Preparing for Emergencies

1
Chapter 9

• Preparing for Emergencies

2
Objectives

• In this chapter, you will
• Understand your systems vulnerabilities and plan
  to protect data and ensure minimum downtime
• Manage the power supply to your Linux computer
• Check the integrity of your Linux file systems
• Understand how redundant disk systems can protect
  data

3
Understanding a Systems Vulnerabilities
Valued Gateway Client

• Computers continue to break down
• System administrators handle crises
• System administrators outline policies and
  procedures for emergencies

4
Creating a Disaster Plan
Valued Gateway Client

• A disaster plan describes how an organization
  will respond to various threats to its
  information systems
• A disaster plan is based on an evaluation of
  different risk levels and the value of having
  systems up and their data available
• A disaster plan includes
• Information about your computer systems
• Immediate steps that you will take to reduce
  their vulnerability
• Steps required to get them running again when
  various problems occur

5
Preventing Downtime
Valued Gateway Client

• Downtime Occasions when an organizations
  computer systems cannot respond to requests for
  information
• To prevent downtime
• Remove single points of failure
• Make vulnerable components redundant

6
Understanding High Availability
Valued Gateway Client

• Fault-tolerant computer systems that run
  continuously provide high availability
• High availability systems use special hardware
  and software to improve statistical uptime (up to
  99.999)
• A high availability cluster uses a resource group
  to handle computing tasks

7
Creating Rescue Disks
Valued Gateway Client

• A rescue disk is a floppy disk that you can use
  to boot a Linux-based computer
• A boot disk lets you start your Linux system when
  the hard disk or boot manager has been damaged

8
Maintaining Software Masters
Valued Gateway Client

• Software masters are original copies of an
  application supplied by software vendors
• Software masters and manuals should be carefully
  stored and maintained to allow recovery after a
  disaster

9
Managing the Computers Power Supply
Valued Gateway Client

• The power supply converts the AC power from a
  wall socket to the low-voltage DC power used by
  computer chips, disk drives, and other
  peripherals
• For occasions when the power supply does fail
• Keep a second power supply
• Purchase a server with a built-in backup power
  supply

10
A Computers Power Supply
11
Providing Consistent Power to a Linux System
Valued Gateway Client

• Surge suppressors protect the computers power
  supply by removing the voltage spikes and minor
  power irregularities
• When the power actually fails, an uninterruptible
  power supply (UPS) is needed

12
A Surge Suppressor Removes Voltage Spikes
13
A Computer Connected to a UPS
14
Automating Linux Shutdown
Valued Gateway Client

• A UPS allows time for a system administrator to
  shut down a Linux system gracefully
• UPS devices have a serial communications port
  that signals the Linux system in the event of a
  power outage
• In response, the Linux system executes simple
  scripts to determine what actions to take

15
Events Tracked by a UPS Devicevia a Serial Port
16
A Serial Connection between a UPS and a Linux
System
17
A Gnome-Based Graphical Interface to the Network
UPS Tools Package
18
Responding to a Power Outage
Valued Gateway Client

• Utilities that manage a UPS typically interact
  with the init program to send signals
• Init checks the /etc/inittab configuration file
  to determine what action to take
• All three signals cause to use the shutdown
  command
• shutdown f h 2
• shutdown -c
• shutdown h now

19
Checking File System Integrity
Valued Gateway Client

• The fsck utility checks the integrity of a Linux
  file system
• Each Linux ext2 or ext3 file system includes a
  superblock where parameters about the file system
  are stored
• An inode contains parameters that describe each
  file in a file system.
• File records hold a files name and point to an
  inode
• Directory records are files that contain a list
  of filenames and corresponding inode numbers

20
Checking File System Integrity
Valued Gateway Client

• A link allows two or more file records to refer
  to the same physical data stored in a file system
• A symbolic link is a file record that includes a
  path and filename, but not an inode number
• A hard link is a file record that includes a
  filename and inode, just like a regular file
  record

21
Using the fsck Utility
Valued Gateway Client

• The fsck utility is run automatically at boot
  time to check all mounted file systems
• To force a complete check of a file system, use
  the fsck command with the f option from a
  command line
• The fsck utility attempts to repair file system
  errors

