape into the drive at
the end of the day.
• The system performs the regularly scheduled nightly
backup of data onto the tape.
• The system administrator ejects that tape the next day
and stores it for a predetermined period of time.
Drive Mapping
What is Drive Mapping?
• Drive mapping is a useful tool
that allows an administrator to
share resources that are stored
on a server.
• The client computers that are
connected to the network
assign a drive letter that will act
as a direct path to access those
resources stored on a server
over the network.
• After a user identifies a network
resource to be used locally, the
resource can be "mapped" as a
drive.
Mapping Drives
in Windows Networks
• To map a drive with Windows
Explorer, navigate to the folder on
the remote system in Windows
Explorer by selecting Network >
Neighborhood > Server name >
Shared folder name.
• Another way to do this is to
choose the Tools menu, and then
choose Map Network Drive.
• The net use command can be
used instead of mapping drives
through Windows Explorer.
• net use can also be incorporated
into a login script that
automatically runs when the user
logs in to the network.
Mapping Drives in Linux Networks
• A client computer running Linux must be mapped in a slightly
different way.
• Use the mount command to establish a connection to the shared
directory on the server.
• Entering the syntax will map a drive to a Linux/UNIX share.
• The local directory designation that points to the remote share
denoted by the first part of the command is called the directory
mount point.
• The mount point location must already exist before a share can
be mapped to it.
mount servername:/directory/subdirectory /localdirectory
Mapping Drives in Novell Networks
• Mapping a drive to a share on a NetWare server can be done
using Windows Explorer.
• If the NetWare client machines are running Windows 95, 98,
NT, 2000, or XP, follow the same process for mapping drives
on a Microsoft network.
• A drive can be mapped at the command line by using the
map command.
• This map command can also be put in a login script that
executes automatically as the user logs into the network.
Partition and Processes Management
Using fdisk, mkfs, and fsck
• fdisk is a text-based and
requires the use of one-letter
commands to manipulate the
options.
• type m or ? at the fdisk
prompt to obtain a list of the
commands that can be used.
• Once the partition changes
have been made, a filesystem
must be created on the
partition.
• This is also referred to as
formatting the partition.
• Use the mkfs utility to create
a filesystem in Linux.
Using fdisk, mkfs, and fsck
• The fsck utility is used to
check file systems for
errors, which occur more
frequently than the need
to add, remove, or format
partitions.
• It is a good idea to use
this utility often to check
for file system integrity.
Managing System Processes
with cron Jobs
• The way to schedule tasks to run
at regular intervals on a Linux
system is with Cron Programs.
• Also known as Cron jobs, they
schedule system maintenance
tasks that are performed
automatically.
• System Cron jobs are controled via
the /etc/crontab file.
• The file begins with a set of
environmental variables. These
sets certain parameters for the
Cron jobs such as the PATH and
MAILTO
• The other lines in this file specify
the minute, hour, date, month, and
day of the week the job will run.
Core Dumps
• Core Dump is a recording of
the memory that a program
was using at the time it
crashed.
• The purpose of Core Dumps
is to allow programmers to
study the file to figure out
exactly what caused the
program to crash.
Viewing Running Processes
• The processes that are currently
running on a Linux system can
be viewed by using the ps
command.
• The ps command has a variety
of options that can be used with
it to manipulate its output.
• These options can be used
together to display specific
output using the ps command.
• There can be some
considerable output that is
generated when a command
such as ps -a --forest is
executed.
Core Dumps
• The top command functions
much like the Windows 2000
Performance tool by providing
detailed information regarding
CPU and RAM usage.
• Sometimes a process will cause
the system to lock up.
• The kill command can be used to
terminate the process.
• The signal option represents the
specified signal that is sent to the
process.
• There are 63 different parameters
that can be entered for the signal
that is sent to the process.
Assigning Permissions for Processes
• Typically, programs have the same types of permission and
can read the same files as the user who runs that program.
• There are certain programs that require additional
permission to be run by certain users.
