Tài liệu Troubleshooting Startup Problems phần 1 doc

Troubleshooting Startup Problems
Windows Server 2003 is certainly the most reliable version of Windows, possessing a
level of robustness that its predecessors lack, even Windows 2000 and Windows XP.
Does this mean startup problems cannot occur in Windows Server 2003? No, it doesn't.
No existing operating system can be considered crash- or corruption-proof, not even
specialized operating systems used by military organizations. Any operating system can
be rendered unbootable, and the newest release of Windows is no exception. This issue
becomes extremely important in a large-scale corporate network, which depends on the
availability of the network operating system (OS)-particularly for network servers. You
must be prepared to handle anything that goes terribly wrong and to avoid feeling
helpless. A significant part of this chapter is dedicated to troubleshooting startup
problems. These, I admit, are the most frustrating ones, especially if your system won't
boot when you have a lot of work to do. So, what should you do in an emergency? First,
don't panic. Next, try to detect what is preventing the operating system from booting.
Since boot sequence in Windows Server 2003 closely resembles that in Windows
2000/XP, most (but not all) techniques described here can be applied to all three versions.
A detailed description of Windows Server 2003 boot sequence was provided in Chapter
6. Therefore, I will provide only a short explanation of the boot sequence, and then
proceed with problem detection and troubleshooting.
Table 12.1
lists Windows Server 2003 startup phases with brief descriptions of the
processes that take place at each stage of the normal boot process.

Table 12.1: Windows Server 2003 Startup Process
Startup stage Description (x86-based systems)
POST routine CPU initiates the system board POST routines. POST routines of the
individual adapters start after the motherboard POST is
accomplished successfully.
Initial startup
process
The system searches for a boot device according to the boot order

setting stored in CMOS. If the boot device is a hard disk, Ntldr starts.
Operating system
load
Ntldr switches the CPU to protected mode, starts the file system,
then reads the contents of the Boot.ini file. This information
determines the startup options and initial boot menu selections.
Hardware detection
and configuration
selection
Ntdetect.com gathers basic hardware configuration data and passes
this information to Ntldr. If more than one hardware profile exists,
Windows XP and Windows Server 2003 attempt to use the correct
one for the current configuration. Notice that if your computer is
ACPI-compliant, Windows XP or Windows Server 2003 ACPI
Table 12.1: Windows Server 2003 Startup Process
Startup stage Description (x86-based systems)
functionality will be used for device enumeration and initialization.
(More information on this topic was provided in Chapter 5
.)
Kernel loading Ntldr passes the information collected by Ntdetect.com to
Ntoskrnl.exe. Ntoskrnl then loads the kernel, HAL, and registry
information. A status bar at the bottom portion of the screen indicates
progress.
Operating system
logon process
Networking-related components (such as TCP/IP) load
asynchronously with other services, and the Begin Logon prompt
appears on screen. After a user logs on successfully, Windows
updates the Last Known Good Configuration information to reflect
the current state.

New devices are
detected by Plug
and Play
If Windows XP or Windows Server 2003 detects new devices, they
are assigned system resources. The operating system extracts the
required driver files from the Driver.cab file. If this file is not found,
Windows XP or Windows Server 2003 prompts the user to provide
them. Device detection occurs asynchronously with the operating
system logon process.

Diagnosing Startup Problems
Fortunately, boot failures in Windows XP and Windows Server 2003 are rare, especially
if you perform regular maintenance and take preventive measures against disaster.
However, problems still can arise. As with any other operating system, they might be
caused both by hardware malfunctions and by software errors. If the problem is severe
enough, the system stops booting and displays an error message. A brief list of error
messages and their meanings is presented in Table 12.2
. Although this list is by no means
comprehensive, it covers the most common problems that can cause startup failures of
Windows NT-based operating systems, including Windows 2000, Windows XP, and
Windows Server 2003.

