IT Essentials II Network Operating Systems

1
IT Essentials IINetwork Operating Systems

• Chapter 13
• Troubleshooting the Operating System

2
Identifying Problems

• Most problems can be assigned to the following
• Hardware A component has malfunctioned, or
  expected but not present.
• Kernel A bug or lack of functionality, i.e.
  module not loaded in the system kernel sometimes
  causes problems of ambiguous origin.
• Application software User level application
  software or command utilities may behave
  strangely, or simply collapse.
• Configuration System services or application
  software may be misconfigured.
• User error One of the most frequent sources of
  error conditions is caused by computer users
  attempting to do something the wrong way.
• All of the above can be categorized as
• Consistent one that is reliably and demonstrably
  occurring again and again.
• Inconsistent those that occur only sporadically,
  or under indeterminate conditions.

3
Identifying Application Problems

• Common signs of application bugs are
• Failure to execute
• program wont start up at all, main file might
  not have permission to execute.
• Or seem to start, but fails to initialize
  entirely, and either exits or stalls part way up.
• Program crashes data not saved
• Sometimes error messages are recorded
• Sometimes a core file is left behind, indicating
  the application itself suffered a catastrophic
  failure
• variant of this locked up program, application
  left running unable to proceed
• Resource Exhaustion
• refer primarily to CPU time, memory, and disk
  space
• An application consumes too much memory and
  ultimately begin to swap so badly that the whole
  system is affected
• Program Specific Misbehaviour
• To do with the running program itself

4
Configuration Problems

• Can present themselves n many ways, i.e. poor
  screen resolution when running a high end monitor
  and graphics card, Xconfigurator program may need
  to be run
• Programs that depend on networking services are
  particularly liable to cause problems
• first place to look is in the configuration file
  /etc/fstab
• If configuration problem only happens to one
  person in a group, it is liable to be caused by
  something that person did.

5
System Tools and Utilities

• Utilities return information about how the system
  or a file should be configured, but they which
  exact file or system configuration is
  misconfigured
• setserial utility provides information and set
  options for the serial ports on the system
• lpq command that helps resolve printing
  problems, display all the jobs that are waiting
  to be printed
• ifconfig - entered at the shell to return the
  current network interface configuration of the
  system
• route displays or sets the information on the
  routing of the system

6
Fixing Persistent Problems and Log Files

• Most log files are located in the /var/log
  directory or a subdirectory
• Log files can be used to
• Monitoring System Loads - Server need to handle
  requests efficiently. Log files can be used to
  determine what requests are being made that might
  cause the server to run slowly
• Intrusion Attempts and Detection - examination of
  system log files can help in finding out how and
  where the intrusion occurred, as well as what
  changes the attacker made to the system or
  network
• Normal System Functioning - log files can be
  examined to ensure the system is functioning
  normally. If something is wrong the information
  in the log files can help identify and eliminate
  them.
• Missing Entries- any log files missing entries
  can indicate something on the server is not
  functioning properly or is misconfigured
• Error Messages - many log files will contain
  various error messages that can be used to locate
  and identify problems or misconfiguration within
  the server

7
Ftab and Lilo Boot Errors

• dmesg command can be used to display the recent
  kernel messages, also known as the kernel ring
  buffer. i.e. using variants of this command you
  can find details about drivers etc.
• LILO boot loader is the first piece of code that
  takes control of the boot process form the BIOS.
  It loads the Linux kernel, and then passes
  control entirely to the Linux kernel, if LILO is
  not working properly the system wont boot.
  Following are some of the LILO error code
• No error - codes LILO hasnt loaded
• L error-code - LILO has started to boot but it is
  unable to boot the 2nd stage boot loader.
  (error-code two-digit number generated by BIOS).
• LI - LILO has started and the 1st and 2nd stage
  loaders have been loaded, but the 2nd stage
  loader wont run
• LI101010 LILO has been loaded and running
  properly but it cannot locate the kernel image
• LIL - 1st and 2nd stage loaders successfully
  loaded and are running, but LILO is unable to
  read the information it needs to work
• LIL? 2nd stage boot loader has been loaded
  correctly but is at an incorrect address
• LILO - LILO has loaded and is running, indicating
  no problem with LILO that is causing the system
  not to boot

8
Booting Without LILO

• When LILO fails completely the following can be
  used
• LOADLIN - DOS utility
• comes with the installation CDs and located in
  dosutils directory.
• to use, use a DOS partition or a DOS boot disk,
  a copy of LOADLIN.EXE, and a copy of the Linux
  kernel.
• Boot from Raw kernel on a floppy
• kernel copied to floppy disk using dd ifvmlinuz
  of/dev/fd0 command, where again vmlinuz is the
  name of the kernel
• LILO on a floppy
• Most preferred methods as it is the fastest.
• To install LILO on a floppy edit lilo.conf file.
  Changing the boot line to boot/dev/fd0.

9
Emergency Boot System

• Linux provides an emergency system copy of LILO,
  to boot system if original fails.
• Known as the Emergency Boot System.
• To use this copy of LILO configuration changes
  must be made in lilo.conf. The steps to take to
  make these changes are listed as follows
• Change where the regular disks root partition is
  mounted.
• Mount it somewhere in the emergency boot system
  like /mnt/std.
• Ensure the /boot directory is in its own
  partition. Mount it instead of or in addition to
  the root partition.
• Last changes the kernel images and other boot
  options to what is normal.
• i.e. the boot and root options should point to
  the regular hard disk.