• Regular users cannot execute the su command because it
requires root account privileges.
• Programs such as these are run using the SUID or SGID
bit, which allows these programs to be run under the
permission of another user.
Monitoring Resources
Disk Management
• By regularly using errorchecking and
defragmentation programs
and continually managing free
disk space, the system
administrator can maintain
healthy hard drives.
• One preventive disk
management tool available to
system administrators is the
use of "quotas" for user
accounts.
• A quota acts as a storage
ceiling that limits the amount
of data each user can store
on the network.
Memory Usage
• Memory diagnostic tools that
allow RAM intensive
applications to be discovered,
and stopped if necessary, are
typically built into most NOS
platforms.
• System administrators can
compensate for the lack of
physical memory through the
use of "virtual memory".
• Virtual memory allocates
space on the hard drive
(swap partition or file) and
treats it as an extension of the
system RAM.
CPU Usage
• All information used by the NOS, including the NOS itself,
is processed millions of times per second by the CPU to
display this information to the user.
• Built-in tools are commonly provided to allow system
administrators to monitor the current level of CPU activity.
• This feedback is often presented in terms of the
percentage of the CPU currently being used and is
refreshed at frequent intervals.
Reviewing Daily Logs
• Most computer programs, servers, login processes,
as well as the system kernel, record summaries of
their activities in log files.
• These summaries can be used and reviewed for
various things, including software that might be
malfunctioning or attempts to break into the system.
• In Windows 2000, the Computer Management tool
allows users to browse the logged events generated
by the NOS.
Reviewing Daily Logs
• Linux uses log daemons to
control the events that are
entered in the system log.
• Most of the Linux system’s log
files are located in the /var/log
directory.
• The log files that are located in
this directory are maintained
by the system log daemon
(syslogd) and the kernel log
daemon (klogd).
• These two daemons are
configured using the
syslog.conf file.
Checking Resource Usage on Windows 2000
and Windows XP
• System resources are monitored
in Windows 2000 and Windows
XP with the Performance tool.
• This application is found under
the Start Menu > Programs >
Administrative Tools >
Performance menu option.
• Users can then right-click on the
graph and select Add Counters
to specify which system
resources to monitor in the
graph.
Checking Resource Usage on Linux
• The df command is used to
display the amount of disk space
currently available to the various
filesystems on the machine.
• When a directory name is
specified, the du command
returns the disk usage for both the
contents of the directory and the
contents of any subdirectories
beneath it.
• The top command functions much
like the Windows 2000
Performance tool by providing
detailed information regarding
CPU and RAM usage.
Analyzing and Optimizing
Network Performance
Key Concepts in Analyzing and Optimizing
Network Performance
•
•
•
The network administrator
should make time to devise a
proactive plan for managing the
network.
This plan enables the detection
of small problems before they
become large ones.
The three key concepts in
analyzing and optimizing
network performance include:
–
–
–
Bottlenecks
Baselines
Best practices
Bottleneck
• A bottleneck is the point in
the system that limits the
data throughput, which is the
amount of data that can flow
through the network.
• The primary performancemonitoring tool for
Microsoft’s Windows 2000
Server is called
Performance.
• Performance can monitor
nearly all hardware and
software components on a
Windows 2000 server.
Bottleneck
• The various versions of the UNIX/Linux operating
systems have command-line utilities that can be used
to monitor performance of the UNIX/Linux network
server.
• The primary tools are sar, vmstat, iostat, and ps.
• The flags used by these commands can vary among
the different versions of UNIX/Linux.
• Use the UNIX/Linux man command to get specifics
about the use of these commands.
• The information displayed by the man command also
tells how to interpret the output generated by the
command.
Baselines
• The baseline measurements should
include the following statistics:
– Processor, Memory, Disk
subsystem, Network queue
length
• Determine how efficiently a network
is performing by comparing various
measurements to the same
measurements taken at an earlier
time.
• This point of comparison is called a
baseline, which is the level of
performance that is acceptable
when the system is handling a
typical workload.