Table 12.2: Startup Problem Symptoms
Startup problem
symptom
Possible cause
The POST routine emits
a
series of beeps and
displays

The system self-test routines stopped because of improperly
installed devices.
To recover from hardware problems, carefully review the
Table 12.2: Startup Problem Symptoms
Startup problem
symptom
Possible cause
error messages, for
example:
Hard disk error.
Hard disk absent/
failed.
documentation supplied with your system and perform the
basic hardware checks. Verify that all cables are attached
correctly and all internal adapters are installed properly. Make
sure that all peripheral devices (such as keyboards) necessary to
complete the POST without error messages are installed and
functioning. If applicable, verify that you have configured
correctly all jumpers or dual in-line package (DIP) switches.
Jumpers and DIP switches are especially important for hard
disks. Run diagnostic software to detect hardware malfunction,
and replace the faulty device.
Unfortunately, the topic of troubleshooting hardware problems
goes beyond the range of problems discussed in this book. It
deserves a separate comprehensive volume. However, I can
recommend some resources on the topic that would help you
make sense of the BIOS error codes:
•
BIOS Survival Guide, available at
/>os_sg.htm

•
Definitions and Solutions for BIOS Error Beeps and
Messages/Codes, available at

CMOS or NVRAM
settings are not retained
The CMOS memory is faulty, data is corrupt, or the battery
needs replacing.
Master boot record
(MBR)-related error
messages similar to the
following:
Missing operating
system.
Insert a system
diskette and restart
the system.
The MBR is corrupt.
The easiest method of recovering the damaged MBR is
provided by Recovery Console (the methods of starting
Recovery Console were discussed in Chapter 2
). Once you are
in Recovery Console, use the FIXMBR command to repair the
MBR.
The FIXMBR command uses the following syntax:
Fixmbr [device_name]
The parameter device_name specifies the drive on which you
Table 12.2: Startup Problem Symptoms
Startup problem
symptom

Possible cause
need to repair the damaged MBR. For example:
fixmbr \Device\HardDiskO
If the device_name parameter is omitted, the new MBR will be
written to the boot device, from which your primary system is
loaded. Notice that you'll be prompted to confirm your
intention to continue if an invalid partition table is detected.
Partition table-related
error message similar to
the following:
Invalid partition
table.

A disk-read error
occurred.
The partition table is invalid.
You can recover from this problem using the DiskProbe
Resource Kit utility or any third-party low-level disk editor.
Note that to prevent this problem, you must create a backup
copy of the MBR beforehand. (You can use the DiskProbe tool
for this purpose.) Detailed information on this topic can found
in the Resource Kit documentation.
If the MBR on the disk used to start Windows is corrupt, most
likely you will be unable to start Windows XP or Windows
Server 2003 (and, consequently, DiskProbe). Therefore, before
proceeding any further, you'll need to start Recovery Console
to replace the damaged MBR.
Boot failure caused by
disk or file system
corruption, not related to

damaged MBR or
partition table
Start Recovery Console and run the CHKDSK command to
repair the disk. If this proves to be insufficient, you will need to
take additional actions to fully recover the damaged file
system.
Windows XP or
Windows Server 2003
cannot start after you
have installed another
operating system
The Windows XP or Windows Server 2003 boot sector was
overwritten by the other operating system's setup program.
Recovery Console provides the FIXBOOT command that
enables you to restore the overwritten boot sector.
Missing Boot.ini,
Ntoskrnl.exe, or
Ntdetect.com files (x86-
based systems)
Required startup files are missing or damaged, or entries in the
Boot.ini are pointing to the wrong partition.
Start the Recovery Console and use available commands, such
as REN, DEL, or COPY, to restore working copies of boot
files.
Table 12.2: Startup Problem Symptoms
Startup problem
symptom
Possible cause
Bootstrap loader error
messages similar to the

following:
Couldn't find loader

Please insert another
disk.
Ntldr is missing or corrupt.
If Ntldr or any other file required to boot the system is missing
or corrupt, start Recovery Console and copy the required file.
Windows NT-based OS
cannot start and displays
message similar to the
one provided below:
Windows could not
start because the
following file is
missing or corrupt:


\WINNT\SYSTEM32\C
ON
FIG\SYSTEM

You can attempt to
repair this file by
starting
Windows Setup using
the original Setup
floppy disk of CD-
ROM.

Select 'r' at the
first screen to
repair.
This and similar error messages specifying different file names
indicate that the boot failure was caused by a damaged registry
hive(s) or by invalid registry settings.
First, try to boot using the safe mode startup option. If your
attempt has failed, try the Last Known Good boot option. If
you still can't boot successfully, start Recovery Console and
use the COPY command to restore known good registry files
(for example, those located in the %SystemRoot%\Repair
folder) to the %SystemRoot%\System32\Config folder.
If the problem is related to settings for a specific service or
driver, you also may be able to use Recovery Console's
DISABLE command to disable the offending service or driver.
Boot failure caused by a
video display driver
problem
Use the safe mode startup option, then repair or replace the
driver.
Boot failure caused by As a first line of defense, try to boot in safe mode and disable