LILO Bootlabel An example of the GUI
configuration screen in which root mount is
changed
10
Emergency Boot Disks

• There are various types of Linux Boot Disks
  available
• Linux Installation Disks included in media used
  to install the OS in the first place. At lilo
  prompt type linux rescue
• Toms Root/Boot Disk (tomsrtbt), has a bootable
  root file. downloadable from the Internet and
  will fit onto a floppy disk.
• ZipSlack - available for Slackware Linux. Can be
  installed on a small partition or on a removable
  drive slightly too big to fit a floppy.
• Demo Linux or SuSE Evaluation - one of the better
  emergency boot disk utilities available because
  it is the most complete, must be burned onto
  CD-ROM
• Custom Boot Disk required if the system being
  worked on contains hardware that needs special
  drivers or other specialities.
• Simplest method of creation modify an existing
  boot disks, by adding required extras.

11
Package Dependency Problems

• Some packages require other packages or libraries
  to run.
• Linux will usually notify users if a package has
  dependencies
• A few examples of events that can cause
  dependency problems and conflicts are listed
  below
• Missing libraries or support programs Libraries
  are a type of support code used by many programs,
  as if they were part of the program itself.
• Incompatible libraries or support programs -
  different versions of libraries and support
  programs available and correspond to current and
  past versions of programs installed. The correct
  version therefore needs to be used.
• Duplicate Files or Features - can cause programs
  to not function correctly.

12
Solutions to Package Dependency Problems

• Force the Installation
• If the error is on a package the user has
  manually compiled the source code for, then
  installation can be forced.
• Note The xxxxxxxx.rpm represents any rpm package
  
• Modify the SystemThe correct and recommended
  method solutions is to modify the system so that
  it has the necessary dependencies needed to run
  properly.
• Rebuild the Problem Package from Source CodeIn
  some instances it may be necessary to rebuild the
  package from source code if there are dependency
  error messages showing up.
• Some dependencies are caused by recompilation of
  the program and dependencies changing.
• To rebuild an RPM call the rpm with the
  rebuild command

13
Application Failure

• Difficult to spot, as they dont present in an
  obvious way. No error message will be given
  outlining the problem. Problems include

14
Troubleshooting Loss of Network connectivity

• Most basic networking problem is the inability of
  two computers to communicate. This can be due to
  hardware and/or software problem.
• First rule check for physical connectivity.
• ensure cables are properly plugged in at both
  ends
• the network adapter is functioning (check the
  link light)
• the hub status lights are on,
• and no simple hardware malfunctions have happened
  

15
Operator Error
16
TCP/IP Utilities and Troubleshooting Steps
Connectivity testing Ping Traceroute
17
Other Tools

• Windows 2000 Diagnostic tools
• Netdiag runs a standard set of network tests and
  generates a report of the results
• Pathping a combination of the ping command and
  the tracert command
• Wake-On-LAN (WOL) - used to enable an
  administrator to power up a computer by sending a
  signal (magic packet) to the NIC with WOL
  technology.

18
Disaster Recovery
Risk Analysis
19
RAID Redundacy
As well as disk, other components in the server
can be configured for redundancy, including.
Power supplies UPS Cooling fans Network
interface adapters Processors
20
Clustering

• Group of independent computers working together
  as a single system, two main types
• shared-device model, applications running in a
  cluster can access any hardware resource
  connected to any node/server in the cluster
• nothing-shared model, each node has ownership of
  a resource, so there is no competition for the
  resources
• Used to ensure mission-critical applications and
  resources are as highly available as possible
• Advantages of Clustering
• Fault tolerance can support the failure of
  components up to a complete computer without
  impacting the capability to support
  mission-critical applications.
• High availability the cluster will not be
  unavailable for reasons such as maintenance,
  upgrades, or configuration changes. A correctly
  configured cluster will come very close to 100
  availability.
• Scalability resources can be added to the
  cluster transparently to the system users.
• Easier manageability servers can be managed as a
  group, dramatically reducing the number and
  amount of management tasks when compared to an
  equivalent number of standalone servers.

• Disadvantages of clustering
• Clusters can be significantly more expensive than
  the equivalent standalone servers, due to the
  additional software and specialized hardware.
• More complex than setting up a server.

21
Hot Swapping, Warm Swapping and Hot Spares

• Hot swap (also known as hot pluggable) - the
  capability to add and remove from a computer
  while its running and have the operating system
  automatically recognize the change.
• i.e. hard drives, particularly useful in
  conjunction with a hardware RAID controller.
• Hot-spare - a component kept on hand in case of
  an equipment failure.
• Examples include
• disk drives,
• RAID controllers,
• NICs,
• and any other critical component that could be
  used to replace a failed component.
• In some mission-critical environments, an entire
  server can be designated hot spares.
• Warm swap - compromise between hot swap and hot
  spare. Generally done in conjunction with hard
  drive failure.
• Shut down the disk array before the drive can be
  replaced. And stop all I/O for that array, users
  cannot access the system.
• Referred to as a warm swap because the server
  does not have to be powered down to replace the
  drive.

22
Disaster Recovery Steps
Testing the Plan
23
The Facility Goes Down

• If the facility goes down, due to
• a natural disaster such as an earthquake or a
  flood,
• sabotage such as a bomb,
• or even just an extended power outage.
• A place to resume critical business activities
  maybe required, a disaster-recovery site. Two
  types of disaster-recovery sites are commonly
  used in the industry.

24
Questions?