Determining Internet
Connection Speed
• The speed of a connection is
limited by its lowest-speed
component or the bottleneck.
• This means that even if the
equipment is capable of a
50-kbps connection, the
connection will be at the
slower speed if the remote
modem supports only 33.6kbps.
Network Monitoring Software
• The network monitor that comes with Windows NT and
Windows 2000 is a functional and useful tool for
performing routine protocol analysis.
• Network Monitor can be used to display the individual
frames of captured data.
• The figure shows that packets for several different
protocols have been captured, including TCP, UDP, and
SMB.
Network Monitoring Software
• The Sniffer products enable
sophisticated filtering based
on pattern matches, IP/IPX
addresses, and so on.
• Sniffer Pro includes a traffic
generator to assist in testing
new devices or applications.
• It can be used to simulate
network traffic or to
measure response times
and hop counts.
• Sniffer uses a dashboardstyle interface.
Network Management Software
• Microsoft SMS is a high-end
network management
package that provides
hardware and software
inventory by installing the
client agent on target
computers.
• One of the most useful
features of SMS is its software
distribution feature.
• The distribution package
contains the information used
by SMS to coordinate the
distribution of the software.
Network Management Software
• ManageWise includes an
alarm/notification feature,
NetWare LANalyzer agent, the
management agent, Intel
LANDesk Manager, Desktop
Manager, and LANDesk virus
protection.
• Tivoli is capable of providing a
complete view of the network
topology.
• Reporting tools enable
customization of the view in
which the data is presented, and
"smart sets" can be created that
group data logically.
Management Software for Small and MediumSized Networks
• SNMP is a protocol that is
included in most
implementations of TCP/IP. It
has several advantages:
– Simplicity, low cost, relative
ease of implementation,
low overhead on the
network, supported by
most network hardware
devices
• It uses a hierarchical database
called a Management
Information Base (MIB) to
organize the information it
gathers about the network.
Management Software for Small
and Medium-Sized Networks
• Common Management Information Protocol (CMIP)
was designed to improve on SNMP and expand its
functionality.
• It works in much the same way as SNMP, but it has
better security features.
• Also, it enables notification when specified events
occur.
• Since the overhead for CMIP is considerably greater
than that which is required for SNMP, it is less widely
implemented.
• CMIP is based on the OSI protocol suite.
Management Service Provider (MSP)
• A new development in network management is the
Management Service Provider (MSP).
• A company subscribes to an MSP service, which
provides performance monitoring and network
management.
• This saves the organization the cost of buying,
installing, and learning to use monitoring and
management software.
SNMP Concepts and Components
•
•
•
There are two main parts to SNMP:
1. Management side – The management station is the centralized
location from which users can manage SNMP.
2. Agent – The agent station is the piece of equipment from which
issuers are trying to extract data.
At least one management system is needed to even be able to use
the SNMP Service.
Certain commands can be issued specifically at the management
system.
SNMP Concepts and Components
• The SNMP agent is responsible for complying with
the requests and responding to the SNMP manager
accordingly.
• Generally, the agent is a router, server, or hub.
• The agent is usually a passive component only
responding to a direct query.
• In one particular instance, however, the agent is the
initiator, acting on its own without a direct query.
• This special case is called a trap.
• A trap is set up from the management side on the
agent.
SNMP Structure and Functions
• The data that the management system
requests from an agent is contained in a
Management Information Base (MIB).
• The namespace for MIB objects is
hierarchical.
• It is structured in this manner so that each
manageable object can be assigned a
globally unique name.
• After the SNMP agent is installed, it must be
configured with an SNMP community name.
• The default SNMP community name is Public.
SNMP Structure and Functions
•
•
•
•
These three steps show what SNMP is really doing
A community is a group of hosts running the SNMP Service
to which they all belong.
There really is no established security with SNMP. The data
is not encrypted.
Notification is possible using SNMP and will vary on an
operating system–by–operating system basis.