22
Hard and Symbolic Links in a Linux File System
23
Defragmenting a File System
Valued Gateway Client

• All file systems are divided into units of
  storage called blocks
• Defragmenting is called disk optimization
• Defragmenting optimizes access time to files and
  reduces wear on the hard disk

24
Tuning a Linux File System
Valued Gateway Client

• The dumpe2fs command provides information about
  the superblock and data blocks of a Linux ext2 or
  ext3 file system
• You can use the tune2fs utility both to read and
  to change file system parameters

25
Understanding Redundant Disk Systems
Valued Gateway Client

• RAID (redundant arrays of inexpensive disks) a
  storage technique using multiple inexpensive hard
  disks arranged in a predefined pattern (an array)
  to improve performance, increase fault tolerance,
  or both
• You can add RAID to your system using a separate
  hardware device

26
Defining RAID Levels
Valued Gateway Client

• RAID levels differ in the amount of fault
  tolerance, the speed of reading or writing data,
  and the cost of implementation
• RAID levels include features
• Mirroring
• Duplexing
• Stripping
• Parity

27
RAID-Linear
Valued Gateway Client

• RAID-Linear A storage technique in which
  multiple physical devices are combined into a
  single logical device
• It does not provide any redundancy or fault
  tolerance, nor does it improve system performance

28
RAID-Linear
29
RAID-0 (Striping)
Valued Gateway Client

• RAID-0 A RAID level that uses striping to
  improve disk performance without adding any fault
  tolerance
• A data storage technique, in which a single block
  of data is divided into pieces and stored on more
  than one hard disk, is called striping

30
RAID-0, or Data Striping
31
RAID-1 (Disk Mirroring and Duplexing)
Valued Gateway Client

• RAID-1 A RAID level that uses disk mirroring to
  significantly improve fault tolerance and disk
  read performance
• Mirroring a redundancy technique in which the
  contents of two file systems contain identical
  information
• Duplexing a redundancy technique in which hard
  disks are accessed via different hard disk
  controllers

32
RAID-3 (Striping with Parity)
Valued Gateway Client

• RAID-3 A RAID level that uses striping with
  parity information to improve performance and
  increase fault tolerance
• Parity is a technique that allows corrupted data
  to be reconstructed using an extra piece of
  information that is created as the data is stored

33
RAID-3, Striping with Parity
34
RAID-5 (Striping and Parity)
Valued Gateway Client

• RAID-5 A RAID level in which striping with
  parity is spread across all disks in the RAID
  array compared to RAID-3, in which the parity
  information is stored on a single hard disk
• Write caching a feature of some storage systems
  in which information to be written to a RAID file
  system is stored in memory temporarily in order
  to improve the overall read/write performance of
  the file system

35
RAID-5, Parity and Striping over Multiple Disks
36
Using Hardware-Based RAID
Valued Gateway Client

• Hardware-based RAID A RAID array that is
  contained in a separate hardware device (a RAID
  subsystem) and is controlled by a CPU and other
  components separate from the CPU of the Linux
  system
• Hot-swapping Removing and replacing a failed
  hard drive or other component without turning off
  the power to the device

37
Summary

• A disaster plan describes how an organization
  will respond to various threats to its
  information systems
• Fault tolerance is achieved by removing single
  points of failure and making vulnerable
  components redundant
• High availability systems use special hardware
  and software to improve statistical uptime
• Surge suppressors and UPS devices provide clean
  power to a computer system
• A UPS can inform a Linux system of power outages
  via a serial cable, while supplying power from
  its batteries until the system can be gracefully
  shut down

38
Summary

• The fsck utility checks the integrity of a Linux
  file system at boot time or whenever the fsck
  utility is run from a command line
• Each Linux ext2 or ext3 file system includes a
  superblock where parameters about the file system
  are stored
• An inode contains parameters that describe each
  file in a file system, while file records hold a
  files name and point to an inode
• RAID systems let you improve speed, fault
  tolerance, or both, depending on the
  configuration you choose
• RAID levels include features such as mirroring,
  duplexing, striping, and parity